Download a
Free via
For HPC, contact [email protected]
Group | Test1 | Test2 |
0 | 86 | 83 |
0 | 93 | 79 |
0 | 85 | 81 |
0 | 83 | 80 |
0 | 91 | 76 |
1 | 94 | 79 |
1 | 91 | 94 |
1 | 83 | 84 |
1 | 96 | 81 |
1 | 95 | 75 |
* First enter the data manually; input str10 sex test1 test2 "Male" 86 83 "Male" 93 79 "Male" 85 81 "Male" 83 80 "Male" 91 76 "Female" 94 79 "Fem ale" 91 94 "Fem ale" 83 84 "Fem ale" 96 81 "Fem ale" 95 75 end
* Next run a paired t-test; ttest test1 == test2
* Create a scatterplot; twoway ( scatter test2 test1 if sex == "Male" ) ( scatter test2 test1 if sex == "Fem ale" ), legend (lab(1 "Male" ) lab(2 "Fem ale" ))
* First enter the data manually; data example; input sex $ test1 test2; datalines ; M 86 83 M 93 79 M 85 81 M 83 80 M 91 76 F 94 79 F 91 94 F 83 84 F 96 81 F 95 75 ; run ;
* Next run a paired t-test; proc ttest data = example; paired test1*test2; run ;
* Create a scatterplot; proc sgplot data = example; scatter y = test1 x = test2 / group = sex; run ;
# Manually enter the data into a dataframe dataset <- data.frame(sex = c("Male", "Male", "Male", "Male", "Male", "Female", "Female", "Female", "Female", "Female"), test1 = c( 86 , 93 , 85 , 83 , 91 , 94 , 91 , 83 , 96 , 95 ), test2 = c( 83 , 79 , 81 , 80 , 76 , 79 , 94 , 84 , 81 , 75 ))
# Now we will run a paired t-test t.test(dataset$test1, dataset$test2, paired = TRUE )
# Last let's simply plot these two test variables plot(dataset$test1, dataset$test2, col = c("red","blue")[dataset$sex]) legend("topright", fill = c("blue", "red"), c("Male", "Female"))
# Making the same graph using ggplot2 install.packages('ggplot2') library(ggplot2) mygraph <- ggplot(data = dataset, aes(x = test1, y = test2, color = sex)) mygraph + geom_point(size = 5) + ggtitle('Test1 versus Test2 Scores')
sex = { 'Male' , 'Male' , 'Male' , 'Male' , 'Male' , 'Female' , 'Female' , 'Female' , 'Female' , 'Female' }; t1 = [86,93,85,83,91,94,91,83,96,95]; t2 = [83,79,81,80,76,79,94,84,81,75];
% paired t-test [h,p,ci,stats] = ttest(t1,t2)
% independent samples t-test sex = categorical(sex); [h,p,ci,stats] = ttest2(t1(sex== 'Male' ),t1(sex== 'Female' ))
plot(t1,t2, 'o' ) g = sex== 'Male' ; plot(t1(g),t2(g), 'bx' ); hold on; plot(t1(~g),t2(~g), 'ro' )
& Syntax | Gradual | Moderate | Moderate Scope | Good | Custom Tables, ANOVA and Multivariate Analysis | |
& Syntax | Gradual | Strong | Moderate Scope | Great | Design of Experiments, Quality Control, Model Fit | |
Menus & | Moderate | Strong | Broad Scope | Good | Panel Data, Mixed Models, Survey Data Analysis | |
Syntax | Steep | Very Strong | Very Broad Scope | Very Good | Large Datasets, Reporting, Password Encryption, Components for Specific Fields | |
Syntax | Steep | Very Strong | Very Broad Scope | Excellent | Graphic Packages, Machine Learning, Predictive Modeling | |
Syntax | Steep | Very Strong | Limited Scope | Excellent | Simulations, Multidimensional Data, Image and Signal Processing |
*The primary interface is bolded in the case of multiple interface types available.
Home Market Research
Content Index
Types of data in research, finding patterns in the qualitative data, methods used for data analysis in qualitative research, preparing data for analysis, methods used for data analysis in quantitative research, considerations in research data analysis, what is data analysis in research.
Definition of research in data analysis: According to LeCompte and Schensul, research data analysis is a process used by researchers to reduce data to a story and interpret it to derive insights. The data analysis process helps reduce a large chunk of data into smaller fragments, which makes sense.
Three essential things occur during the data analysis process — the first is data organization . Summarization and categorization together contribute to becoming the second known method used for data reduction. It helps find patterns and themes in the data for easy identification and linking. The third and last way is data analysis – researchers do it in both top-down and bottom-up fashion.
LEARN ABOUT: Research Process Steps
On the other hand, Marshall and Rossman describe data analysis as a messy, ambiguous, and time-consuming but creative and fascinating process through which a mass of collected data is brought to order, structure and meaning.
We can say that “the data analysis and data interpretation is a process representing the application of deductive and inductive logic to the research and data analysis.”
Researchers rely heavily on data as they have a story to tell or research problems to solve. It starts with a question, and data is nothing but an answer to that question. But, what if there is no question to ask? Well! It is possible to explore data even without a problem – we call it ‘Data Mining’, which often reveals some interesting patterns within the data that are worth exploring.
Irrelevant to the type of data researchers explore, their mission and audiences’ vision guide them to find the patterns to shape the story they want to tell. One of the essential things expected from researchers while analyzing data is to stay open and remain unbiased toward unexpected patterns, expressions, and results. Remember, sometimes, data analysis tells the most unforeseen yet exciting stories that were not expected when initiating data analysis. Therefore, rely on the data you have at hand and enjoy the journey of exploratory research.
Create a Free Account
Every kind of data has a rare quality of describing things after assigning a specific value to it. For analysis, you need to organize these values, processed and presented in a given context, to make it useful. Data can be in different forms; here are the primary data types.
Learn More : Examples of Qualitative Data in Education
Data analysis and qualitative data research work a little differently from the numerical data as the quality data is made up of words, descriptions, images, objects, and sometimes symbols. Getting insight from such complicated information is a complicated process. Hence it is typically used for exploratory research and data analysis .
Although there are several ways to find patterns in the textual information, a word-based method is the most relied and widely used global technique for research and data analysis. Notably, the data analysis process in qualitative research is manual. Here the researchers usually read the available data and find repetitive or commonly used words.
For example, while studying data collected from African countries to understand the most pressing issues people face, researchers might find “food” and “hunger” are the most commonly used words and will highlight them for further analysis.
LEARN ABOUT: Level of Analysis
The keyword context is another widely used word-based technique. In this method, the researcher tries to understand the concept by analyzing the context in which the participants use a particular keyword.
For example , researchers conducting research and data analysis for studying the concept of ‘diabetes’ amongst respondents might analyze the context of when and how the respondent has used or referred to the word ‘diabetes.’
The scrutiny-based technique is also one of the highly recommended text analysis methods used to identify a quality data pattern. Compare and contrast is the widely used method under this technique to differentiate how a specific text is similar or different from each other.
For example: To find out the “importance of resident doctor in a company,” the collected data is divided into people who think it is necessary to hire a resident doctor and those who think it is unnecessary. Compare and contrast is the best method that can be used to analyze the polls having single-answer questions types .
Metaphors can be used to reduce the data pile and find patterns in it so that it becomes easier to connect data with theory.
Variable Partitioning is another technique used to split variables so that researchers can find more coherent descriptions and explanations from the enormous data.
LEARN ABOUT: Qualitative Research Questions and Questionnaires
There are several techniques to analyze the data in qualitative research, but here are some commonly used methods,
LEARN ABOUT: 12 Best Tools for Researchers
The first stage in research and data analysis is to make it for the analysis so that the nominal data can be converted into something meaningful. Data preparation consists of the below phases.
Data validation is done to understand if the collected data sample is per the pre-set standards, or it is a biased data sample again divided into four different stages
More often, an extensive research data sample comes loaded with errors. Respondents sometimes fill in some fields incorrectly or sometimes skip them accidentally. Data editing is a process wherein the researchers have to confirm that the provided data is free of such errors. They need to conduct necessary checks and outlier checks to edit the raw edit and make it ready for analysis.
Out of all three, this is the most critical phase of data preparation associated with grouping and assigning values to the survey responses . If a survey is completed with a 1000 sample size, the researcher will create an age bracket to distinguish the respondents based on their age. Thus, it becomes easier to analyze small data buckets rather than deal with the massive data pile.
LEARN ABOUT: Steps in Qualitative Research
After the data is prepared for analysis, researchers are open to using different research and data analysis methods to derive meaningful insights. For sure, statistical analysis plans are the most favored to analyze numerical data. In statistical analysis, distinguishing between categorical data and numerical data is essential, as categorical data involves distinct categories or labels, while numerical data consists of measurable quantities. The method is again classified into two groups. First, ‘Descriptive Statistics’ used to describe data. Second, ‘Inferential statistics’ that helps in comparing the data .
This method is used to describe the basic features of versatile types of data in research. It presents the data in such a meaningful way that pattern in the data starts making sense. Nevertheless, the descriptive analysis does not go beyond making conclusions. The conclusions are again based on the hypothesis researchers have formulated so far. Here are a few major types of descriptive analysis methods.
For quantitative research use of descriptive analysis often give absolute numbers, but the in-depth analysis is never sufficient to demonstrate the rationale behind those numbers. Nevertheless, it is necessary to think of the best method for research and data analysis suiting your survey questionnaire and what story researchers want to tell. For example, the mean is the best way to demonstrate the students’ average scores in schools. It is better to rely on the descriptive statistics when the researchers intend to keep the research or outcome limited to the provided sample without generalizing it. For example, when you want to compare average voting done in two different cities, differential statistics are enough.
Descriptive analysis is also called a ‘univariate analysis’ since it is commonly used to analyze a single variable.
Inferential statistics are used to make predictions about a larger population after research and data analysis of the representing population’s collected sample. For example, you can ask some odd 100 audiences at a movie theater if they like the movie they are watching. Researchers then use inferential statistics on the collected sample to reason that about 80-90% of people like the movie.
Here are two significant areas of inferential statistics.
These are sophisticated analysis methods used to showcase the relationship between different variables instead of describing a single variable. It is often used when researchers want something beyond absolute numbers to understand the relationship between variables.
Here are some of the commonly used methods for data analysis in research.
LEARN ABOUT: Best Data Collection Tools
LEARN MORE: Descriptive Research vs Correlational Research The sheer amount of data generated daily is frightening. Especially when data analysis has taken center stage. in 2018. In last year, the total data supply amounted to 2.8 trillion gigabytes. Hence, it is clear that the enterprises willing to survive in the hypercompetitive world must possess an excellent capability to analyze complex research data, derive actionable insights, and adapt to the new market needs.
LEARN ABOUT: Average Order Value
QuestionPro is an online survey platform that empowers organizations in data analysis and research and provides them a medium to collect data by creating appealing surveys.
Aug 14, 2024
Aug 13, 2024
Aug 9, 2024
Software for data analysis.
Meet one-on-one with a research data specialist from the library or NUIT.
Request a Consult
Open-source software.
Both R and Python are free and open source. NUIT's Research Data Services offers installation guidelines:
Northwestern provides access to licensed software in the library computer labs and on NUWorkspace , a virtual desktop. NUIT also makes free or discounted software licenses available. In addition to these campus-wide resources, your department may have software licenses you can access.
Stata | SPSS | SAS | Matlab | StatTransfer | Excel | |
---|---|---|---|---|---|---|
| * | * | ||||
-- | -- | |||||
-- | -- |
Run a free plagiarism check in 10 minutes, generate accurate citations for free.
Methodology
Published on June 12, 2020 by Pritha Bhandari . Revised on June 22, 2023.
Quantitative research is the process of collecting and analyzing numerical data. It can be used to find patterns and averages, make predictions, test causal relationships, and generalize results to wider populations.
Quantitative research is the opposite of qualitative research , which involves collecting and analyzing non-numerical data (e.g., text, video, or audio).
Quantitative research is widely used in the natural and social sciences: biology, chemistry, psychology, economics, sociology, marketing, etc.
Quantitative research methods, quantitative data analysis, advantages of quantitative research, disadvantages of quantitative research, other interesting articles, frequently asked questions about quantitative research.
You can use quantitative research methods for descriptive, correlational or experimental research.
Correlational and experimental research can both be used to formally test hypotheses , or predictions, using statistics. The results may be generalized to broader populations based on the sampling method used.
To collect quantitative data, you will often need to use operational definitions that translate abstract concepts (e.g., mood) into observable and quantifiable measures (e.g., self-ratings of feelings and energy levels).
Research method | How to use | Example |
---|---|---|
Control or manipulate an to measure its effect on a dependent variable. | To test whether an intervention can reduce procrastination in college students, you give equal-sized groups either a procrastination intervention or a comparable task. You compare self-ratings of procrastination behaviors between the groups after the intervention. | |
Ask questions of a group of people in-person, over-the-phone or online. | You distribute with rating scales to first-year international college students to investigate their experiences of culture shock. | |
(Systematic) observation | Identify a behavior or occurrence of interest and monitor it in its natural setting. | To study college classroom participation, you sit in on classes to observe them, counting and recording the prevalence of active and passive behaviors by students from different backgrounds. |
Secondary research | Collect data that has been gathered for other purposes e.g., national surveys or historical records. | To assess whether attitudes towards climate change have changed since the 1980s, you collect relevant questionnaire data from widely available . |
Note that quantitative research is at risk for certain research biases , including information bias , omitted variable bias , sampling bias , or selection bias . Be sure that you’re aware of potential biases as you collect and analyze your data to prevent them from impacting your work too much.
Professional editors proofread and edit your paper by focusing on:
See an example
Once data is collected, you may need to process it before it can be analyzed. For example, survey and test data may need to be transformed from words to numbers. Then, you can use statistical analysis to answer your research questions .
Descriptive statistics will give you a summary of your data and include measures of averages and variability. You can also use graphs, scatter plots and frequency tables to visualize your data and check for any trends or outliers.
Using inferential statistics , you can make predictions or generalizations based on your data. You can test your hypothesis or use your sample data to estimate the population parameter .
First, you use descriptive statistics to get a summary of the data. You find the mean (average) and the mode (most frequent rating) of procrastination of the two groups, and plot the data to see if there are any outliers.
You can also assess the reliability and validity of your data collection methods to indicate how consistently and accurately your methods actually measured what you wanted them to.
Quantitative research is often used to standardize data collection and generalize findings . Strengths of this approach include:
Repeating the study is possible because of standardized data collection protocols and tangible definitions of abstract concepts.
The study can be reproduced in other cultural settings, times or with different groups of participants. Results can be compared statistically.
Data from large samples can be processed and analyzed using reliable and consistent procedures through quantitative data analysis.
Using formalized and established hypothesis testing procedures means that you have to carefully consider and report your research variables, predictions, data collection and testing methods before coming to a conclusion.
Despite the benefits of quantitative research, it is sometimes inadequate in explaining complex research topics. Its limitations include:
Using precise and restrictive operational definitions may inadequately represent complex concepts. For example, the concept of mood may be represented with just a number in quantitative research, but explained with elaboration in qualitative research.
Predetermined variables and measurement procedures can mean that you ignore other relevant observations.
Despite standardized procedures, structural biases can still affect quantitative research. Missing data , imprecise measurements or inappropriate sampling methods are biases that can lead to the wrong conclusions.
Quantitative research often uses unnatural settings like laboratories or fails to consider historical and cultural contexts that may affect data collection and results.
If you want to know more about statistics , methodology , or research bias , make sure to check out some of our other articles with explanations and examples.
Research bias
Quantitative research deals with numbers and statistics, while qualitative research deals with words and meanings.
Quantitative methods allow you to systematically measure variables and test hypotheses . Qualitative methods allow you to explore concepts and experiences in more detail.
In mixed methods research , you use both qualitative and quantitative data collection and analysis methods to answer your research question .
Data collection is the systematic process by which observations or measurements are gathered in research. It is used in many different contexts by academics, governments, businesses, and other organizations.
Operationalization means turning abstract conceptual ideas into measurable observations.
For example, the concept of social anxiety isn’t directly observable, but it can be operationally defined in terms of self-rating scores, behavioral avoidance of crowded places, or physical anxiety symptoms in social situations.
Before collecting data , it’s important to consider how you will operationalize the variables that you want to measure.
Reliability and validity are both about how well a method measures something:
If you are doing experimental research, you also have to consider the internal and external validity of your experiment.
Hypothesis testing is a formal procedure for investigating our ideas about the world using statistics. It is used by scientists to test specific predictions, called hypotheses , by calculating how likely it is that a pattern or relationship between variables could have arisen by chance.
If you want to cite this source, you can copy and paste the citation or click the “Cite this Scribbr article” button to automatically add the citation to our free Citation Generator.
Bhandari, P. (2023, June 22). What Is Quantitative Research? | Definition, Uses & Methods. Scribbr. Retrieved August 13, 2024, from https://www.scribbr.com/methodology/quantitative-research/
Other students also liked, descriptive statistics | definitions, types, examples, inferential statistics | an easy introduction & examples, what is your plagiarism score.
Home » Quantitative Research – Methods, Types and Analysis
Table of Contents
Quantitative research is a type of research that collects and analyzes numerical data to test hypotheses and answer research questions . This research typically involves a large sample size and uses statistical analysis to make inferences about a population based on the data collected. It often involves the use of surveys, experiments, or other structured data collection methods to gather quantitative data.
Quantitative Research Methods are as follows:
Descriptive research design is used to describe the characteristics of a population or phenomenon being studied. This research method is used to answer the questions of what, where, when, and how. Descriptive research designs use a variety of methods such as observation, case studies, and surveys to collect data. The data is then analyzed using statistical tools to identify patterns and relationships.
Correlational research design is used to investigate the relationship between two or more variables. Researchers use correlational research to determine whether a relationship exists between variables and to what extent they are related. This research method involves collecting data from a sample and analyzing it using statistical tools such as correlation coefficients.
Quasi-experimental research design is used to investigate cause-and-effect relationships between variables. This research method is similar to experimental research design, but it lacks full control over the independent variable. Researchers use quasi-experimental research designs when it is not feasible or ethical to manipulate the independent variable.
Experimental research design is used to investigate cause-and-effect relationships between variables. This research method involves manipulating the independent variable and observing the effects on the dependent variable. Researchers use experimental research designs to test hypotheses and establish cause-and-effect relationships.
Survey research involves collecting data from a sample of individuals using a standardized questionnaire. This research method is used to gather information on attitudes, beliefs, and behaviors of individuals. Researchers use survey research to collect data quickly and efficiently from a large sample size. Survey research can be conducted through various methods such as online, phone, mail, or in-person interviews.
Here are some commonly used quantitative research analysis methods:
Statistical analysis is the most common quantitative research analysis method. It involves using statistical tools and techniques to analyze the numerical data collected during the research process. Statistical analysis can be used to identify patterns, trends, and relationships between variables, and to test hypotheses and theories.
Regression analysis is a statistical technique used to analyze the relationship between one dependent variable and one or more independent variables. Researchers use regression analysis to identify and quantify the impact of independent variables on the dependent variable.
Factor analysis is a statistical technique used to identify underlying factors that explain the correlations among a set of variables. Researchers use factor analysis to reduce a large number of variables to a smaller set of factors that capture the most important information.
Structural equation modeling is a statistical technique used to test complex relationships between variables. It involves specifying a model that includes both observed and unobserved variables, and then using statistical methods to test the fit of the model to the data.
Time series analysis is a statistical technique used to analyze data that is collected over time. It involves identifying patterns and trends in the data, as well as any seasonal or cyclical variations.
Multilevel modeling is a statistical technique used to analyze data that is nested within multiple levels. For example, researchers might use multilevel modeling to analyze data that is collected from individuals who are nested within groups, such as students nested within schools.
Quantitative research has many applications across a wide range of fields. Here are some common examples:
Here are some key characteristics of quantitative research:
Here are some examples of quantitative research in different fields:
Here is a general overview of how to conduct quantitative research:
Here are some situations when quantitative research can be appropriate:
The purpose of quantitative research is to systematically investigate and measure the relationships between variables or phenomena using numerical data and statistical analysis. The main objectives of quantitative research include:
Quantitative research is used in many different fields, including social sciences, business, engineering, and health sciences. It can be used to investigate a wide range of phenomena, from human behavior and attitudes to physical and biological processes. The purpose of quantitative research is to provide reliable and valid data that can be used to inform decision-making and improve understanding of the world around us.
There are several advantages of quantitative research, including:
There are several limitations of quantitative research, including:
Researcher, Academic Writer, Web developer
An official website of the United States government
The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.
The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.
Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .
Zulfiqar ali.
Department of Anaesthesiology, Division of Neuroanaesthesiology, Sheri Kashmir Institute of Medical Sciences, Soura, Srinagar, Jammu and Kashmir, India
1 Department of Anaesthesiology and Critical Care, Vijayanagar Institute of Medical Sciences, Bellary, Karnataka, India
Statistical methods involved in carrying out a study include planning, designing, collecting data, analysing, drawing meaningful interpretation and reporting of the research findings. The statistical analysis gives meaning to the meaningless numbers, thereby breathing life into a lifeless data. The results and inferences are precise only if proper statistical tests are used. This article will try to acquaint the reader with the basic research tools that are utilised while conducting various studies. The article covers a brief outline of the variables, an understanding of quantitative and qualitative variables and the measures of central tendency. An idea of the sample size estimation, power analysis and the statistical errors is given. Finally, there is a summary of parametric and non-parametric tests used for data analysis.
Statistics is a branch of science that deals with the collection, organisation, analysis of data and drawing of inferences from the samples to the whole population.[ 1 ] This requires a proper design of the study, an appropriate selection of the study sample and choice of a suitable statistical test. An adequate knowledge of statistics is necessary for proper designing of an epidemiological study or a clinical trial. Improper statistical methods may result in erroneous conclusions which may lead to unethical practice.[ 2 ]
Variable is a characteristic that varies from one individual member of population to another individual.[ 3 ] Variables such as height and weight are measured by some type of scale, convey quantitative information and are called as quantitative variables. Sex and eye colour give qualitative information and are called as qualitative variables[ 3 ] [ Figure 1 ].
Classification of variables
Quantitative or numerical data are subdivided into discrete and continuous measurements. Discrete numerical data are recorded as a whole number such as 0, 1, 2, 3,… (integer), whereas continuous data can assume any value. Observations that can be counted constitute the discrete data and observations that can be measured constitute the continuous data. Examples of discrete data are number of episodes of respiratory arrests or the number of re-intubations in an intensive care unit. Similarly, examples of continuous data are the serial serum glucose levels, partial pressure of oxygen in arterial blood and the oesophageal temperature.
A hierarchical scale of increasing precision can be used for observing and recording the data which is based on categorical, ordinal, interval and ratio scales [ Figure 1 ].
Categorical or nominal variables are unordered. The data are merely classified into categories and cannot be arranged in any particular order. If only two categories exist (as in gender male and female), it is called as a dichotomous (or binary) data. The various causes of re-intubation in an intensive care unit due to upper airway obstruction, impaired clearance of secretions, hypoxemia, hypercapnia, pulmonary oedema and neurological impairment are examples of categorical variables.
Ordinal variables have a clear ordering between the variables. However, the ordered data may not have equal intervals. Examples are the American Society of Anesthesiologists status or Richmond agitation-sedation scale.
Interval variables are similar to an ordinal variable, except that the intervals between the values of the interval variable are equally spaced. A good example of an interval scale is the Fahrenheit degree scale used to measure temperature. With the Fahrenheit scale, the difference between 70° and 75° is equal to the difference between 80° and 85°: The units of measurement are equal throughout the full range of the scale.
Ratio scales are similar to interval scales, in that equal differences between scale values have equal quantitative meaning. However, ratio scales also have a true zero point, which gives them an additional property. For example, the system of centimetres is an example of a ratio scale. There is a true zero point and the value of 0 cm means a complete absence of length. The thyromental distance of 6 cm in an adult may be twice that of a child in whom it may be 3 cm.
Descriptive statistics[ 4 ] try to describe the relationship between variables in a sample or population. Descriptive statistics provide a summary of data in the form of mean, median and mode. Inferential statistics[ 4 ] use a random sample of data taken from a population to describe and make inferences about the whole population. It is valuable when it is not possible to examine each member of an entire population. The examples if descriptive and inferential statistics are illustrated in Table 1 .
Example of descriptive and inferential statistics
The extent to which the observations cluster around a central location is described by the central tendency and the spread towards the extremes is described by the degree of dispersion.
The measures of central tendency are mean, median and mode.[ 6 ] Mean (or the arithmetic average) is the sum of all the scores divided by the number of scores. Mean may be influenced profoundly by the extreme variables. For example, the average stay of organophosphorus poisoning patients in ICU may be influenced by a single patient who stays in ICU for around 5 months because of septicaemia. The extreme values are called outliers. The formula for the mean is
where x = each observation and n = number of observations. Median[ 6 ] is defined as the middle of a distribution in a ranked data (with half of the variables in the sample above and half below the median value) while mode is the most frequently occurring variable in a distribution. Range defines the spread, or variability, of a sample.[ 7 ] It is described by the minimum and maximum values of the variables. If we rank the data and after ranking, group the observations into percentiles, we can get better information of the pattern of spread of the variables. In percentiles, we rank the observations into 100 equal parts. We can then describe 25%, 50%, 75% or any other percentile amount. The median is the 50 th percentile. The interquartile range will be the observations in the middle 50% of the observations about the median (25 th -75 th percentile). Variance[ 7 ] is a measure of how spread out is the distribution. It gives an indication of how close an individual observation clusters about the mean value. The variance of a population is defined by the following formula:
where σ 2 is the population variance, X is the population mean, X i is the i th element from the population and N is the number of elements in the population. The variance of a sample is defined by slightly different formula:
where s 2 is the sample variance, x is the sample mean, x i is the i th element from the sample and n is the number of elements in the sample. The formula for the variance of a population has the value ‘ n ’ as the denominator. The expression ‘ n −1’ is known as the degrees of freedom and is one less than the number of parameters. Each observation is free to vary, except the last one which must be a defined value. The variance is measured in squared units. To make the interpretation of the data simple and to retain the basic unit of observation, the square root of variance is used. The square root of the variance is the standard deviation (SD).[ 8 ] The SD of a population is defined by the following formula:
where σ is the population SD, X is the population mean, X i is the i th element from the population and N is the number of elements in the population. The SD of a sample is defined by slightly different formula:
where s is the sample SD, x is the sample mean, x i is the i th element from the sample and n is the number of elements in the sample. An example for calculation of variation and SD is illustrated in Table 2 .
Example of mean, variance, standard deviation
Most of the biological variables usually cluster around a central value, with symmetrical positive and negative deviations about this point.[ 1 ] The standard normal distribution curve is a symmetrical bell-shaped. In a normal distribution curve, about 68% of the scores are within 1 SD of the mean. Around 95% of the scores are within 2 SDs of the mean and 99% within 3 SDs of the mean [ Figure 2 ].
Normal distribution curve
It is a distribution with an asymmetry of the variables about its mean. In a negatively skewed distribution [ Figure 3 ], the mass of the distribution is concentrated on the right of Figure 1 . In a positively skewed distribution [ Figure 3 ], the mass of the distribution is concentrated on the left of the figure leading to a longer right tail.
Curves showing negatively skewed and positively skewed distribution
In inferential statistics, data are analysed from a sample to make inferences in the larger collection of the population. The purpose is to answer or test the hypotheses. A hypothesis (plural hypotheses) is a proposed explanation for a phenomenon. Hypothesis tests are thus procedures for making rational decisions about the reality of observed effects.
Probability is the measure of the likelihood that an event will occur. Probability is quantified as a number between 0 and 1 (where 0 indicates impossibility and 1 indicates certainty).
In inferential statistics, the term ‘null hypothesis’ ( H 0 ‘ H-naught ,’ ‘ H-null ’) denotes that there is no relationship (difference) between the population variables in question.[ 9 ]
Alternative hypothesis ( H 1 and H a ) denotes that a statement between the variables is expected to be true.[ 9 ]
The P value (or the calculated probability) is the probability of the event occurring by chance if the null hypothesis is true. The P value is a numerical between 0 and 1 and is interpreted by researchers in deciding whether to reject or retain the null hypothesis [ Table 3 ].
P values with interpretation
If P value is less than the arbitrarily chosen value (known as α or the significance level), the null hypothesis (H0) is rejected [ Table 4 ]. However, if null hypotheses (H0) is incorrectly rejected, this is known as a Type I error.[ 11 ] Further details regarding alpha error, beta error and sample size calculation and factors influencing them are dealt with in another section of this issue by Das S et al .[ 12 ]
Illustration for null hypothesis
Numerical data (quantitative variables) that are normally distributed are analysed with parametric tests.[ 13 ]
Two most basic prerequisites for parametric statistical analysis are:
However, if the distribution of the sample is skewed towards one side or the distribution is unknown due to the small sample size, non-parametric[ 14 ] statistical techniques are used. Non-parametric tests are used to analyse ordinal and categorical data.
The parametric tests assume that the data are on a quantitative (numerical) scale, with a normal distribution of the underlying population. The samples have the same variance (homogeneity of variances). The samples are randomly drawn from the population, and the observations within a group are independent of each other. The commonly used parametric tests are the Student's t -test, analysis of variance (ANOVA) and repeated measures ANOVA.
Student's t -test
Student's t -test is used to test the null hypothesis that there is no difference between the means of the two groups. It is used in three circumstances:
where X = sample mean, u = population mean and SE = standard error of mean
where X 1 − X 2 is the difference between the means of the two groups and SE denotes the standard error of the difference.
The formula for paired t -test is:
where d is the mean difference and SE denotes the standard error of this difference.
The group variances can be compared using the F -test. The F -test is the ratio of variances (var l/var 2). If F differs significantly from 1.0, then it is concluded that the group variances differ significantly.
Analysis of variance
The Student's t -test cannot be used for comparison of three or more groups. The purpose of ANOVA is to test if there is any significant difference between the means of two or more groups.
In ANOVA, we study two variances – (a) between-group variability and (b) within-group variability. The within-group variability (error variance) is the variation that cannot be accounted for in the study design. It is based on random differences present in our samples.
However, the between-group (or effect variance) is the result of our treatment. These two estimates of variances are compared using the F-test.
A simplified formula for the F statistic is:
where MS b is the mean squares between the groups and MS w is the mean squares within groups.
Repeated measures analysis of variance
As with ANOVA, repeated measures ANOVA analyses the equality of means of three or more groups. However, a repeated measure ANOVA is used when all variables of a sample are measured under different conditions or at different points in time.
As the variables are measured from a sample at different points of time, the measurement of the dependent variable is repeated. Using a standard ANOVA in this case is not appropriate because it fails to model the correlation between the repeated measures: The data violate the ANOVA assumption of independence. Hence, in the measurement of repeated dependent variables, repeated measures ANOVA should be used.
When the assumptions of normality are not met, and the sample means are not normally, distributed parametric tests can lead to erroneous results. Non-parametric tests (distribution-free test) are used in such situation as they do not require the normality assumption.[ 15 ] Non-parametric tests may fail to detect a significant difference when compared with a parametric test. That is, they usually have less power.
As is done for the parametric tests, the test statistic is compared with known values for the sampling distribution of that statistic and the null hypothesis is accepted or rejected. The types of non-parametric analysis techniques and the corresponding parametric analysis techniques are delineated in Table 5 .
Analogue of parametric and non-parametric tests
Median test for one sample: The sign test and Wilcoxon's signed rank test
The sign test and Wilcoxon's signed rank test are used for median tests of one sample. These tests examine whether one instance of sample data is greater or smaller than the median reference value.
This test examines the hypothesis about the median θ0 of a population. It tests the null hypothesis H0 = θ0. When the observed value (Xi) is greater than the reference value (θ0), it is marked as+. If the observed value is smaller than the reference value, it is marked as − sign. If the observed value is equal to the reference value (θ0), it is eliminated from the sample.
If the null hypothesis is true, there will be an equal number of + signs and − signs.
The sign test ignores the actual values of the data and only uses + or − signs. Therefore, it is useful when it is difficult to measure the values.
Wilcoxon's signed rank test
There is a major limitation of sign test as we lose the quantitative information of the given data and merely use the + or – signs. Wilcoxon's signed rank test not only examines the observed values in comparison with θ0 but also takes into consideration the relative sizes, adding more statistical power to the test. As in the sign test, if there is an observed value that is equal to the reference value θ0, this observed value is eliminated from the sample.
Wilcoxon's rank sum test ranks all data points in order, calculates the rank sum of each sample and compares the difference in the rank sums.
Mann-Whitney test
It is used to test the null hypothesis that two samples have the same median or, alternatively, whether observations in one sample tend to be larger than observations in the other.
Mann–Whitney test compares all data (xi) belonging to the X group and all data (yi) belonging to the Y group and calculates the probability of xi being greater than yi: P (xi > yi). The null hypothesis states that P (xi > yi) = P (xi < yi) =1/2 while the alternative hypothesis states that P (xi > yi) ≠1/2.
Kolmogorov-Smirnov test
The two-sample Kolmogorov-Smirnov (KS) test was designed as a generic method to test whether two random samples are drawn from the same distribution. The null hypothesis of the KS test is that both distributions are identical. The statistic of the KS test is a distance between the two empirical distributions, computed as the maximum absolute difference between their cumulative curves.
Kruskal-Wallis test
The Kruskal–Wallis test is a non-parametric test to analyse the variance.[ 14 ] It analyses if there is any difference in the median values of three or more independent samples. The data values are ranked in an increasing order, and the rank sums calculated followed by calculation of the test statistic.
Jonckheere test
In contrast to Kruskal–Wallis test, in Jonckheere test, there is an a priori ordering that gives it a more statistical power than the Kruskal–Wallis test.[ 14 ]
Friedman test
The Friedman test is a non-parametric test for testing the difference between several related samples. The Friedman test is an alternative for repeated measures ANOVAs which is used when the same parameter has been measured under different conditions on the same subjects.[ 13 ]
Chi-square test, Fischer's exact test and McNemar's test are used to analyse the categorical or nominal variables. The Chi-square test compares the frequencies and tests whether the observed data differ significantly from that of the expected data if there were no differences between groups (i.e., the null hypothesis). It is calculated by the sum of the squared difference between observed ( O ) and the expected ( E ) data (or the deviation, d ) divided by the expected data by the following formula:
A Yates correction factor is used when the sample size is small. Fischer's exact test is used to determine if there are non-random associations between two categorical variables. It does not assume random sampling, and instead of referring a calculated statistic to a sampling distribution, it calculates an exact probability. McNemar's test is used for paired nominal data. It is applied to 2 × 2 table with paired-dependent samples. It is used to determine whether the row and column frequencies are equal (that is, whether there is ‘marginal homogeneity’). The null hypothesis is that the paired proportions are equal. The Mantel-Haenszel Chi-square test is a multivariate test as it analyses multiple grouping variables. It stratifies according to the nominated confounding variables and identifies any that affects the primary outcome variable. If the outcome variable is dichotomous, then logistic regression is used.
Numerous statistical software systems are available currently. The commonly used software systems are Statistical Package for the Social Sciences (SPSS – manufactured by IBM corporation), Statistical Analysis System ((SAS – developed by SAS Institute North Carolina, United States of America), R (designed by Ross Ihaka and Robert Gentleman from R core team), Minitab (developed by Minitab Inc), Stata (developed by StataCorp) and the MS Excel (developed by Microsoft).
There are a number of web resources which are related to statistical power analyses. A few are:
It is important that a researcher knows the concepts of the basic statistical methods used for conduct of a research study. This will help to conduct an appropriately well-designed study leading to valid and reliable results. Inappropriate use of statistical techniques may lead to faulty conclusions, inducing errors and undermining the significance of the article. Bad statistics may lead to bad research, and bad research may lead to unethical practice. Hence, an adequate knowledge of statistics and the appropriate use of statistical tests are important. An appropriate knowledge about the basic statistical methods will go a long way in improving the research designs and producing quality medical research which can be utilised for formulating the evidence-based guidelines.
Conflicts of interest.
There are no conflicts of interest.
31k Accesses
2 Citations
Data analysis tools for quantitative studies are addressed in the areas of: (a) enhancements for data acquisition, (b) simple to sophisticated analysis techniques, and (c) extended exploration of relationships in data, often with visualization of results. Examples that are interwoven with data and findings from published research studies are used to illustrate the use of the tools in the service of established research goals and objectives. The authors contend that capabilities have greatly expanded in all three areas over the past 30 years, and especially during the past two decades.
This is a preview of subscription content, log in via an institution to check access.
Subscribe and save.
Tax calculation will be finalised at checkout
Purchases are for personal use only
Institutional subscriptions
Note that multilevel analysis could also be used for detailed examination of this type of research question, and for separating out effects at different levels of a multilevel design. However, other issues such as having sufficient degrees of freedom to develop robust solutions also enter with multilevel designs. One practical consideration is the lack of broad-scale researcher access to multilevel analysis software, as of 2011. Multilevel approaches such as Hierarchical Linear Modeling (HLM) (Roberts & Herrington, 2005 ) are destined to gain in popularity in the coming years.
American Psychological Association. (2001). Publication manual of the American Psychological Association (5th ed.). Washington, DC: Author.
Google Scholar
Anderson, R., & O’Connor, B. (2009). Reconstructing bellour: Automating the semiotic analysis of film. Bulletin of the American Society for Information Science and Technology. Special Issue on Visual Representation, Search and Retrieval: Ways of Seeing, 35 (5), pp. 31–40, June/July 2009. Retrieved from http://www.asis.org/Bulletin/Jun-09/JunJul09_Anderson_OConnor.html .
Becker, L. A. (1999). Calculate d and r using means and standard deviations . Online calculator. Retrieved from http://www.uccs.edu/~faculty/lbecker/ .
Bialo, E. R., & Sivin-Kachala, J. (1996). The effectiveness of technology in schools: A summary of recent research. School Library Media Quarterly, 25 (1), 51–57.
Campbell, D. T., & Stanley, J. C. (1963). Experimental and quasi-experimental designs for research . Dallas, TX: Houghton Mifflin.
Cartwright, D. (2005). MySQL 5.0 open source database . Techworld, November 11, 2005. Retrieved from http://review.techworld.com/software/346/mysql-50-open-source-database/ .
Cattell, R. B. (1950). Personality. A systematic theoretical and factual study . New York, NY: McGraw-Hill.
Book Google Scholar
*Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates.
Cumming, G. (2003, February). Confidence intervals: Understanding, picturing, and using interval estimates . Keynote at the Educational Research Exchange 2003. College of Education, University of North Texas.
DeVellis, R. (2003). Scale development: Theory and applications (2nd ed.). Thousand Oaks, CA: Sage.
*Dunn - Rankin, P. (1978). The visual characteristics of words. Scientific American , 238 (1), 122–130.
Dunn-Rankin, P. (1983). Scaling methods . Mahwah, NJ: Lawrence Erlbaum.
Dunn-Rankin, P., Knezek, G., Wallace, S., & Zhang, S. (2004). Scaling methods (2nd ed.). Mahwah, NJ: Lawrence Erlbaum.
Frawley, W., Piatetsky-Shapiro, G., & Matheus, C. (1992). Knowledge discovery in databases: An overview. AI Magazine, 13 , 213–228.
*Gibbons, J. D. (1976). Nonparametric methods for quantitative analysis . New York, NY: Holt, Rinehart and Winston.
Gibson, D. (2009). Designing a computational model of learning. In R. Ferdig (Ed.), Handbook of research on effective electronic gaming in education (Vol. 2, pp. 671–701). Hershey, PA: Information Science Reference.
Gonzalez-Sanchez, J., Christopherson, R., Chavez-Echeagaray, M., Gibson, D., Atkinson, R., & Burleson, W. (2011). How to Do Multimodal Detection of Affective States? Proceedings from the 2011 IEEE International Conference on Advanced Learning Technologies. Athens, GA.
GraphPad Software. (2005). QuickCalcs Online Calculators for Scientists: sign and binomial test calculator . Retrieved from http://www.graphpad.com/quickcalcs/binomial1.cfm .
Ihaka, R., & Gentleman, R. (1996). R: A language for data analysis and graphics. Journal of Computational and Graphical Statistics, 5 , 299–314.
*Joreskog, K. G., & Van Thillo, M. (1972). LISREL: A general computer program for estimating a linear structural equation system involving multiple indicators of unmeasured variables (RB-72-56) . Princeton, NJ: Educational Testing Service. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/17478411 .
Knezek, G., & Christensen, R. (2001). Evolution of an online data acquisition system. In J. Price, D. A. Willis, N. Davis, & J. Willis (Eds.), Proceedings of the Society for Information Technology in Teacher Education (SITE) 12th International Conference 2001, Orlando, FL (pp. 1999–2001). Norfolk, VA: Association for the Advancement of Computing in Education.
Knezek, G., & Christensen, R. (2002). Technology, pedagogy, professional development and reading achievement: KIDS project findings 2001–2002 . Denton, TX: Institute for the Integration of Technology into Teaching and Learning (IITTL).
Knezek, G., & Christensen, R. (2008a). Effect of technology-based reading programs on first and second grade achievement. Computers in the Schools, 24 (3), 23–41.
Article Google Scholar
Knezek, G., & Christensen, R. (2008b). The importance of information technology attitudes and competencies in primary and secondary education. International Handbook of Information Technology in Primary and Secondary Education (Springer International Handbooks of Education), 20 , 321–328.
Knezek, G., & Miyashita, K. (1991). Computer-related attitudes of primary school students in Japan and the U.S.A. Educational Technology Research (Japan), 14 , 17–23.
Kulik, C. C., & Kulik, J. A. (1991). Effectiveness of computer-based instruction: An updated analysis. Computers in Human Behavior, 7 , 75–94.
Microsoft Research. (2010). WorldWide Telescope . Retrieved from http://www.worldwidetelescope.org/Home.aspx .
Morales, C. (2007). Testing predictive models of technology integration in Mexico and the United States. Computers in the Schools, 24 (3/4), 153, 157.
Naik, P., Wedel, M., Bacon, L., Bodapati, A., Bradlow, E., Kamakura, W., Kreulen, J., Lenk, P., Madigan, D. M., & Montgomery, A. (2008). Challenges and opportunities in high-dimensional choice data analyses. Marketing Letters, 19 , pp. 201–213. doi 10.1007/s11002-008-9036-3 . Retrieved from https://faculty.fuqua.duke.edu/~kamakura/bio/WagnerKamakuraResearch.htm .
*Onwuegbuzie, A. J., & Teddlie, C. (2003). A framework for analyzing data in mixed methods research. In A. Tashakkori, & C. Teddlie (Eds.), Handbook of mixed methods in social and behavioral research (pp. 351–383). Thousand Oaks, CA: Sage.
Physics Forums. (2007). Matlab vs. mathematica . (November 7, 2007 posting). Retrieved from http://www.physicsforums.com/showthread.php?t=196740 .
QI Macros. (n.d). Box Plots Show Variation . Retrieved from http://www.qimacros.com/spc-charts/box plot2.php?gclid = CP2qpbOKgqgCFcZw5QodD12WrA.
Roberts, J. K., & Herrington, R. (2005). Demonstration of software programs for estimating multilevel measurement models. Journal of Applied Measurement, 6 (3), 255–272.
SAS. (2011). About SAS . Retrieved from http://www.sas.com/company/about/index.html .
Schmidt, M., & Lipson, H. (2009). Cornell Creative Machines Lab: Eureqa . Retrieved from http://ccsl.mae.cornell.edu/eureqa .
Schulz-Zander, R., Pfeifer, M., & Voss, A. (2008). Observation measures for determining attitudes and competencies toward technology. International Handbook of Information Technology in Primary and Secondary Education (Springer International Handbooks of Education), 20 (4), 367–379.
Scientific Software International, Inc. (2011). Retrieved April 4, 2013, from http://www.ssicentral.com/ .
*Shumacker, R., & Lomax, R. (2004). A Beginner’s guide to structural equation modeling . Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
SPSS. (2010). IBM SPSS statistics family . Retrieved from http://www.spss.com/ .
Tashakkori, A., & Teddlie, C. (2010). Handbook of mixed methods in social and behavioral research (2nd ed.). Thousand Oaks, CA: Sage.
The MathWorks, Inc. (2011). MATLAB . Retrieved from http://www.mathworks.com/products/matlab/description1.html .
The R Project for Statistical Computing. (2011). About R . Retrieved from http://www.r-project.org/ .
*Thompson, B. (1998). Statistical significance and effect size reporting: Portrait of a possible future. Research in the Schools, 5 (2), 33–38.
*Trochim, W. M. K. (2006). General issues in scaling. In: Research methods knowledge base . Retrieved from http://www.socialresearchmethods.net/kb/scalgen.php .
Tyler-Wood, T. L., Ellison, A., Lim, O., & Periathiruvadi, S. (2011). Bringing up girls in science (BUGS): The effectiveness of an afterschool environmental science program for increasing female student’s interest in science careers. Journal of Science Education and Technology, 21 (1), 46–55.
Tyler-Wood, T., Knezek, G., Christensen, R., Morales, C., & Dunn-Rankin, P. (2005, April 13). Scaling three versions of the Stanford-Binet Intelligence Test: Examining ceiling effects for identifying giftedness. Distinguished Paper Award for the Hawaii Educational Research Association. Presented at HERA in Hawaii in January 2005 and in the distinguished papers at the American Educational Research Association in Montreal, Canada.
Vallejo, M. A., Jordan, C. M., Diaz, M. I., Comeche, M. I., & Ortega, J. (2007). Psychological assessment via the Internet: A reliability and validity study of online (vs paper-and-pencil) versions of the General Health Questionnaire-28 (GHQ-28) and the Symptoms Check-List-90-Revised (SCL-90-R). Journal of Medical Internet Research, 9 (1), e2.
Wang, G., & Wang, Y. (2009). 3DM: Domain-oriented data-driven data mining. Fundamenta Informaticae 90 (2009), pp. 395–426, 395. doi 10.3233/FI-2009-0026 .
Wolfram Alpha. (2011). Wolfram Alpha computational knowledge engine . Retrieved from http://www.wolframalpha.com/ .
Zwilliger, D. (2003). CRC standard mathematical tables and formulae (31st ed.). Boca Raton, FL: Chapman & Hall.
Download references
Authors and affiliations.
University of North Texas, 3940 Elm St. #G150, Denton, TX, 76207-7102, USA
Gerald A. Knezek Ph.D. & Rhonda Christensen Ph.D.
You can also search for this author in PubMed Google Scholar
Correspondence to Gerald A. Knezek Ph.D. .
Editors and affiliations.
, Department of Learning Technologies, C, University of North Texas, North Elm 3940, Denton, 76207-7102, Texas, USA
J. Michael Spector
W. Sunset Blvd. 1812, St. George, 84770, Utah, USA
M. David Merrill
, Centr. Instructiepsychol.&-technologie, K.U. Leuven, Andreas Vesaliusstraat 2, Leuven, 3000, Belgium
Research Drive, Iacocca A109 111, Bethlehem, 18015, Pennsylvania, USA
M. J. Bishop
Reprints and permissions
© 2014 Springer Science+Business Media New York
Knezek, G.A., Christensen, R. (2014). Tools for Analyzing Quantitative Data. In: Spector, J., Merrill, M., Elen, J., Bishop, M. (eds) Handbook of Research on Educational Communications and Technology. Springer, New York, NY. https://doi.org/10.1007/978-1-4614-3185-5_17
DOI : https://doi.org/10.1007/978-1-4614-3185-5_17
Published : 22 May 2013
Publisher Name : Springer, New York, NY
Print ISBN : 978-1-4614-3184-8
Online ISBN : 978-1-4614-3185-5
eBook Packages : Humanities, Social Sciences and Law Education (R0)
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.
Provided by the Springer Nature SharedIt content-sharing initiative
Policies and ethics
Research approach for quantitative vs. qualitative research.
Home » Research Approach for Quantitative vs. Qualitative Research
Research methodologies are crucial in shaping our understanding of phenomena, influencing both academic and practical outcomes. Methodological distinctions between quantitative and qualitative research greatly impact how data is collected, analyzed, and interpreted. Recognizing these differences allows researchers to choose appropriate methods that align with their objectives and target populations.
Quantitative research emphasizes numerical data and statistical analysis, seeking to establish patterns and test hypotheses through measurable variables. In contrast, qualitative research focuses on understanding human experiences and social phenomena through detailed observations and interviews. By grasping the methodological distinctions, researchers can enhance the validity and reliability of their studies, ultimately contributing to deeper insights and informed decision-making.
Quantitative research is distinguished by its reliance on numerical data and statistical analysis, setting it apart from qualitative methods. Researchers often use structured tools, such as surveys or experiments, to gather quantifiable data. This data can be analyzed using various statistical methods, allowing for the identification of patterns and relationships. Such methodological distinctions are vital in forming clear conclusions based on measurable evidence, contributing to decision-making processes.
In contrast, qualitative research emphasizes understanding human experiences and perspectives through open-ended questions and unstructured approaches. While both methodologies have their strengths, it is essential to recognize the unique contributions of quantitative research. Its focus on quantifiable results helps to ensure objectivity and reliability, providing a solid foundation for further analytical endeavors. Understanding these methodological distinctions enables researchers to select the most appropriate approach for their specific research inquiries.
Data collection techniques vary significantly between qualitative and quantitative research, reflecting distinct methodological distinctions. In qualitative research, techniques such as interviews, focus groups, and observations enable researchers to gather in-depth insights. These methods allow for open-ended responses, which help in understanding participants' thoughts, behaviors, and experiences.
Conversely, quantitative research relies on structured tools like surveys and experiments, which facilitate the collection of numerical data. This approach aims to quantify variables and ultimately identify relationships, enabling hypothesis testing. By employing both qualitative and quantitative methods, researchers can create a more comprehensive understanding of their study subject. The choice of technique profoundly influences the research outcome, highlighting the importance of selecting the appropriate method based on the research goals.
Statistical analysis and interpretation play pivotal roles in discerning the methodological distinctions between quantitative and qualitative research. Quantitative research relies on statistical methods to process numerical data, enabling researchers to identify patterns and test hypotheses. In contrast, qualitative research emphasizes understanding phenomena through non-numerical data, such as interviews and observations, often requiring thematic or content analysis for interpretation.
The methodological distinctions also dictate the tools employed for analysis. For quantitative approaches, researchers often utilize software for statistical computations and visual representations of data. Qualitative analysis, however, focuses on deriving meaning and insights from textual information, often utilizing coding strategies. Each method’s interpretative framework influences not only how data is collected but also the subsequent conclusions derived, shaping the research output's validity and reliability. This understanding enhances the research's overall impact and informs best practices for conducting robust analyses across different research paradigms.
Qualitative research focuses on understanding human experiences and the meanings individuals attach to those experiences. Its methodological distinctions set it apart from quantitative approaches, emphasizing depth over breadth. Data collection methods such as interviews, focus groups, and participant observations allow researchers to gather rich narratives that illuminate complex social phenomena. This depth creates a nuanced understanding of participant perspectives, enabling the extraction of themes and patterns inherent in the data.
Moreover, qualitative research prioritizes context and rich descriptions, capturing the variability of human behavior. Unlike quantitative research, which seeks to measure and quantify, qualitative methods emphasize subjective meaning. This approach promotes exploration and discovery, allowing researchers to adapt their inquiries based on emerging findings. Through these methodological distinctions, qualitative research offers valuable insights that inform theory and practice, contributing to a holistic understanding of diverse experiences.
Thematic analysis and interpretation play a crucial role in understanding qualitative data. By identifying patterns and themes, researchers can gain deeper insights into the perspectives and experiences of participants. This process requires careful coding of data, where segments are categorized based on recurring ideas. Methodological distinctions become evident here, as qualitative analysis focuses on context and meaning, contrasting with the more structured approach of quantitative research.
In executing thematic analysis, researchers typically follow several stages. First, they familiarize themselves with the data through thorough reading. Next, they generate initial codes that capture significant features. Following coding, themes are constructed, allowing for interpretation of the results in relation to the research questions. Finally, researchers refine these themes, ensuring they accurately represent the data. Each of these steps underscores the relevance of methodological distinctions in effectively analyzing and interpreting qualitative research.
In conclusion, understanding methodological distinctions between quantitative and qualitative research is essential for effective inquiry. Each approach offers unique insights and caters to different research questions. Quantitative research excels at measuring and analyzing numerical data, establishing patterns and relationships through statistical techniques. Conversely, qualitative research delves into the rich, subjective experiences of individuals, uncovering deeper meanings and nuanced perspectives.
Choosing the right approach hinges on your objectives, context, and the nature of the questions posed. A clear understanding of each methodology's strengths enables researchers to select the most suitable framework. Ultimately, synthesizing these distinctions fosters a more comprehensive understanding of research outcomes and supports informed decision-making in diverse fields.
On this Page
You may also like, risk sensitivity analysis: ai tools for market research 2024.
Sensitivity risk analysis: how to use ai tools in 2024.
Unlock Insights from Interviews 10x faster
Check your thesis for plagiarism in 10 minutes, generate your apa citations for free.
Statistical analysis means investigating trends, patterns, and relationships using quantitative data . It is an important research tool used by scientists, governments, businesses, and other organisations.
To draw valid conclusions, statistical analysis requires careful planning from the very start of the research process . You need to specify your hypotheses and make decisions about your research design, sample size, and sampling procedure.
After collecting data from your sample, you can organise and summarise the data using descriptive statistics . Then, you can use inferential statistics to formally test hypotheses and make estimates about the population. Finally, you can interpret and generalise your findings.
This article is a practical introduction to statistical analysis for students and researchers. We’ll walk you through the steps using two research examples. The first investigates a potential cause-and-effect relationship, while the second investigates a potential correlation between variables.
Step 1: write your hypotheses and plan your research design, step 2: collect data from a sample, step 3: summarise your data with descriptive statistics, step 4: test hypotheses or make estimates with inferential statistics, step 5: interpret your results, frequently asked questions about statistics.
To collect valid data for statistical analysis, you first need to specify your hypotheses and plan out your research design.
The goal of research is often to investigate a relationship between variables within a population . You start with a prediction, and use statistical analysis to test that prediction.
A statistical hypothesis is a formal way of writing a prediction about a population. Every research prediction is rephrased into null and alternative hypotheses that can be tested using sample data.
While the null hypothesis always predicts no effect or no relationship between variables, the alternative hypothesis states your research prediction of an effect or relationship.
A research design is your overall strategy for data collection and analysis. It determines the statistical tests you can use to test your hypothesis later on.
First, decide whether your research will use a descriptive, correlational, or experimental design. Experiments directly influence variables, whereas descriptive and correlational studies only measure variables.
Your research design also concerns whether you’ll compare participants at the group level or individual level, or both.
First, you’ll take baseline test scores from participants. Then, your participants will undergo a 5-minute meditation exercise. Finally, you’ll record participants’ scores from a second math test.
In this experiment, the independent variable is the 5-minute meditation exercise, and the dependent variable is the math test score from before and after the intervention. Example: Correlational research design In a correlational study, you test whether there is a relationship between parental income and GPA in graduating college students. To collect your data, you will ask participants to fill in a survey and self-report their parents’ incomes and their own GPA.
When planning a research design, you should operationalise your variables and decide exactly how you will measure them.
For statistical analysis, it’s important to consider the level of measurement of your variables, which tells you what kind of data they contain:
Many variables can be measured at different levels of precision. For example, age data can be quantitative (8 years old) or categorical (young). If a variable is coded numerically (e.g., level of agreement from 1–5), it doesn’t automatically mean that it’s quantitative instead of categorical.
Identifying the measurement level is important for choosing appropriate statistics and hypothesis tests. For example, you can calculate a mean score with quantitative data, but not with categorical data.
In a research study, along with measures of your variables of interest, you’ll often collect data on relevant participant characteristics.
Variable | Type of data |
---|---|
Age | Quantitative (ratio) |
Gender | Categorical (nominal) |
Race or ethnicity | Categorical (nominal) |
Baseline test scores | Quantitative (interval) |
Final test scores | Quantitative (interval) |
Parental income | Quantitative (ratio) |
---|---|
GPA | Quantitative (interval) |
In most cases, it’s too difficult or expensive to collect data from every member of the population you’re interested in studying. Instead, you’ll collect data from a sample.
Statistical analysis allows you to apply your findings beyond your own sample as long as you use appropriate sampling procedures . You should aim for a sample that is representative of the population.
There are two main approaches to selecting a sample.
In theory, for highly generalisable findings, you should use a probability sampling method. Random selection reduces sampling bias and ensures that data from your sample is actually typical of the population. Parametric tests can be used to make strong statistical inferences when data are collected using probability sampling.
But in practice, it’s rarely possible to gather the ideal sample. While non-probability samples are more likely to be biased, they are much easier to recruit and collect data from. Non-parametric tests are more appropriate for non-probability samples, but they result in weaker inferences about the population.
If you want to use parametric tests for non-probability samples, you have to make the case that:
Keep in mind that external validity means that you can only generalise your conclusions to others who share the characteristics of your sample. For instance, results from Western, Educated, Industrialised, Rich and Democratic samples (e.g., college students in the US) aren’t automatically applicable to all non-WEIRD populations.
If you apply parametric tests to data from non-probability samples, be sure to elaborate on the limitations of how far your results can be generalised in your discussion section .
Based on the resources available for your research, decide on how you’ll recruit participants.
Your participants are self-selected by their schools. Although you’re using a non-probability sample, you aim for a diverse and representative sample. Example: Sampling (correlational study) Your main population of interest is male college students in the US. Using social media advertising, you recruit senior-year male college students from a smaller subpopulation: seven universities in the Boston area.
Before recruiting participants, decide on your sample size either by looking at other studies in your field or using statistics. A sample that’s too small may be unrepresentative of the sample, while a sample that’s too large will be more costly than necessary.
There are many sample size calculators online. Different formulas are used depending on whether you have subgroups or how rigorous your study should be (e.g., in clinical research). As a rule of thumb, a minimum of 30 units or more per subgroup is necessary.
To use these calculators, you have to understand and input these key components:
Once you’ve collected all of your data, you can inspect them and calculate descriptive statistics that summarise them.
There are various ways to inspect your data, including the following:
By visualising your data in tables and graphs, you can assess whether your data follow a skewed or normal distribution and whether there are any outliers or missing data.
A normal distribution means that your data are symmetrically distributed around a center where most values lie, with the values tapering off at the tail ends.
In contrast, a skewed distribution is asymmetric and has more values on one end than the other. The shape of the distribution is important to keep in mind because only some descriptive statistics should be used with skewed distributions.
Extreme outliers can also produce misleading statistics, so you may need a systematic approach to dealing with these values.
Measures of central tendency describe where most of the values in a data set lie. Three main measures of central tendency are often reported:
However, depending on the shape of the distribution and level of measurement, only one or two of these measures may be appropriate. For example, many demographic characteristics can only be described using the mode or proportions, while a variable like reaction time may not have a mode at all.
Measures of variability tell you how spread out the values in a data set are. Four main measures of variability are often reported:
Once again, the shape of the distribution and level of measurement should guide your choice of variability statistics. The interquartile range is the best measure for skewed distributions, while standard deviation and variance provide the best information for normal distributions.
Using your table, you should check whether the units of the descriptive statistics are comparable for pretest and posttest scores. For example, are the variance levels similar across the groups? Are there any extreme values? If there are, you may need to identify and remove extreme outliers in your data set or transform your data before performing a statistical test.
Pretest scores | Posttest scores | |
---|---|---|
Mean | 68.44 | 75.25 |
Standard deviation | 9.43 | 9.88 |
Variance | 88.96 | 97.96 |
Range | 36.25 | 45.12 |
30 |
From this table, we can see that the mean score increased after the meditation exercise, and the variances of the two scores are comparable. Next, we can perform a statistical test to find out if this improvement in test scores is statistically significant in the population. Example: Descriptive statistics (correlational study) After collecting data from 653 students, you tabulate descriptive statistics for annual parental income and GPA.
It’s important to check whether you have a broad range of data points. If you don’t, your data may be skewed towards some groups more than others (e.g., high academic achievers), and only limited inferences can be made about a relationship.
Parental income (USD) | GPA | |
---|---|---|
Mean | 62,100 | 3.12 |
Standard deviation | 15,000 | 0.45 |
Variance | 225,000,000 | 0.16 |
Range | 8,000–378,000 | 2.64–4.00 |
653 |
A number that describes a sample is called a statistic , while a number describing a population is called a parameter . Using inferential statistics , you can make conclusions about population parameters based on sample statistics.
Researchers often use two main methods (simultaneously) to make inferences in statistics.
You can make two types of estimates of population parameters from sample statistics:
If your aim is to infer and report population characteristics from sample data, it’s best to use both point and interval estimates in your paper.
You can consider a sample statistic a point estimate for the population parameter when you have a representative sample (e.g., in a wide public opinion poll, the proportion of a sample that supports the current government is taken as the population proportion of government supporters).
There’s always error involved in estimation, so you should also provide a confidence interval as an interval estimate to show the variability around a point estimate.
A confidence interval uses the standard error and the z score from the standard normal distribution to convey where you’d generally expect to find the population parameter most of the time.
Using data from a sample, you can test hypotheses about relationships between variables in the population. Hypothesis testing starts with the assumption that the null hypothesis is true in the population, and you use statistical tests to assess whether the null hypothesis can be rejected or not.
Statistical tests determine where your sample data would lie on an expected distribution of sample data if the null hypothesis were true. These tests give two main outputs:
Statistical tests come in three main varieties:
Your choice of statistical test depends on your research questions, research design, sampling method, and data characteristics.
Parametric tests make powerful inferences about the population based on sample data. But to use them, some assumptions must be met, and only some types of variables can be used. If your data violate these assumptions, you can perform appropriate data transformations or use alternative non-parametric tests instead.
A regression models the extent to which changes in a predictor variable results in changes in outcome variable(s).
Comparison tests usually compare the means of groups. These may be the means of different groups within a sample (e.g., a treatment and control group), the means of one sample group taken at different times (e.g., pretest and posttest scores), or a sample mean and a population mean.
The z and t tests have subtypes based on the number and types of samples and the hypotheses:
The only parametric correlation test is Pearson’s r . The correlation coefficient ( r ) tells you the strength of a linear relationship between two quantitative variables.
However, to test whether the correlation in the sample is strong enough to be important in the population, you also need to perform a significance test of the correlation coefficient, usually a t test, to obtain a p value. This test uses your sample size to calculate how much the correlation coefficient differs from zero in the population.
You use a dependent-samples, one-tailed t test to assess whether the meditation exercise significantly improved math test scores. The test gives you:
Although Pearson’s r is a test statistic, it doesn’t tell you anything about how significant the correlation is in the population. You also need to test whether this sample correlation coefficient is large enough to demonstrate a correlation in the population.
A t test can also determine how significantly a correlation coefficient differs from zero based on sample size. Since you expect a positive correlation between parental income and GPA, you use a one-sample, one-tailed t test. The t test gives you:
The final step of statistical analysis is interpreting your results.
In hypothesis testing, statistical significance is the main criterion for forming conclusions. You compare your p value to a set significance level (usually 0.05) to decide whether your results are statistically significant or non-significant.
Statistically significant results are considered unlikely to have arisen solely due to chance. There is only a very low chance of such a result occurring if the null hypothesis is true in the population.
This means that you believe the meditation intervention, rather than random factors, directly caused the increase in test scores. Example: Interpret your results (correlational study) You compare your p value of 0.001 to your significance threshold of 0.05. With a p value under this threshold, you can reject the null hypothesis. This indicates a statistically significant correlation between parental income and GPA in male college students.
Note that correlation doesn’t always mean causation, because there are often many underlying factors contributing to a complex variable like GPA. Even if one variable is related to another, this may be because of a third variable influencing both of them, or indirect links between the two variables.
A statistically significant result doesn’t necessarily mean that there are important real life applications or clinical outcomes for a finding.
In contrast, the effect size indicates the practical significance of your results. It’s important to report effect sizes along with your inferential statistics for a complete picture of your results. You should also report interval estimates of effect sizes if you’re writing an APA style paper .
With a Cohen’s d of 0.72, there’s medium to high practical significance to your finding that the meditation exercise improved test scores. Example: Effect size (correlational study) To determine the effect size of the correlation coefficient, you compare your Pearson’s r value to Cohen’s effect size criteria.
Type I and Type II errors are mistakes made in research conclusions. A Type I error means rejecting the null hypothesis when it’s actually true, while a Type II error means failing to reject the null hypothesis when it’s false.
You can aim to minimise the risk of these errors by selecting an optimal significance level and ensuring high power . However, there’s a trade-off between the two errors, so a fine balance is necessary.
Traditionally, frequentist statistics emphasises null hypothesis significance testing and always starts with the assumption of a true null hypothesis.
However, Bayesian statistics has grown in popularity as an alternative approach in the last few decades. In this approach, you use previous research to continually update your hypotheses based on your expectations and observations.
Bayes factor compares the relative strength of evidence for the null versus the alternative hypothesis rather than making a conclusion about rejecting the null hypothesis or not.
Hypothesis testing is a formal procedure for investigating our ideas about the world using statistics. It is used by scientists to test specific predictions, called hypotheses , by calculating how likely it is that a pattern or relationship between variables could have arisen by chance.
The research methods you use depend on the type of data you need to answer your research question .
Statistical analysis is the main method for analyzing quantitative research data . It uses probabilities and models to test predictions about a population from sample data.
Other students also liked, a quick guide to experimental design | 5 steps & examples, controlled experiments | methods & examples of control, between-subjects design | examples, pros & cons, more interesting articles.
Human-ai collaboration in thematic analysis using chatgpt: a user study and design recommendations.
Detection of gpt-4 generated text in higher education: combining academic judgement and software to identify generative ai tool misuse, ai and the anthropological imagination: rethinking education in the digital age.
Generative discrimination: what happens when generative ai exhibits bias, and what can be done about it, the role of generative ai in qualitative research: gpt-4's contributions to a grounded theory analysis, deepfakes and higher education: a research agenda and scoping review of synthetic media, is chatgpt transforming academics' writing style, related papers.
Showing 1 through 3 of 0 Related Papers
Grab your spot at the free arXiv Accessibility Forum
Help | Advanced Search
Title: generative ai tools in academic research: applications and implications for qualitative and quantitative research methodologies.
Abstract: This study examines the impact of Generative Artificial Intelligence (GenAI) on academic research, focusing on its application to qualitative and quantitative data analysis. As GenAI tools evolve rapidly, they offer new possibilities for enhancing research productivity and democratising complex analytical processes. However, their integration into academic practice raises significant questions regarding research integrity and security, authorship, and the changing nature of scholarly work. Through an examination of current capabilities and potential future applications, this study provides insights into how researchers may utilise GenAI tools responsibly and ethically. We present case studies that demonstrate the application of GenAI in various research methodologies, discuss the challenges of replicability and consistency in AI-assisted research, and consider the ethical implications of increased AI integration in academia. This study explores both qualitative and quantitative applications of GenAI, highlighting tools for transcription, coding, thematic analysis, visual analytics, and statistical analysis. By addressing these issues, we aim to contribute to the ongoing discourse on the role of AI in shaping the future of academic research and provide guidance for researchers exploring the rapidly evolving landscape of AI-assisted research tools and research.
Subjects: | Human-Computer Interaction (cs.HC); Artificial Intelligence (cs.AI) |
Cite as: | [cs.HC] |
(or [cs.HC] for this version) | |
Focus to learn more arXiv-issued DOI via DataCite |
Access paper:.
Code, data and media associated with this article, recommenders and search tools.
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .
Published on 14.8.2024 in Vol 26 (2024)
Authors of this article:
1 Department of Computational Medicine and Bioinformatics, University of Michigan Medical School, Ann Arbor, MI, United States
2 Department of Health Outcomes and Biomedical Informatics, University of Florida, Gainesville, FL, United States
3 Center for Research in Perinatal Outcomes, University of Florida, Gainesville, FL, United States
4 Department of Physiology and Aging, University of Florida, Gainesville, FL, United States
5 Department of Obstetrics & Gynecology, University of Florida, Gainesville, FL, United States
Lana X Garmire, PhD
Department of Computational Medicine and Bioinformatics
University of Michigan Medical School
Room 3366, Building 520, NCRC
1600 Huron Parkway
Ann Arbor, MI, 48105
United States
Phone: 1 734 615 0514
Email: [email protected]
Background: Preeclampsia is a potentially fatal complication during pregnancy, characterized by high blood pressure and the presence of excessive proteins in the urine. Due to its complexity, the prediction of preeclampsia onset is often difficult and inaccurate.
Objective: This study aimed to create quantitative models to predict the onset gestational age of preeclampsia using electronic health records.
Methods: We retrospectively collected 1178 preeclamptic pregnancy records from the University of Michigan Health System as the discovery cohort, and 881 records from the University of Florida Health System as the validation cohort. We constructed 2 Cox-proportional hazards models: 1 baseline model using maternal and pregnancy characteristics, and the other full model with additional laboratory findings, vitals, and medications. We built the models using 80% of the discovery data, tested the remaining 20% of the discovery data, and validated with the University of Florida data. We further stratified the patients into high- and low-risk groups for preeclampsia onset risk assessment.
Results: The baseline model reached Concordance indices of 0.64 and 0.61 in the 20% testing data and the validation data, respectively, while the full model increased these Concordance indices to 0.69 and 0.61, respectively. For preeclampsia diagnosed at 34 weeks, the baseline and full models had area under the curve (AUC) values of 0.65 and 0.70, and AUC values of 0.69 and 0.70 for preeclampsia diagnosed at 37 weeks, respectively. Both models contain 5 selective features, among which the number of fetuses in the pregnancy, hypertension, and parity are shared between the 2 models with similar hazard ratios and significant P values. In the full model, maximum diastolic blood pressure in early pregnancy was the predominant feature.
Conclusions: Electronic health records data provide useful information to predict the gestational age of preeclampsia onset. Stratification of the cohorts using 5-predictor Cox-proportional hazards models provides clinicians with convenient tools to assess the onset time of preeclampsia in patients.
Preeclampsia is a pregnancy-associated condition characterized by new-onset hypertension and proteinuria, typically diagnosed after 20 weeks of gestation in approximately 3%-5% of all pregnancies [ 1 ]. As one of the leading causes of maternal mortality and morbidity worldwide, it can lead to a more serious condition called eclampsia if left untreated [ 2 ]. Timely identification of preeclampsia is a key factor in pregnancy risk management and subsequent treatment. Current medical practice guideline recommends prevention therapy of low-dose aspirin on women at high risk for preeclampsia before the 13-week gestation period [ 3 ]. However, preeclampsia does not typically manifest itself clinically until after 20 weeks of gestation, through clinical markers such as blood pressure (BP), urinary protein excretion, mean arterial pressure, and placental growth factor levels. Moreover, the gestational age of preeclampsia onset can vary greatly across pregnancies [ 3 ]. Preeclampsia diagnosed before 34 weeks of gestation is called early-onset preeclampsia, and late-onset preeclampsia is diagnosed after 34 weeks [ 4 ]. To allow for maximal efficiency of prevention therapy, tools that accurately predict the onset time of preeclampsia and the patient risk will be extremely beneficial.
Previous studies have identified some qualitative risk factors of preeclampsia, including preeclampsia in a previous pregnancy, a multifetal pregnancy, chronic hypertension, kidney disease, diabetes before pregnancy, autoimmune disorders, as well as demographic factors including obesity, advanced maternal age, and race [ 5 ]. However, the quantitative importance of these risk factors relative to one another has not been adequately investigated. Haile et al [ 6 ] discuss how maternal age, weight, and history of preeclampsia significantly drive preeclampsia onset time, but many additional factors remain undefined. There is an unmet need to provide clinicians with tools to accurately identify which mothers are at risk for preeclampsia, and further identify when they will develop preeclampsia.
Prognosis modeling using population-level health data provides opportunities to systematically address both issues mentioned above [ 7 ]. These new models enable the investigation of risk factors (features) that may affect the gestational age at preeclampsia diagnosis, using the hazard ratio (HR), which indicates the importance of the risk factors. Each model outputs risk factors that influence preeclampsia development and predicts the gestational age at preeclampsia diagnosis for patients using the weighted impact of each feature. In addition, patients can be stratified into low-risk and high-risk preeclampsia groups, accompanied by differences in risk factors (features). These developed and validated prognosis models will allow clinicians to practically identify when an at-risk mother might develop preeclampsia and reveal any features associated with the onset time of preeclampsia that are not included in the current guidelines.
The discovery cohort for this project was obtained from the University of Michigan (UM) Medicine Healthcare System. All deidentified pregnancy records between the years 2015 and 2021, with at least one preeclampsia diagnosis, based on the ICD-10-CM ( International Classification of Diseases, Tenth Revision, Clinical Modification ) codes, were extracted (Table S1 in Multimedia Appendix 1 ). Patients who were diagnosed with competing conditions (Table S1 in Multimedia Appendix 1 ) were removed from the cohort. Patients who did not have any electronic medical record (EMR) in the UM system within 20 weeks of the start of their pregnancy were also removed. Since preeclampsia is clinically defined after 20 weeks, all patients with a preeclampsia diagnosis before 20 weeks of gestation were dropped from the discovery cohort. A total of 1178 pregnancies remained in the UM discovery cohort after this data selection.
Following the same inclusion and exclusion criteria, the validation cohort was generated from the University of Florida Health System and contained 881 preeclamptic pregnancies from 2015 to 2021. The Integrated Data Repository managed the deidentification and transfer of patient data to the researchers.
The electronic medical records include medical history, obstetric diagnostic codes entered during each unique pregnancy, demographics, medications, laboratory results, and vital signs (Table S2 in Multimedia Appendix 1 ). The baseline model initially used age at the start of pregnancy, race, pregnancy start date, date of the first preeclampsia diagnosis, gravidity, parity, and previous history of preeclampsia at the trimester it was diagnosed. In addition, medical histories based on ICD-10-CM diagnosis codes were extracted using the Elixhauser Comorbidities definitions [ 8 ]. Current diagnoses entered within 20 weeks of gestation were extracted using the same ICD-10-CM diagnosis codes and definitions.
The full model includes all features in the baseline model. In addition, laboratory results, vital signs, and medications ordered before 20 weeks of gestation were also added. Laboratory tests that included a complete blood count were considered (Table S2 in Multimedia Appendix 1 ). Vital signs included diastolic and systolic BP. Laboratory findings and vitals collected from the start of pregnancy (0 days gestation) to 20 weeks (140 days) gestation were included. The mean, maximum, minimum, and SD for each laboratory value were calculated. Medication records were retrieved based on previous reports that medications prescribed during pregnancy may be related to preeclampsia development [ 9 ]. Patients who did not have any laboratory finding or vital data collected and entered in the EMR system within the first 20 weeks (~15%) were assigned as “missing”. These missing values were imputed using the predictive mean matching algorithm from the R package “mice” [ 10 ], which has been shown to produce the least-biased results for data sets that use feature selection [ 11 - 13 ]. The standards for missing data used for multiple imputations were followed, and imputation was performed on only the variables with no more than 20% missingness [ 14 ]. All numeric variables were log-transformed to adjust for skewness. Each feature in the medical history, clinical diagnosis, and medication categories was computed as a binary category: 1 for presence, and 0 for absence, to reduce feature dimensionality and improve interpretability. All analysis was conducted using R (version 4.2.2; The R Foundation) [ 15 ]. Data cleaning was carried out using the packages “dplyr” [ 16 ] and “gtsummary” [ 17 ].
The UM discovery data set was randomly divided into a training set (80%) and a hold-out testing set (20%) after multiple imputations on missing variables. A Cox-proportional hazards model with Least Absolute Shrinkage and Selection Operator (LASSO) regularization was conducted through 5-fold cross-validation, using the “glmnet” [ 18 ] package in R. We used cross-validation to select the optimal LASSO hyperparameter (lambda) that gave the smallest mean squared error and then performed bootstrapping with 1000 replicates to calculate a concordance index (C-index) and 95% CIs for each data set (training, testing, and validation). The baseline model had an optimal lambda of 0.0058 (Figure S1A in Multimedia Appendix 2 ) and the full model had an optimal lambda of 0.0066 (Figure S1B in Multimedia Appendix 2 ). The baseline model had 31 features and the full model had 92 features before selection. Following regularized feature selection using the LASSO method on the training data sets, both final models have 5 selected features. The output of the Cox-PH model is the log hazard ratio, also called the prognosis index (PI), which depicts the relative risk of a patient when compared with the baseline hazard of the population. The full model was constructed in the same way as the baseline model.
External validation on each finalized model (baseline and full models) was done through collaboration with the University of Florida (UF), where the electronic health record (EHR) data and patient characteristics are different. Each feature chosen by the model was able to be identified in the UF validation cohort except for the nonsteroidal anti-inflammatory drug (NSAID) medication prescription, which was not available at the time of collection.
The performance of each model was evaluated using the C-index with bootstrapping of 1000 replicates to calculate 95% CI and P values from log-rank tests. The C-index is a metric to compare the discriminative power of a risk prediction model that describes the frequency of concordant pairs among all pairs of patients included in the model construction [ 19 ]. We used the C-index calculated from the “cindex” [ 20 ] function. Low- and high-risk pregnancies were stratified based on the median PI score of the model, and Kaplan-Meier curves were plotted for each risk group. Their differences were tested with log-rank tests using the training data set, hold-out testing data set, and the validation data set separately to evaluate the discriminative power of the model. The log-rank test is a significance test in survival analysis, with the null hypothesis that 2 groups have identical distributions of survival time. Any log-rank P value below .05 is considered statistically significant in these analyses. Feature importance was evaluated in the Cox-PH model by their HR P values. HR describes the relative contribution of a feature to the patient’s PI. In the context of our model, HRs above 1 shorten the gestational age of preeclampsia diagnosis, while HRs below 1 lengthen it.
We further measured model performance by calculating the sensitivity and specificity for each model, classified by predicting preeclampsia diagnosis by 34 and 37 weeks, respectively. We also plotted the area under the curve (AUC) from each testing data set for both models at both time points, using the “pROC” [ 21 ] package in R.
The institutional review board (IRB) of the UM Medical School (HUM#00168171) and the UF (#201601899) approved the original data collection and the use of the discovery cohort. All authors have permission for the use of this data. IRB approval was not required for the secondary analysis presented here, as it was deemed exempt. [ 22 ].
The overall study design is shown in Figure 1 . The discovery cohort was extracted from patient records in the UM Health System from 2015 to 2022 with ICD-10 ( International Statistical Classification of Diseases, Tenth Revision ) code access. All patients with a preeclampsia diagnosis after 20 weeks of gestation were included in the cohort, and other exclusion criteria are detailed in the Methods section. The finalized UM discovery cohort consists of EMRs from 1178 pregnancies. Using the same inclusion and exclusion criteria, 881 pregnancies were identified in the validation data set from UF. The patient characteristics for each cohort are listed in Table 1 . The average maternal age was 30.2 years (SD 5.67) in the discovery cohort and 29.1 years (SD 6.18) in the validation cohort. The mean gestational age of preeclampsia onset was 251 (SD 25.4) days for the discovery cohort and 257 (SD 25.9) days for the validation cohort. We constructed and validated 2 models using this data: (1) a baseline model using only patient medical history, demographics, and diagnoses of any new medical issues within the first 20 weeks of gestation; and (2) a full model including those features from the baseline model, as well as additional information on medication, laboratory findings, and vitals within the first 20 weeks of pregnancy.
Characteristics | Discovery cohort (N=1178) | Validation cohort (N=881) | |
Maternal age (years), mean (SD) | 30.2 (5.67) | 29.1 (6.18) | |
Gravidity, mean (SD) | 2.31 (1.74) | 2.82 (2.04) | |
Parity, mean (SD) | 0.68 (1.12) | 1.17 (1.5) | |
Number of fetuses, mean (SD) | 1.07 (0.26) | 1.04 (0.22) | |
Gestational age at PE onset (days), mean (SD) | 251 (25.4) | 257 (25.9) | |
Current smoker, n (%) | 61 (5) | 112 (13) | |
Current alcohol user, n (%) | 311 (26) | 184 (21) | |
African American | 195 (17) | 335 (38) | |
Asian | 74 (6) | 19 (2) | |
Hispanic | 58 (5) | 4 (1) | |
History of PE | 184 (16) | 117 (13) | |
History of PE diagnosed in the second trimester | 66 (6) | 3 (<1) | |
Uncomplicated type I diabetes | 34 (3) | 19 (2) | |
Uncomplicated type II diabetes | 62 (5) | 22 (3) | |
Uncomplicated hypertension | 201 (17) | 81 (9) | |
Kidney disease | 14 (1) | 1 (<1) | |
Depression | 265 (22) | 19 (2) | |
Mood and anxiety disorder | 318 (27) | 0 |
a PE: preeclampsia.
A baseline model was first built using medical history, demographics, and ICD-10-CM diagnosis codes of new medical conditions entered during the first 20 weeks of pregnancy. To build and test the model, we randomly split the data into an 80:20 ratio for training and testing data sets, and the Cox-PH model with LASSO (L1) regularization was built with the UM training data under 5-fold cross-validation. Alternatively, we explored ElasticNet (combined L1 and L2 regularization) as well as L2 penalization. However, the LASSO (L1) model overall performs better with higher C-indices and fewer features over these alternatives. We therefore chose LASSO as the regularization method (Table S3 in Multimedia Appendix 1 ).
We then applied this model to the 20% UM hold-out testing data and external UF validation cohort. The C-indices for the training, hold-out testing, and external validation data of the baseline model are 0.62 (95% CI 0.61-0.63), 0.64 (95% CI 0.60-0.65), and 0.61 (95% CI 0.59-0.63), confirming its validity. Table 2 shows the baseline model’s C-index and corresponding 95% CI values for each data set. To further facilitate interpretation, we classified each preeclampsia diagnosis prediction by the timeline of its occurrence, specifically by gestational weeks 34 and 37, using the UM hold-out testing data set. Such simple binary classification shows a sensitivity of 0.74, specificity of 0.50, and AUC of 0.65 for preeclampsia diagnosed at 34 weeks ( Table 2 ). It has improved performance for preeclampsia diagnosis by 37 weeks, with a sensitivity of 0.82, specificity of 0.50, and AUC of 0.69 ( Table 2 and Multimedia Appendix 3 ).
Five features were selected for the baseline model. Their respective HRs and rankings in the multivariate Cox-PH are depicted in Figure 2 A and Table 3 . By the descending order of HR, these features are the number of fetuses in pregnancy of interest (HR 25.2; P <.001), parity (HR 2.08; P <.001), history of uncomplicated hypertension (HR 2.01; P <.001), history of uncomplicated type II diabetes (HR 1.87; P <.001), and a mood or anxiety disorder (HR 1.24; P =.01). All features increase preeclampsia risk and shorten the gestational age of preeclampsia diagnosis.
Model version | 34 weeks | 37 weeks | |||||
Metrics | Sensitivity | Specificity | AUC | Sensitivity | Specificity | AUC | |
Baseline | 0.74 | 0.50 | 0.65 | 0.82 | 0.50 | 0.70 | |
Full | 0.98 | 0.51 | 0.70 | 0.86 | 0.50 | 0.70 |
Features | Hazard ratio (95% CI) | value |
Number of fetuses | 25.2 (10.7-59.4) | <.001 |
Parity | 2.08 (1.54-2.81) | <.001 |
History of uncomplicated hypertension | 2.01 (1.68-2.40) | <.001 |
History of uncomplicated type II diabetes | 1.87 (1.41-2.49) | <.001 |
Mood and anxiety disorder | 1.24 (1.07-1.43) | .01 |
To evaluate the discriminative power of this model, patients from the training data set were dichotomized into high- and low-risk groups by stratifying the samples using the median of the predicted PI (PI=1.17) from the model. The 2 risk groups showed significant differences in prognosis ( Figure 2 B and Table S4 in Multimedia Appendix 1 ). The high-risk group was characterized by higher parity and number of fetuses, while the low-risk pregnancies had no prevalence of hypertension ( P <.001) or diabetes ( P <.001). The median PI value above was applied to categorize samples into high versus low-risk groups in the hold-out (PI=1.17) and validation data (PI=2.38), similar to others [ 23 - 25 ]. As shown in Figures 2 C and 2D, the KM curves on these 2 risk groups are also significantly different ( P <.001).
We next evaluated a model with the addition of laboratory findings, vitals, and medications prescribed in the first 20 weeks of gestation to the clinical data used in the baseline model. We constructed the new Cox-PH model, or the “full model,” in the same manner as the baseline model and obtained a 5-feature Cox-PH model ( Figure 3 A). Similar to the baseline model, LASSO regularization shows better overall performance than ElasticNet and L2 regularization and is chosen as the default (Table S3 in Multimedia Appendix 1 ). This new model reaches the C-indices of 0.66 (95% CI 0.64-0.67) and 0.69 (95% CI 0.64-0.70) for the training and hold-out testing data sets, respectively. It also yields a C-index of 0.61 (95% CI 0.60-0.63) on the UF validation cohort, despite missing 1 feature (NSAID medication) in the UF cohort. Table 2 lists the full-model C-indices and 95% CIs for each data set. Similar to the baseline model, to help interpretation, we classified each preeclampsia diagnosis prediction using the timeline of preeclampsia occurrence by gestational weeks 34 and 37, respectively, using the UM hold-out testing data set. It yields a sensitivity of 0.98, specificity of 0.51, and AUC of 0.70 for correctly predicting preeclampsia by week 34 ( Table 2 ). The model has an improved correct diagnosis by week 37, with a sensitivity of 0.86, specificity of 0.50, and AUC of 0.70 ( Table 2 and Multimedia Appendix 3 ).
The full model also yields 5 features, all with positive HRs ( Figure 3 A and Table 4 ). In descending order of HR, these features are maximum diastolic blood pressure (HR 21.7; P <.001), number of fetuses in current pregnancy (HR 21.1; P <.001), parity (HR 1.81; P <.001), history of uncomplicated hypertension (HR 1.79; P <.001), and NSAID medication prescription (HR 1.35; P <.001). Three of these features, namely the number of fetuses, history of uncomplicated hypertension, and parity features were also selected by the baseline model ( Figure 3 B). Table S5 in Multimedia Appendix 1 shows each of the features and their HRs in a univariate analysis. Their HRs across the baseline and full models remain very similar and had P values less than .05, suggesting that they are all significant in predicting preeclampsia onset time regardless of the other additional input information. Maximum diastolic BP and NSAID medication prescription are newly selected features unique to the full model ( Figures 3 A and 3B).
Like the baseline model, we stratified patients into high- versus low-risk groups using the median predicted PI value of 5.15 from the training data set ( Figure 3 C). The high-risk group was characterized by higher parity, a higher number of fetuses, and higher maximum diastolic BP (Table S4 in Multimedia Appendix 1 ). In contrast, the low-risk group had no history of hypertension and rare use of NSAID medication. BP had the most statistically significant difference ( P <.001), as expected. The same median threshold was applied to the 20% hold-out testing data set (PI=5.08) and validation data (PI=5.18) for dichotomization ( Figures 3 D and 3E). KM curves on these 2 risk groups in the testing set have even more significant differences in their gestational age at diagnosis ( P <.001). Both models are to be used by entering patient information in the predictors to predict when the patient may develop preeclampsia.
To determine the potential impact of missing data on modeling results, we explored building a baseline and full model with only cases that had complete BP data—the main selected feature in the full model. Table S6 in Multimedia Appendix 1 shows the selected features of both of these models. The complete cases baseline model had a training C-index of 0.63 and a testing C-index of 0.64. The complete cases full model had a training C-index of 0.67 and a testing C-index of 0.65. Due to similar performance and selected features, it can be safely assumed that imputation had little impact on the finalized models.
Features | Hazard ratio (95% CI) | value |
Maximum diastolic blood pressure | 21.7 (7.93-59.8) | <.001 |
Number of fetuses | 21.1 (9.88-45.1) | <.001 |
Parity | 1.81 (1.37-2.39) | <.001 |
History of uncomplicated hypertension | 1.79 (1.53-2.11) | <.001 |
NSAID medication | 1.35 (1.15-1.58) | <.001 |
a NSAID: nonsteroidal anti-inflammatory drug.
This paper is the first of its kind to implement and externally validate a prognosis-predicting model for preeclampsia onset time using EHR data from the first 20 weeks of pregnancy [ 26 ]. These models confirmed that factors such as BP in the first 20 weeks of pregnancy, the number of fetuses, parity, and previous history of hypertension are associated with earlier preeclampsia onset time. Moreover, comorbidities such as gestational diabetes and anxiety, as well as NSAID medication, shorten preeclampsia onset time. The similar performance across validation and development data sets provides confidence in the accuracy of the predictive outputs.
A recent study stratified patients with preeclampsia by gestational age to build classification models, resulting in many models that are difficult for clinicians to select from [ 27 ]. Moreover, these classification models cannot predict the gestational age of onset for an individual patient, thus failing to assist clinicians in making early decisions on delivery plans and proper antenatal care. Unlike most other accurate preeclampsia onset time prediction models, our models only use EMR data from the first 20 weeks of pregnancy and do not require advanced testing inputs, such as biomarkers [ 27 ], enabling earlier use in clinics. In a systematic review of 68 preeclampsia prediction models [ 27 ], only 6% (4/68) of them were externally validated, and those not requiring complex biomarker features had much lower AUCs (0.58-0.61) than the models presented here (AUC 0.65-0.70), highlighting the accuracy of our models once validated against a different patient population.
Due to the difficulty in predicting preeclampsia, accurate models that can identify women at high risk for preeclampsia can provide early targeted treatment as well as increased surveillance to reduce adverse outcomes [ 28 ]. The models here not only confirm the importance of some previously known risk factors, such as the number of fetuses, history of hypertension, and parity but also assign quantitative scores (weights) on the importance of these risk factors relative to each other. This is a significant advancement from most of the other studies focusing on a single risk factor. It also provides clinicians as well as pregnant women with quantitative tools to assess the onset time of preeclampsia more accurately, beyond the qualitative assessment of risks. Risk factors with higher weights can take a higher priority for clinicians to identify potential patients with preeclampsia. The fact that maximum diastolic BP had the highest HR in the full model confirms the importance of monitoring BP as early as possible, even before preeclampsia is diagnosed clinically [ 29 ]. More importantly, it identifies additional alarming factors to be considered in predicting preeclampsia diagnosis at gestational age, such as mood and anxiety disorder.
Further risk stratification of the survival models had slightly low specificity values in predicting the dichotomous diagnosis of preeclampsia at 34 and 37 weeks, suggesting that the continuous risk diagnosis has overall better performance compared with the simple binary prediction. However, the stratification may offer an easier way to identify women who may benefit more significantly from prevention therapy and need more medical attention from doctors for the possibility of preeclampsia. EHR-based models can serve as a screening test. For the patients that are potentially false positive for preeclampsia due to the lower specificity of the model, additional confirmative diagnostic tests using very specific biomarkers should be done, as practiced clinically.
Earlier studies using all pregnant women also revealed that mood and anxiety disorders increase the risk of preeclampsia [ 30 ]. We further show that within patients with preeclampsia, mood and anxiety disorders shorten the onset time of preeclampsia. This provides more context for clinicians to identify pregnant patients who present mood and anxiety disorders and provide preventative care to reduce preeclampsia onset risk. The molecular mechanism linking mood and anxiety disorders with preeclampsia is worth further research. We also show that NSAID use is positively associated with earlier onset of preeclampsia. However, aspirin is a common NSAID used by pregnant women at risk for preeclampsia early in pregnancy [ 31 ]. It was suggested that NSAID use may serve as a proxy for the interaction of many unmeasured risk factors [ 32 ]. Thus, the positive association of NSAID to the earlier onset of preeclampsia may indicate that it is a marker of high-risk preeclampsia in the population, rather than the cause of it.
A particular strength of the models here is their simplicity despite being quantitative. The models can also be generalized to different medical centers and hospitals, given the good accuracy when validated by vastly different institutions with different protocols, data collection, and data storage. There is a growing need for evidence-based and effective tools for clinicians to screen women at high risk of preeclampsia early in pregnancy, in the first and early second trimesters. This model supplies this need for early prediction models that previous models have not been able to fulfill [ 33 ]. Most clinical models recently published include many predictors from biomarkers and ultrasound markers that need special procedures [ 34 ], further suggesting that a simpler model on routinely collected clinical data is valuable to be implemented in a clinical setting. The main strength of this modeling for clinical use proposed here is providing more context in screening patients at risk for preeclampsia.
Our ultimate goal is to implement these models in the health care system, for example, starting from the University of Michigan. Potential challenges for implementing these models in a clinical setting include institutional buy-in, installation of the software in a HIPAA (Health Insurance Portability and Accountability Act)-compliant computing environment, and explaining the meaning of risk factors and model results to patients informatively without overly stressing them. In addition, these models may potentially require more active updating for improving accuracy, by considering additional multicenter data. Also, the current Cox-PH model is not designed to include longitudinal observations, limiting the kind of input variables to be incorporated into the model. Future work may benefit from more sophisticated modeling approaches [ 35 ]. Besides EHR, other omics information such as genetics, genomics, proteomics, and metabolomics using maternal blood samples [ 34 ] may be used, if they are available, to improve the model performance. However, implementing multimodal and complex models like this in the clinical setting is additionally challenging and would require more advanced modeling that can calculate individual risk scores for clinical application. It is also important to note the use of EHR data to extract medication prescriptions does not accurately capture the actual use or adherence of the medication by patients, and future research could be strengthened by combining data sources that provide such information.
In conclusion, this study reports prognosis models to predict the onset gestational age of preeclampsia with EMR data before the first 20 weeks of pregnancy. They identify clinical and physiological factors that clinicians should monitor as indicators of early preeclampsia development.
The authors would like to thank UM Precision Health for providing technical support for data extraction in this study, the UF Integrated Data Repository, and the UF Health Office of the Chief Data Office for providing the analytic data set for this project. DJL was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (K01DK115632) and the UF Clinical and Translational Science Institute (UL1TR001427). LXG was supported by grants (K01ES025434) awarded through funds provided by the trans-National Institutes of Health Big Data to Knowledge initiative (R01 LM012373 and LM012907 awarded by the National Library of Medicine, and R01 HD084633 awarded by National Institute of Child Health and Human Development). ADM is supported by the National Center for Advancing Translational Science (5TL1TR001428). No funding sources listed were involved in the study design, collection, analysis, and interpretation of data, writing of the report, or decision to submit for publication.
The data sets generated during and/or analyzed during this study are not publicly available due to the presence of patient-protected health information. Data are available upon reasonable request and must be submitted on an individual basis to the home institution. Table S2 in Multimedia Appendix 1 lists all the EHR features extracted from the UM system that were considered in the starting model.
LXG conceived this project and supervised the study. HKB conducted data analysis and wrote the manuscript. XY collaborated on data extraction of the University of Michigan cohort. ADM and DJL collaborated on validation using the University of Florida cohort. ADM provided clinical assessments and assistance. All authors have read, revised, and approved the manuscript.
None declared.
Supplementary tables.
Lambdas from Least Absolute Shrinkage and Selection Operator (LASSO) regularization from the baseline and full preeclampsia (PE) prediction models. (A) Scatterplot of tested lambda values and associated errors from baseline model LASSO regularization. (B) Scatterplot of tested lambda values and associated errors from full model LASSO regularization.
AUC values of preeclampsia (PE) diagnosed at 34 and 37 weeks for the baseline and full PE prediction models. (A) Plot of the sensitivity, specificity, and area under the curve (AUC) values for the baseline model of predicting PE diagnosed at 34 weeks (red, AUC=0.654) and 37 weeks (green, AUC=0.694) for the testing dataset. (B) Plot of the sensitivity, specificity, and AUC values for the full model of predicting PE diagnosed at 34 weeks (red, AUC=0.697) and 37 weeks (green, AUC=0.700) for the testing data set.
area under the curve |
blood pressure |
concordance index |
electronic health record |
electronic medical record |
Health Insurance Portability and Accountability Act |
hazard ratio |
International Statistical Classification of Diseases, Tenth Revision |
International Classification of Diseases, Tenth Revision, Clinical Modification |
institutional review board |
Least Absolute Shrinkage and Selection Operator |
nonsteroidal anti-inflammatory drug |
prognosis index |
University of Florida |
University of Michigan |
Edited by A Mavragani; submitted 15.05.23; peer-reviewed by S Nagavally, D Heider, B Puladi; comments to author 22.11.23; revised version received 17.01.24; accepted 30.05.24; published 14.08.24.
©Hailey K Ballard, Xiaotong Yang, Aditya D Mahadevan, Dominick J Lemas, Lana X Garmire. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 14.08.2024.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research (ISSN 1438-8871), is properly cited. The complete bibliographic information, a link to the original publication on https://www.jmir.org/, as well as this copyright and license information must be included.
Objectives Mycophenolate mofetil (MMF) and azathioprine (AZA) are immunomodulatory treatments in interstitial lung disease (ILD). This systematic review aimed to evaluate the efficacy of MMF or AZA on pulmonary function in ILD.
Design Population included any ILD diagnosis, intervention included MMF or AZA treatment, outcome was delta change from baseline in per cent predicted forced vital capacity (%FVC) and gas transfer (diffusion lung capacity of carbon monoxide, %DLco). The primary endpoint compared outcomes relative to placebo comparator, the secondary endpoint assessed outcomes in treated groups only.
Eligibility criteria Randomised controlled trials (RCTs) and prospective observational studies were included. No language restrictions were applied. Retrospective studies and studies with high-dose concomitant steroids were excluded.
Data synthesis The systematic search was performed on 9 May. Meta-analyses according to drug and outcome were specified with random effects, I 2 evaluated heterogeneity and Grading of Recommendations, Assessment, Development and Evaluation evaluated certainty of evidence. Primary endpoint analysis was restricted to RCT design, secondary endpoint included subgroup analysis according to prospective observational or RCT design.
Results A total of 2831 publications were screened, 12 were suitable for quantitative synthesis. Three MMF RCTs were included with no significant effect on the primary endpoints (%FVC 2.94, 95% CI −4.00 to 9.88, I 2 =79.3%; %DLco −2.03, 95% CI −4.38 to 0.32, I 2 =0.0%). An overall 2.03% change from baseline in %FVC (95% CI 0.65 to 3.42, I 2 =0.0%) was observed in MMF, and RCT subgroup summary estimated a 4.42% change from baseline in %DL CO (95% CI 2.05 to 6.79, I 2 =0.0%). AZA studies were limited. All estimates were considered very low certainty evidence.
Conclusions There were limited RCTs of MMF or AZA and their benefit in ILD was of very low certainty. MMF may support preservation of pulmonary function, yet confidence in the effect was weak. To support high certainty evidence, RCTs should be designed to directly assess MMF efficacy in ILD.
PROSPERO registration number CRD42023423223.
Data are available in a public, open access repository. We cited published study.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/ .
https://doi.org/10.1136/bmjresp-2023-002163
Request permissions.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Mycophenolate mofetil (MMF) and azathioprine (AZA) are two immunomodulatory drugs used in the treatment of connective tissue disease with both drugs having mechanisms that target lymphocytes. While increasingly used in treatment of interstitial lung disease (ILD), there is limited evidence for the efficacy of MMF or AZA in improving outcomes.
We undertook a systematic review and meta-analysis to assess whether administration MMF or AZA in ILD was associated with changes in pulmonary function and gas transfer. There was an unclear benefit of MMF on ILD. There was no significant difference in outcome when compared with placebo or standard of care. A minor increase in per cent predicted forced vital capacity and diffusion lung capacity of carbon monoxide from baseline was observed in MMF. Studies on AZA were limited.
Findings may provide indication of an attenuation on lung function decline, however, all estimates should be considered weak evidence with a high likelihood that additional trials may change effect estimates in a manner sufficient to influence decision-making. The limited number of controlled studies in MMF and AZA highlight an important need for additional well-designed randomised controlled trials to directly test their efficacy in ILD.
Interstitial lung disease (ILD) is a diverse group of conditions that affect the interstitial structure of the lungs. These diseases can be characterised by progressive lung damage, resulting in symptoms such as dyspnoea, decreased exercise tolerance and a diminished quality of life. 1 Forced vital capacity (FVC) and the diffusion lung capacity of carbon monoxide (DL CO ) are widely used to assess the severity of disease and predict prognosis of people with ILD. 2
Mycophenolate mofetil (MMF) and azathioprine (AZA) are two immunomodulatory drugs commonly used in the treatment of connective tissue disease (CTD) and associated ILD (CTD-ILD). MMF works by blocking the de novo synthesis of DNA, thereby inhibiting the proliferation of lymphocytes. AZA is a purine analogue that hinders purine synthesis and becomes incorporated into DNA during the anabolic process. Similar to MMF, this mechanism of action makes both drugs more specific for targeting lymphocytes, as lymphocytes do not have a salvage pathway in DNA synthesis. 3
There is limited evidence for the safety or efficacy of MMF or AZA in improving outcomes for people with ILD. 4 This systematic review and meta-analysis aims to assess whether the administration of MMF or AZA in ILD is associated with changes in pulmonary function and gas transfer, and to synthesise evidence of safety profiles.
The prespecified protocol was submitted to PROSPERO on 3 May 2023 and registered on 16 May 2023 (CRD42023423223). The search strategy was last performed on 9 May 2023.
The population was defined as people with ILD (Idiopathic pulmonary fibrosis (IPF), chronic hypersensitivity pneumonia and all CTD-ILD, including systemic scleroderma) the intervention was MMF or AZA; the comparator was placebo or standard of care; the primary outcomes were per cent predicted FVC (%FVC) and DL CO (%DL CO ). Adverse events, respiratory symptoms, quality of life and mortality were investigated as secondary outcomes. Relevant studies were searched in Medline and Embase using comprehensive search terms ( online supplemental documents 1 and 2 ). Relevant ongoing trials were searched on clinicaltrials.gov ( online supplemental document 3 ).
Inclusion criteria.
Eligible studies included interventional randomised controlled trials (RCTs) and observational prospective studies of adults (>18 years old) diagnosed with any ILD, where at least one arm was treated with MMF or AZA. Low doses of steroids concomitant with or prior to MMF or AZA treatment were allowed, while we excluded studies with concomitant high-dose therapies (≥20 mg/day of prednisone or equivalent). Finally, we excluded studies that did not report %FVC or %DL CO . No language restrictions were applied.
Two authors (FL and LF) independently assessed the titles and abstracts of the identified studies according to the eligibility criteria. Subsequently, two authors (FL and LF) evaluated the full text of the selected articles to determine their inclusion. Any disagreements were resolved through discussion and consensus with a third author (IS) resolving any remaining disagreements.
Data were independently extracted using a proforma and confirmed by two authors (FL and LF). Extracted data included study design, authors, year of publication; patient data namely age, reported sex or gender, duration of disease at the time of evaluation, aetiology of the disease and intervention characteristics, including MMF or AZA treatment, dose and duration of treatments. Primary outcomes of interest, %FVC and %DL CO , were extracted, along with any secondary outcomes reported, at baseline and follow-up time point closest to 12 months.
Continuous primary outcomes were collected as mean and SD at baseline and follow-up time points. When studies reported other summary values, these were converted to mean and SD. 5 Secondary outcomes reported as dichotomous and categorical variables were extracted as ratio and/or per cent.
Two authors (FL and LF) independently used the Cochrane ‘Risk of Bias’ assessment tool 2.0 to evaluate the included RCTs prior to quantitative synthesis. 6 Risk of bias in the observational prospective studies was assessed using the Newcastle-Ottawa Quality Assessment Scale. 7 To assess the risk of bias in single-arm observational cohorts, specifically for evaluation of ‘selection bias’ and ‘comparative bias’ on the Newcastle-Ottawa Quality Assessment Scale, baseline time points were considered as the ‘not exposed cohort’ and the follow-up time point as the ‘exposed cohort’. Studies that were determined to have a high risk of bias were excluded from quantitative synthesis.
When two or more studies were available for a specific treatment, a random effects meta-analysis with inverse-variance was performed to evaluate the effect of the treatment on %FVC and %DL CO values. Estimates were expressed as weighted mean difference (WMD) with 95% CI.
Where there were sufficient RCT data, the primary endpoint analysis assessed the delta difference in %FVC and %DL CO at follow-up from baseline in respiratory function for MMF or AZA relative to the comparator. In a secondary endpoint analysis, the difference in %FVC and %DL CO between follow-up and baseline in people receiving of MMF or AZA was compared. Analyses were performed according to drug, prespecified subgroup analyses were performed according to study design (RCT or prospective observation study) and follow-up time (6 months or 12 months and over).
Heterogeneity was evaluated using I 2 statistic to interpret the proportion of the total variability that was due to between-study heterogeneity, as well as inspection of forest plots. All analyses were performed by using Stata SE V.17.0.
The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach was used to assess the certainty of evidence in effect estimates from RCT data exclusively. The level of certainty was evaluated as high, moderate, low or very low, considering factors of risk of bias, inconsistency, indirectness, imprecision and publication bias. 8 Publication bias was inspected with asymmetry in funnel plots and Egger’s test.
Representatives from the Action for Pulmonary Fibrosis charity were involved in the design and dissemination of this systematic review. Members of the REMAP-ILD Consortium include charity representatives.
A total of 2831 publications from Embase and Medline were identified. After removal of duplicates and evaluating the titles and abstracts, 23 studies were assessed for eligibility. Among these, 11 studies were excluded due to retrospective design (n=2), incompleteness (n=2), lack of the outcome of interest (n=2) or the presence of concomitant treatment with high doses of steroids (n=5) ( figure 1 , online supplemental table 1 ). A total of 13 studies were eligible for qualitative synthesis ( table 1 ). 9–21 Separately, four ongoing MMF studies were identified, including one phase II RCT, two open-label trials and one prospective cohort study; two studies address pulmonary involvement of systemic sclerosis, one study recruits participants with fibrotic hypersensitivity pneumonitis and one study focuses on idiopathic inflammatory myopathy ILD ( online supplemental document 3 ).
Preferred reporting items for systematic review and meta-analysis (PRISMA) flow of study search and inclusion. AZA, azathioprine; MMF, mycophenolate mofetil.
Reported study characteristics of included cohorts
A moderate risk of bias was observed for the blinding of outcome assessment in all the included RCTs, 12 14 15 19–21 as there were no mentioned strategies to blind the pulmonary function test evaluations ( figure 2A ). Roig et al 21 and Zhang et al 20 were considered at high risk of bias in terms of blinding of participants and personnel, as they compared intravenous and oral (per os) treatments without implementing a double dummy strategy. Due to the high risks of bias across a number of domains and insufficient data reporting, the study by Roig et al 21 was excluded from quantitative synthesis. In the assessment of prospective observational studies, six studies 10 11 13 16–18 had selection bias in the ascertainment of exposure, but all studies were considered adequate ( figure 2B , online supplemental table 2 ).
Qualitative synthesis: risk of bias. (A) Risk of bias in RCTs assessed using Cochrane ROB2.0 tool. (B) Risk of bias assessed using Newcastle-Ottawa Quality assessment scale for cohort studies. Green has been assessed as: three or four stars in selection bias; two stars in comparability, three stars in outcome. Yellow has been assessed as: two stars in selection bias; one star in comparability, two stars in outcome. RCTs, randomised controlled trial; ROB2.0, Risk of Bias 2.0.
MMF or AZA were tested in a total of four trials, with three trials using MMF 15 19 20 and one trial using AZA. 14 Only MMF trials were included in primary analysis with a total of 249 participants, of which 119 were in the intervention arm and 130 were in the comparator arm ( figure 3A ). In primary analysis, the overall delta change in %FVC values from baseline to follow-up was not significantly different between the intervention and comparator arms (WMD 2.94, 95% CI −4.00 to 9.88, I 2 =79.3%). Significant heterogeneity was observed and the estimate was interpreted to have very low certainty ( table 2 , online supplemental figure 1A ).
Primary endpoint analysis of efficacy on pulmonary function relative to comparator. (A) Forest plot of difference in %FVC in treatment of MMF versus comparators at follow-up. (B) Forest plot of difference in %DLco in treatment of MMF versus comparators at follow-up. Positive values indicate improvement relative to comparator, negative values indicate decline relative to comparator. Presented with cohort size (N) for intervention and comparator, weighted mean difference (WMD) and 95% CI. Follow-up time reported in months. %DLco, per cent predicted diffusion lung capacity of carbon monoxide; %FVC, per cent predicted forced vital capacity; MMF, mycophenolate mofetil.
GRADE approach to rate certainty of effect estimates
The overall delta change in %DL CO from baseline to follow-up was not significantly different in the interventional arm compared with the comparator arm (WMD %DLco −2.03, 95% CI −4.38 to 0.32, I 2 =0.0% ( figure 3B ). Heterogeneity was not observed and the estimate was interpreted to have very low certainty ( table 2 , online supplemental figure 2B ).
A total of 6 prospective observational studies 9–11 16–18 and 5 RCTs 12 14 15 19 20 were included in secondary analysis of the difference between follow-up and baseline in %FVC, including a combined sample of 267 evaluated at baseline and 244 at follow-up, representing 7.5% loss to follow up. In prespecified subgroup analysis by drug ( online supplemental figure 3A ), treatment with AZA suggested a decline in %FVC with treatment, although this was not statistically significant (two studies; WMD −6.14, 95% CI −12.88 to 0.61, I 2 =48.3%). Treatment with MMF was observed to have a small and significant increase in %FVC value at follow-up (nine studies; WMD 2.03, 95% CI 0.65 to 3.42, I 2 =0.0%). Additional subgroup analyses performed on MMF treatment observed similar effect sizes according to study design and very low certainty of evidence ( figure 4A , table 2 ), while a greater effect of MMF was observed at follow-up of 12 months or over with no significant heterogeneity between time points ( online supplemental figure 4A ).
Secondary endpoint analysis of efficacy on pulmonary function compared with baseline. Subgroup analysis of MMF overall and summary estimates presented by study design of trial or prospective observational study. 4 (A) Forest plot of change in %FVC at follow-up versus baseline. (B) Forest plot of change in %DLco versus baseline. Positive values indicate improvement relative to baseline, negative values indicate decline relative to baseline. Presented with cohort size (N) for intervention and comparator, weighted mean difference (WMD) and 95% CIs. Follow-up time reported in months. %DLco, per cent predicted diffusion lung capacity of carbon monoxide; %FVC, per cent predicted forced vital capacity; MMF, mycophenolate mofetil.
Data from a total of 7 observational studies 9–11 13 16–18 and 5 RCTs 12 14 15 19 20 were available for analysis of %DL CO , including 262 and 234 patients, respectively, at baseline and follow-up representing a 10.7% loss to follow up. In subgroup analysis by drug ( online supplemental figure 3B ), treatment with AZA suggested a decline (two studies; −5.72, 95% CI −13.79 to 2.34, I 2 =49.8%), while treatment with MMF suggested an increase (10 studies; 1.62, 95% CI −1.70 to 4.94, I 2 =60.5%), although effect estimates did not reach significance and substantial heterogeneity was observed. Additional subgroup analyses performed on MMF treatment observed a significant decline in %DL CO in prospective observation studies (WMD −1.36, 95% CI −2.37 to −0.36, I 2 =0.0%) and a significant improvement in RCTs (WMD 4.42, 95% CI 2.05 to 6.79; I 2 =0.0%), with substantial heterogeneity between subgroups and very low certainty in evidence ( figure 4B , table 2 ). Subgroup analysis on follow-up time did not observe a significant effect in %DL CO with no significant heterogeneity observed between groups ( figure 4B ).
All the studies reported adverse events. The most frequent adverse events in the treated arms were diarrhoea and pneumonia, followed by lympho/leucopenia, anaemia and skin infection ( online supplemental table 3 ).
Four studies reported on respiratory symptoms. 11 12 15 18 In the study by Mankikian et al , no significant difference was observed in the change from baseline in dyspnoea and cough between the treated patients and the placebo group. Naidu et al reported an improvement in respiratory symptoms in both arms of the study, with no significant difference between the treatment and control groups. Liossis et al reported an improvement in respiratory symptoms compared with baseline after administration of MMF. Vaiarello et al evaluated symptoms during a cardiopulmonary exercise test before and after MMF treatment, observing no significant difference in dyspnoea measured by the Borg scale.
Two studies reported change in quality of life. 12 15 Mankikian et al and Naidu et al evaluated the change of quality of life between the interventional and the control arm using respectively the SF-36 V.1.3 questionnaire and the Medical Outcome Survey SF-36 V.2. Both these studies reported no difference in the QoL in MMF arm compared with control. None of the included studies reported on mortality.
This systematic review and meta-analysis suggested an unclear benefit of MMF or AZA on FVC or DL CO in people with ILD. Secondary endpoint analysis of change over time stratified by treatment suggested a minor increase in %FVC or %DL CO compared with baseline in MMF treated groups. The review highlighted a limited number of trials and prospective observational studies that directly tested the effect of MMF or AZA on lung function in the current literature, particularly precluding interpretations on the efficacy of AZA.
All estimates based on MMF RCT data were of very low GRADE certainty of evidence. Risk of bias was deemed moderate as one trial included unblinded participants, one study was post hoc analysis of trial data, and all trials had potential issues in blinding of outcome assessment. Heterogeneity and differences in the direction of effect across RCTs contributed to inconsistency. Imprecision was considered high due to limited RCTs, small samples and small effect sizes with wide CIs. Indirectness was deemed moderate as studies included different diagnoses. There was no strong evidence of publication bias. While these findings provide some indication of the effect, all estimates should be considered weak evidence with a high likelihood that additional studies may change effect estimates in a manner sufficient to influence decision-making.
Primary endpoint analysis in MMF observed no significant effect of treatment vs comparator groups for %FVC or %DL CO , although a non-significant effect in %DL CO favoured comparator. In contrast, secondary endpoint analysis suggested that MMF treatments could improve on baseline pulmonary function, although this may be insufficient relative to placebo. In further subanalyses restricted to MMF, greater improvement in %FVC was observed at longer follow-up, with no difference according to study design. Conversely, greater improvement in %DL CO was observed in trial designs, with no difference according to follow-up timing. While heterogeneity was minimised in subgroup analyses, effect sizes were small.
In the narrative review of adverse events, we found that both treatments were well tolerated, however, studies on real-world data suggest difficulties in tolerability. 4 The most frequent adverse events observed with MMF and AZA treatment included respiratory infections and haematological disorders. It is noteworthy that these adverse events were often mild and did not typically require specific treatment nor differ to events encountered in standard treatments. MMF or AZA interruption due to adverse events led to treatment discontinuation only in a few cases. Symptoms appeared to slightly improve after treatment commenced, but stricter interventional vs placebo studies are needed to assess the real effect on patient-reported outcomes.
The first meta-analysis examining the safety and efficacy of MMF in ILD associated with systemic sclerosis, conducted by Tzouvelekis et al included both retrospective and one prospective study. The outcomes of their study align with our findings, indicating an acceptable safety profile for MMF without clear evidence regarding its effectiveness on pulmonary function. 22 Similarly, network meta-analysis in systemic sclerosis associated ILD did not identify significant treatment efficacy of MMF, nor AZA in combination with cyclosporin-A. 23 Further studies are necessary across ILD diagnoses to ascertain potential efficacy in disease subtypes.
This study employed a comprehensive search strategy and strict inclusion criteria, which focused on prospective designs and trials. To support quality, estimates were specifically provided for trial designs along with GRADE assessment. We did not include restrictions on study language or cohort size. MMF and AZA were evaluated in prespecified subgroup analysis based on drug. Where study designs included other treatments, data were collected to support interpretation of MMF or AZA with omission of the drug in comparator arms. Effects regarding AZA should be interpreted with great caution due to limited studies and insufficient studies for primary analysis. Those involving AZA included an active intervention of Cyclosporin-A in the comparator, with addition of AZA in the treatment group, precluded specific interpretation of AZA alone. The limited representation of AZA in the recent literature may be partially attributed to the results of the PANTHER trial, where AZA in combination with n-acetylcysteine and prednisone led to worse outcomes in patients with IPF. 24 Mankikian et al designed an RCT randomising rituximab+MMF versus MMF, we extracted data only from the MMF arm for secondary endpoints. 12 Furthermore, studies were not consistent in ILD diagnosis inclusion, with the majority of prospective observational studies including systemic sclerosis-associated ILD; trials included IPF, non-specific interstitial pneumonia and CTD-ILD, which may contribute to heterogeneity in effect estimates. While ongoing studies were identified, MMF studies did not included blinded phase III RCTs and no AZA studies were identified.
In conclusion, the beneficial impact of MMF and AZA on pulmonary function in patients with ILD is uncertain with some weak evidence that suggests a need to further investigate the effect of MMF in preserving function. While MMF and AZA were generally well tolerated in patients with ILD, it is important to note that the certainty of effects on pulmonary function was very low. Further well-designed RCTs across diagnoses of fibrotic and inflammatory ILD are necessary to support high certainty evidence.
Patient consent for publication.
Not applicable.
No ethical approval was sought as the study uses summary information from published literature.
We express our gratitude to librarian Jacqueline Kemp, Imperial College London, for her valuable assistance in the development of the search strategy. Additionally, we would like to extend our thanks to Dr Liu Bin, Imperial College London, for providing the translation of Chinese manuscripts.
Supplementary data.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Twitter @istamina, @IPFdoc
FL and IS contributed equally.
Collaborators REMAP-ILD Consortium: Alexandre Biasi Cavalcanti (Hospital of Coracao), Ali Mojibian (Black Tusk Research Group), Amanda Bravery (Imperial College Clinical Trials Unit), Amanda Goodwin (University of Nottingham), Ana Etges (Federal University of Rio Grande do Sul), Ana Sousa Marcelino Boshoff (Imperial College Clinical Trials Unit), Andreas Guenther (Justus-Liebig-University of Giessen), Andrew Briggs (London School of Hygiene and Tropical Medicine), Andrew Palmer (University of Tasmania), Andrew Wilson (University of East Anglia), Anjali Crawshaw (University Hospitals Birmingham), Anna-MariaHoffmann-Vold (Oslo University Hospital), Anne Bergeron (University Hospitals Geneva), Anne Holland (Monash University), Anthony Gordon (Imperial College London), Antje Prasse (Hannover Medical School), Argyrios Tzouvelekis (Yale University), Athina Trachalaki (Imperial College London), Athol Wells (Royal Brompton Hospital), Avinash Anil Nair (Christian Medical College Vellore), Barbara Wendelberger (Berry Consultants), Ben Hope-Gill (Cardiff and Vale University Hospital), Bhavika Kaul (U.S. Department of Veterans Affairs Center for Innovation in Quality, Effectiveness, and Safety; Baylor College of Medicine and University of California San Francisco), Bibek Gooptu (University of Leicester), Bruno Baldi (Pulmonary Division, Heart Institute (InCor), University of Sao Paulo Medical School, Sao Paulo, Brazil), Bruno Crestani (Public Assistance Hospital of Paris), Carisi Anne Polanczyk (Federal University of Rio Grande do Sul), Carlo Vancheri (University of Catania), Carlos Robalo (European Respiratory Society), Charlotte Summers (University of Cambridge), Chris Grainge (University of Newcastle), Chris Ryerson (Department of Medicine and Centre of Heart Lung Innovations, University of British Columbia), Christophe von Garnier (Centre Hospitalier Universitaire Vaudois), Christopher Huntley (University Hospitals Birmingham), Claudia Ravaglia (University of Bologna), Claudia Valenzuela (Hospital Universitario de La Princesa), Conal Hayton (Manchester University Hospital), Cormac McCarthy (University College Dublin), Daniel Chambers (Queensland Health), Dapeng Wang (National Heart and Lung Institute, Imperial College London), Daphne Bablis (Imperial College Clinical Trials Unit), David Thicket (University of Birmingham), David Turner (University of East Anglia), Deepak Talwar (Metro Respiratory Centre Pulmonology & Sleep Medicine), Deji Adegunsoye (University of Chicago), Devaraj Anand (Royal Brompton Hospital), Devesh Dhasmana (University of St. Andrews), Dhruv Parek (Brimingham University), Diane Griffiths (University Hospitals Birmingham), Duncan Richards (Oxford University), Eliana Santucci (Hospital of Coracao), Elisabeth Bendstrup (Aarhus University), Elisabetta Balestro (University of Padua), Eliza Tsitoura (University of Crete), Emanuela Falaschetti (Imperial College London), Emma Karlsen (Black Tusk Research Group), Ena Gupta (University of Vermont Health Network), Erica Farrand (University of California, San Fransisco), Fasihul Khan (University of Nottingham), Felix Chua (Royal Brompton Hospital), Fernando J Martinez (Weill Cornell Medicine), Francesco Bonella (Essen University Hospital), Francesco Lombardi (Division of Pulmonary Medicine, Fondazione Policlinico Universitario Agostino Gemelli IRCCS), Gary M Hunninghake (Brigham and Women's Hospital), Gauri Saini (Nottingham University Hospital), George Chalmers (Glasgow Royal Infirmary), Gisli Jenkins (Imperial College London), Gunnar Gudmundsson (University of Iceland), Harold Collard (University of California, San Francisco), Helen Parfrey (Royal Papworth Hospital NHS Foundation Trust), Helmut Prosch (Medical University of Vienna), Hernan Fainberg (Imperial College London), Huzaifa Adamali (North Bristol NHS Trust), Iain Stewart (National Heart and Lung Institute, Imperial College London), Ian Forrest (Newcastle Hospitals NHS Foundation Trust), Ian Glaspole (Alfred Hospital), Iazsmin Bauer-Ventura (The University of Chicago), Imre Noth (University of Virginia), Ingrid Cox (University of Tasmania), Irina Strambu (University of Medicine and Pharmacy), Jacobo Sellares (Hospital Clínic de Barcelona), James Eaden (Sheffield University Hospitals), Janet Johnston (Manchester Royal Infirmary NHS Foundation Trust), Jeff Swigris (National Jewish Health), John Blaikley (Manchester University), John S Kim (University of Virginia), Jonathan Chung (The University of Chicago), Joseph A Lasky (Tulane & Pulmonary Fibrosis Foundation), Joseph Jacob (University College London), Joyce Lee (University of Colorado), Juergen Behr (Ludwig Maximilian University of Munich), Karin Storrer (Federal University of Sao Paulo), Karina Negrelli (Hospital of Curacao), Katarzyna Lewandowska (Institute of Tuberculosis and Lung Diseases), Kate Johnson (The University of British Colombia), Katerina Antoniou (University of Crete), Katrin Hostettler (University Hospital Basel), Kerri Johannson (University of Calgary), Killian Hurley (Royal College of Surgeons, Ireland), Kirsty Hett (Cardiff and Vale University Health Board), Larissa Schwarzkopf (The Institute for Therapy Research), Laura Fabbri (National Heart and Lung Institute, Imperial College London), Laura Price (Royal Brompton Hospital), Laurence Pearmain (Manchester University), Leticia Kawano-Dourado (Hcor Research Institute, Hospital do Coracao, Sao Paulo, Brazil. 2. Pulmonary Division, University of Sao Paulo, Sao Paulo, Brazil. 3. MAGIC Evidence Ecosystem Foundation, Oslo, Norway), Liam Galvin (European Pulmonary Fibrosis Federation), Lisa G. Spencer (Liverpool University Hospitals NHS Foundation Trust), Lisa Watson (Sheffield University Hospitals), Louise Crowley (Queen Elizabeth Hospital, University Hospitals Birmingham), Luca Richeldi (Agostino Gemelli IRCCS University Hospital Foundation), Lucilla Piccari (Department of Pulmonary Medicine, Hospital del Mar, Barcelona (Spain)), Manuela Funke Chambour (University of Bern), Maria Molina-Molina (IDIBELL Bellvitge Biomedical Research Institute), Mark Jones (Southampton University), Mark Spears (University of Dundee Scotland), Mark Toshner (University of Cambridge), Marlies Wijsenbeek-Lourens (Erasmus University Medical Hospital), Martin Brutsche (Kantonsspital St.Gallen), Martina Vasakova (Faculty Thomayer Hospital), Melanie Quintana (Berry Consultants), Michael Gibbons (University of Exeter), Michael Henry (Cork University Hospital), Michael Keane (University College Dublin), Michael Kreuter (Heidelberg University Hospital), Milena Man Iuliu Hatieganu (University of Medicine and Pharmacy), Mohsen Sadatsafavi (The University of British Colombia), Naftali Kaminski (Yale University), Nazia Chaudhuri (Ulster University), Nick Weatherley (Sheffield University Hospitals), Nik Hirani (The University of Edinburgh), Ovidiu Fira Mladinescu Victor Babes (University of Medicine and Pharmacy), Paolo Spagnolo (University of Padua), Paul Beirne (Leeds Teaching Hospitals NHS Foundation Trust), Peter Bryce (Pulmonary Fibrosis Trust), Peter George (Royal Brompton Hospital), Philip L Molyneaux (Imperial College London), Pilar Rivera Ortega (Interstitial Lung Disease Unit, Department of Respiratory Medicine, Wythenshawe Hospital. Manchester University NHS Foundation Trust. United Kingdom.), Radu Crisan-Dabija (University of Medicine and Pharmacy "Grigore T. Popa" Iasi), Rahul Maida (University of Birmingham), Raphael Borie (Public Assistance Hospital of Paris), Roger Lewis (Berry Consultants), Rui Rolo (Braga Hospital), Sabina Guler (University Hospital of Bern), Sabrina Paganoni (Massachusetts General Hospital), Sally Singh (University of Leicester.), Sara Freitas (University Hospital Coimbra), Sara Piciucchi (Department of Radiology, GB Morgagni Hospital; Azienda USL Romagna), Shama Malik (Action for Pulmonary Fibrosis), Shaney Barratt (North Bristol NHS Trust), Simon Hart (University of Hull), Simone Dal Corso (Monash University), Sophie Fletcher (Southampton University), Stefan Stanel (Manchester University NHS Foundation Trust), Stephen Bianchi (Thornbury Hospital), Steve Jones (Action for Pulmonary Fibrosis), Wendy Adams (Action for Pulmonary Fibrosis).
Contributors FL: protocol development, formal analysis, data curation, writing–original draft. IS: protocol development, formal analysis, methodology, supervision, writing–original draft, guarantor. LF: protocol development, data curation, writing–review and editing. WA: protocol development, writing–review and editing. LK-D: protocol development, writing–review and editing. CJR: protocol development, writing–review and editing. GJ: conceptualisation, protocol development, supervision, writing–review and editing.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests GJ is supported by a National Institute for Health Research (NIHR) Research Professorship (NIHR reference RP-2017-08-ST2-014). GJ is a trustee of Action for Pulmonary Fibrosis and reports personal fees from Astra Zeneca, Biogen, Boehringer Ingelheim, Bristol Myers Squibb, Chiesi, Daewoong, Galapagos, Galecto, GlaxoSmithKline, Heptares, NuMedii, PatientMPower, Pliant, Promedior, Redx, Resolution Therapeutics, Roche, Veracyte and Vicore. CJR reports grants from Boehringer Ingelheim, and honoraria or consulting fees from Boehringer Ingelheim, Pliant Therapeutics, Astra Zeneca, Trevi Therapeutics, Veracyte, Hoffmann-La Roche, Cipla. FL, IS, LF, WA and LK-D report no competing interests.
Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
IMAGES
COMMENTS
Below are the steps to prepare a data before quantitative research analysis: Step 1: Data Collection. Before beginning the analysis process, you need data. Data can be collected through rigorous quantitative research, which includes methods such as interviews, focus groups, surveys, and questionnaires. Step 2: Data Cleaning.
6. Kissmetrics. Kissmetrics is a software for quantitative data analysis that focuses on customer analytics and helps businesses understand user behavior and customer journeys. Kissmetrics lets you track user actions, create funnels to analyze conversion rates, segment your user base, and measure customer lifetime value.
Here, we explore the top 10 quantitative data analysis software options available today. 1. QuestionPro. Known for its robust survey and research capabilities, QuestionPro is a versatile platform that offers powerful data analysis tools tailored for market research, customer feedback, and academic studies.
Quantitative data analysis is one of those things that often strikes fear in students. It's totally understandable - quantitative analysis is a complex topic, full of daunting lingo, like medians, modes, correlation and regression.Suddenly we're all wishing we'd paid a little more attention in math class…. The good news is that while quantitative data analysis is a mammoth topic ...
Google Cloud AutoML contains a suite of tools across categories from structured data to language translation, image and video classification. As more and more organizations adopt machine learning, there will be a growing demand for data analysts who can use AutoML tools to automate their work easily. 7. SAS.
Analysts commonly use tools during the following stages of the data analysis process: Data mining: Data mining helps users find the key characteristics of their data so they can apply this knowledge to real-world problems, and data mining software helps automate this process by looking for patterns and trends within the data.
1. Excel. Microsoft Excel is one of the most common software used for data analysis. In addition to offering spreadsheet functions capable of managing and organizing large data sets, Excel also includes graphing tools and computing capabilities like automated summation or "AutoSum.". Excel also includes Analysis ToolPak, which features data ...
Statistical analysis means investigating trends, patterns, and relationships using quantitative data. It is an important research tool used by scientists, governments, businesses, and other organizations. To draw valid conclusions, statistical analysis requires careful planning from the very start of the research process. You need to specify ...
Here's how to make sense of your company's numbers in just four steps: 1. Collect data. Before you can actually start the analysis process, you need data to analyze. This involves conducting quantitative research and collecting numerical data from various sources, including: Interviews or focus groups.
8. Weight customer feedback. So far, the quantitative data analysis methods on this list have leveraged numeric data only. However, there are ways to turn qualitative data into quantifiable feedback and to mix and match data sources. For example, you might need to analyze user feedback from multiple surveys.
quantitative and geospatial data unstructured text as data. Imagine that you have data for all the deaths of all Medicare beneficiaries in the US 2000-2012 (~half a million person-years) and want to model the effect of air pollution levels on death, controlling for other factors that also affect death (such as smoking, BMI).
SPSS is the most popular quantitative analysis software program used by social scientists. Made and sold by IBM, it is comprehensive, flexible, and can be used with almost any type of data file. However, it is especially useful for analyzing large-scale survey data . It can be used to generate tabulated reports, charts, and plots of ...
Although there are many other methods to collect quantitative data. Those mentioned above probability sampling, interviews, questionnaire observation, and document review are the most common and widely used methods for data collection. With QuestionPro, you can precise results, and data analysis.
Stata was first released in January 1985 as a regression and data management package with 44 commands, written by Bill Gould and Sean Becketti. The name Stata is a syllabic abbreviation of the words statistics and data. The graphical user interface (menus and dialog boxes) was released in 2003. Users. Economics; Sociology; Political Science ...
Definition of research in data analysis: According to LeCompte and Schensul, research data analysis is a process used by researchers to reduce data to a story and interpret it to derive insights. The data analysis process helps reduce a large chunk of data into smaller fragments, which makes sense. Three essential things occur during the data ...
A brief but thorough introduction to analyzing data using Stata software. The makers of Stata maintain this extensive library of short video tutorials. The official user guide, along with manuals and examples for using specific statistical methods in Stata. Beginner-friendly guide to Stata from UCLA's Advanced Research Computing.
Abstract. Quantitative data analysis serves as part of an essential process of evidence-making in health and social sciences. It is adopted for any types of research question and design whether it is descriptive, explanatory, or causal. However, compared with qualitative counterpart, quantitative data analysis has less flexibility.
Revised on June 22, 2023. Quantitative research is the process of collecting and analyzing numerical data. It can be used to find patterns and averages, make predictions, test causal relationships, and generalize results to wider populations. Quantitative research is the opposite of qualitative research, which involves collecting and analyzing ...
Replicable: Quantitative research aims to be replicable, meaning that other researchers should be able to conduct similar studies and obtain similar results using the same methods. Statistical analysis: Quantitative research involves using statistical tools and techniques to analyze the numerical data collected during the research process ...
The article covers a brief outline of the variables, an understanding of quantitative and qualitative variables and the measures of central tendency. An idea of the sample size estimation, power analysis and the statistical errors is given. Finally, there is a summary of parametric and non-parametric tests used for data analysis.
Two commonly used statistical analysis packages described later in this chapter (SPSS and SAS) offer comprehensive data analysis tools for hypothesis testing. Spreadsheet and Relational Database Packages. Many application tools not created for quantitative data research have become sufficiently powerful to be used for that today.
In quantitative data analysis, ... "Research tool" or "Research instrument" is a mean s of collecting data in research, like questionnaires, interv iews, and . observation.
Quantitative Research: Methodological Distinctions and Approach. Quantitative research is distinguished by its reliance on numerical data and statistical analysis, setting it apart from qualitative methods. Researchers often use structured tools, such as surveys or experiments, to gather quantifiable data.
Statistical analysis means investigating trends, patterns, and relationships using quantitative data. It is an important research tool used by scientists, governments, businesses, and other organisations. To draw valid conclusions, statistical analysis requires careful planning from the very start of the research process. You need to specify ...
This study examines the impact of Generative Artificial Intelligence on academic research, focusing on its application to qualitative and quantitative data analysis, and provides insights into how researchers may utilise GenAI tools responsibly and ethically. This study examines the impact of Generative Artificial Intelligence (GenAI) on academic research, focusing on its application to ...
This study examines the impact of Generative Artificial Intelligence (GenAI) on academic research, focusing on its application to qualitative and quantitative data analysis. As GenAI tools evolve rapidly, they offer new possibilities for enhancing research productivity and democratising complex analytical processes. However, their integration into academic practice raises significant questions ...
Background: Preeclampsia is a potentially fatal complication during pregnancy, characterized by high blood pressure and the presence of excessive proteins in the urine. Due to its complexity, the prediction of preeclampsia onset is often difficult and inaccurate. Objective: This study aimed to create quantitative models to predict the onset gestational age of preeclampsia using electronic ...
Objectives Mycophenolate mofetil (MMF) and azathioprine (AZA) are immunomodulatory treatments in interstitial lung disease (ILD). This systematic review aimed to evaluate the efficacy of MMF or AZA on pulmonary function in ILD. Design Population included any ILD diagnosis, intervention included MMF or AZA treatment, outcome was delta change from baseline in per cent predicted forced vital ...