What is regression analysis and when do you use it

TL;DR

• Regression analysis measures how one variable predicts another.

• Linear regression is the most common starting point for student researchers.

• Use it when your data has a clear independent and dependent variable.

• Results must be interpreted carefully — correlation is not causation.

• Most peer-reviewed journals expect regression output to be reported in full.

You have collected your data. You have cleaned it, organised it, and stared at it long enough that the numbers have started to blur. Now comes the question every student researcher eventually faces: how do I actually show that one thing affects another? That is exactly where regression analysis earns its place. It is one of the most widely used statistical methods in published research, and understanding what it does — and when to use it — can be the difference between a paper that gets accepted and one that gets sent back for major revisions.

The problem is that most introductions to regression analysis are written for statistics students, not for researchers who are trying to finish a paper. They lead with formulas and end with more formulas. This post takes a different approach. It explains what regression analysis actually is, when it fits your research question, and how to report it in a way that journal reviewers will take seriously.

If you are still deciding whether your research question is ready for a statistical method this specific, it helps to first understand what peer-reviewed research is and why it sets the standards it does. That context matters before you run a single model.

What Is Regression Analysis

Regression analysis is a statistical method that estimates the relationship between one outcome variable and one or more predictor variables. It tells you how much the outcome changes when a predictor changes, and how confident you can be in that estimate. In plain terms: it helps you say, with evidence, that as X changes, Y tends to change in a predictable direction.

The outcome you are trying to explain is called the dependent variable. The variable you believe is doing the explaining is called the independent variable. In a simple linear regression, you have one of each. In a multiple regression, you have one dependent variable and several independent variables, which lets you control for factors that might otherwise muddy your results.

Consider a concrete example. A student researching study habits and academic performance might ask: does the number of hours spent reviewing notes each week predict a student's exam score? Hours of review would be the independent variable. Exam score would be the dependent variable. Regression analysis would produce a line that best fits the data, a slope that tells you how much the score changes per additional hour, and a measure of how well the model fits overall.

That measure of fit is called R-squared, written as R². It ranges from 0 to 1. An R² of 0.65, for instance, means the independent variable explains 65 percent of the variation in the dependent variable. The remaining 35 percent is explained by factors not included in the model. Understanding what R² does and does not tell you is essential before you report it in a paper.

When Do You Use Regression Analysis in Student Research

Regression analysis is appropriate when your research question asks how much one variable predicts or is associated with another, your data is numerical or can be treated as continuous, and you have enough observations to produce reliable estimates. Most statisticians recommend at least 10 to 20 observations per predictor variable in your model, a guideline discussed in introductory biostatistics texts such as Vittinghoff et al.'s Regression Methods in Biostatistics.

There are three situations where regression analysis is a natural fit for student research projects.

1. You want to quantify a relationship. If your hypothesis states that one variable influences another, regression gives you a number that describes the size and direction of that influence. A positive slope means the variables move together. A negative slope means they move in opposite directions.

2. You want to control for confounding variables. If you suspect that a third variable is affecting your results, multiple regression lets you hold it constant statistically. This is one of the most powerful things regression does, and it is why it appears so frequently in published social science and health research.

3. You want to make a prediction. If your research involves forecasting an outcome based on known inputs, regression produces an equation you can use to generate predictions on new data.

Regression is not the right tool when your variables are purely categorical with no natural order, when your sample size is very small, or when your research question is descriptive rather than relational. In those cases, other methods such as chi-square tests or thematic analysis are more appropriate.

If you are working on a project that involves human subjects, it is also worth confirming whether your study requires Institutional Review Board approval before you collect any data. Understanding what IRB approval is and when students need it can save you from a significant problem at the submission stage.

What Is Regression Analysis Telling You About Causation

Regression analysis shows association, not causation. This is one of the most important things to understand before you write up your results. A statistically significant regression coefficient means the relationship is unlikely to be due to chance. It does not mean that one variable causes the other to change.

This distinction matters because journal reviewers will flag it immediately if your discussion section overclaims. You can say that hours of note review were a significant positive predictor of exam scores. You cannot say, based on regression alone, that reviewing notes causes higher scores. That causal claim requires an experimental design, not an observational one.

If you are planning to submit to a journal that publishes student research, such as the International Journal of High School Research or the Columbia Junior Science Journal, both of which accept quantitative work from pre-university researchers, your methods section will need to describe your regression model clearly and your discussion section will need to stay within what the data actually supports. Reviewers at both journals expect methodological precision from student authors. You can learn more about what the International Journal of High School Research publishes and what the Columbia Junior Science Journal looks for before you decide where to submit.

How to Report Regression Results in a Paper

Reporting regression results correctly is as important as running the analysis correctly. A paper that presents incomplete or poorly formatted statistical output will struggle in peer review regardless of how interesting the findings are.

Here is the standard sequence for reporting a simple linear regression in a student research paper.

1. State the model. Identify your dependent and independent variables clearly in the methods section. Name the software you used, whether that is R, SPSS, Python, or another tool.

2. Report the coefficient and standard error. The regression coefficient (often written as b or beta) tells readers the size of the effect. The standard error tells them how precisely that coefficient is estimated.

3. Report the p-value. The p-value tells readers whether the relationship is statistically significant. The conventional threshold in most fields is p less than 0.05, though some disciplines use stricter cutoffs. Report the exact value rather than just writing "significant."

4. Report R-squared. This tells readers how much of the variation in the dependent variable your model explains.

5. Include a regression table. Most journals expect results to be presented in a formatted table, not embedded only in prose. Check the journal's author guidelines for the specific format they require.

For multiple regression, you also need to check and report whether your predictors are correlated with each other, a problem called multicollinearity. High multicollinearity can make individual coefficients unreliable even when the overall model appears to fit well.

If you are preparing a paper and want structured feedback on your methods section before submission, joining the Publication Compass waitlist gives you early access to an AI platform built to help student researchers navigate exactly this stage of the process.

Common Mistakes Students Make With Regression Analysis

Most errors in student regression work fall into a small number of categories. Knowing them in advance makes them easier to avoid.

The first is running regression on too little data. With fewer than 30 observations and multiple predictors, your results are likely to be unstable. The model may fit your sample well but fail to generalise to any other group. This is called overfitting, and it is one of the first things a reviewer will look for.

The second is ignoring the assumptions of linear regression. Linear regression assumes a linear relationship between variables, that the residuals (the differences between predicted and actual values) are normally distributed, and that the variance of residuals is roughly constant across all values of the predictor. Violating these assumptions does not automatically invalidate your results, but it does require you to address the violation in your methods section or use a different model.

The third is treating a non-significant result as a failed study. A p-value above 0.05 does not mean your research question was wrong. It means that, in your sample, you did not find sufficient evidence of the relationship you hypothesised. That is a legitimate finding. Report it honestly. Understanding what rejection actually means and what to do next applies equally to null results as to rejected manuscripts.

The fourth is skipping the assumptions check entirely. Many student researchers run the model, get a significant p-value, and move straight to writing up results. Before you do that, plot your residuals. Look at a Q-Q plot. Run a test for heteroscedasticity. These checks take twenty minutes and can prevent a reviewer from rejecting your paper on methodological grounds.

What Is Regression Analysis Used For in Published High School Research

Regression analysis appears across disciplines in published student research. In psychology and sociology, it is used to examine how demographic variables predict attitudes or behaviours. In environmental science, it is used to model how pollution levels change with temperature or population density. In economics, it is used to estimate how price changes affect demand.

Publication Compass is a software platform designed to help student researchers move from a completed draft to a submitted paper. It helps identify journals suited to your methodology, provides structured feedback on your methods and results sections, and guides you through the submission process. For a student who has run a regression analysis and is now trying to write it up correctly, that kind of structured support addresses a very specific gap.

Understanding what peer review actually involves is also worth doing before you submit. Knowing what peer review is and what happens to your paper helps you anticipate the kinds of questions reviewers are likely to ask about your statistical methods.

Frequently Asked Questions

What is regression analysis in simple terms

Regression analysis is a statistical method that shows how one variable predicts another. It produces a mathematical equation describing the relationship, a measure of how strong that relationship is, and a test of whether the relationship is likely to be real rather than due to random chance. It is widely used in social science, health research, and economics.

When should a student researcher use regression analysis

Use regression analysis when your research question asks how much one numerical variable predicts another, when you have enough data (generally at least 20 to 30 observations per predictor), and when your variables have a plausible linear relationship. It is not appropriate for purely categorical data or for very small samples.

What is the difference between simple and multiple regression

Simple linear regression uses one independent variable to predict one dependent variable. Multiple regression uses two or more independent variables. Multiple regression is more powerful because it lets you control for the influence of additional factors, but it also requires more data and more careful interpretation of results.

Does a significant p-value mean one variable causes the other

No. A significant p-value means the relationship observed in your data is unlikely to be due to chance. It does not establish causation. Causation requires an experimental design where the researcher controls which participants receive which condition. Observational studies using regression can show association but not cause and effect.

What is regression analysis used for in high school research papers

In high school research, regression analysis is most commonly used in psychology, environmental science, public health, and economics projects. It allows student researchers to move beyond describing their data and make a specific, testable claim about the relationship between variables, which strengthens both the argument and the likelihood of publication.

Where to Go From Here

Regression analysis is a precise tool. Used correctly, it lets you make specific, evidence-backed claims about the relationships in your data. Used carelessly, it produces numbers that look authoritative but mislead. The difference comes down to checking your assumptions, reporting your results in full, and keeping your conclusions within what the data actually supports.

If your paper is at the stage where your analysis is done and your methods section needs to hold up to reviewer scrutiny, the next step is making sure the rest of your submission is equally solid. The Publication Compass blog covers the full arc of the research and publication process, from designing a study to understanding what happens after a paper is accepted.