How To Interpret Your Research Findings (including Regression Results)

You have collected your data, run the analysis, and your screen is now full of tables. Percentages, means, p-values, a regression output that looks like a spreadsheet exploded. And the question every student asks at this point is the same one: what does all of this actually mean? That question is what interpretation is. It is not a second round of analysis. It is the part where you stop describing numbers and start explaining them, in words your reader can follow, and in a way that answers the questions you set out to ask.

Most final year projects lose marks here, not because the analysis was wrong, but because the interpretation was missing. A table with no explanation underneath it is unfinished work. A regression result copied into the chapter with no sentence saying what it means is a number without a story. Your supervisor does not need you to repeat every figure. They need to see that you understand what the figures are saying, and that you can connect them back to your objectives, your literature, and the real world your study is about. This guide explains how to do that, whether your data is numbers, interviews, or a mix of both.

What interpretation means, in plain words

Interpretation is attaching meaning to your results. Analysis tells you what happened in your data. Interpretation tells you why it matters. If your analysis says 68 percent of respondents agreed that online lectures improved their learning, your interpretation explains what that suggests about your research question, whether it agrees with what earlier studies found, and what it means for the people your study is about.

Think of it in three layers. The first layer is what the result is, stated simply. The second layer is what it means for your research question, including whether it supports or contradicts your hypothesis. The third layer is why it might have turned out that way, and what it implies. Most students stop at the first layer. Good interpretation reaches the third, and does it honestly, including when the results are messy or unexpected.

Start with your research questions, not with your tables

Before you write a single sentence of interpretation, go back to Chapter One and read your research questions again. Then read your objectives. Interpretation is organised around those questions, not around the order SPSS printed your output. For each research question, gather the results that answer it, and write about them together. One question, one section, one clear answer. That structure is what keeps your discussion focused, and it is what your examiner will look for first.

This is also why your research questions matter so much. If a question is vague, your interpretation will be vague too. If you have not yet settled your questions, the ProjectNG guide on how to write research questions and hypotheses for your final year project will help you tighten them before you interpret anything. And if you are still deciding how your results fit into the whole project, the guide on how to write Chapter Four of your final year project explains where presentation ends and discussion begins.

Interpreting descriptive results: percentages and means

Most undergraduate projects begin with descriptive results, and they are easier to interpret than students expect. The rule is simple: say what the number shows, then say what it suggests.

Suppose 72 percent of your respondents said they prefer physical lectures to online lectures. Do not just write "72 percent preferred physical lectures" and move on. That is description, not interpretation. Add the meaning: most students in your sample still value face to face teaching, which suggests that the convenience of online learning has not replaced the interaction students get in a classroom. Then connect it to your literature: does this agree with what earlier researchers found, or does it differ, and why might your setting be different?

Means work the same way. If students rated their satisfaction with library services at 3.8 out of 5, explain what that level means in your context. Is 3.8 high for your scale, or only moderate? What does it suggest about the service? A common mistake is to report a mean without saying whether it is high or low, and without comparing it to anything. A number on its own is not a finding. A number with a comparison and a meaning is.

Three checks before you move on

For every descriptive result, ask yourself three questions. First, have I said what the figure means in plain words? Second, have I linked it to one of my research questions? Third, have I compared it with something, either an earlier study, a benchmark, or another group in my own data? If any answer is no, that paragraph is not finished yet.

Interpreting qualitative findings: themes from interviews

If your study used interviews or focus groups, your findings will be themes, not numbers, and the interpretation works differently. A theme is not a heading you invented. It is a pattern that appeared across several participants, supported by what they actually said. When you interpret a theme, name it clearly, explain what it means, and support it with one or two short quotations from your participants. Then explain how the theme answers your research question.

For example, if several students told you they study better at night because the hostel is quieter, the theme might be "quiet environment shapes study time". The interpretation is not the quotation itself. It is your explanation of what the pattern means: that study habits in your setting are shaped by living conditions as much as by personal preference, which has implications for how the university timetables evening study spaces. That move, from what people said to what it means, is the whole point of qualitative interpretation. If you are unsure whether your study is qualitative, quantitative, or a mix, the ProjectNG guide on qualitative versus quantitative research for final year projects will help you decide, and that decision shapes how you interpret everything that follows.

Interpreting regression results, step by step

Regression is where many students panic, because the output looks technical and the numbers seem to carry hidden meanings. In practice, you only need to understand five things, and you can read them in order, from the top of the output to the bottom.

First, look at the model summary and find R-squared. This tells you what proportion of the variation in your dependent variable your model explains. An R-squared of 0.52 means your independent variables together explain about 52 percent of the variation. That is a useful, honest statement. It also means 48 percent is not explained by your model, and saying so shows you understand the result rather than overselling it.

Second, check the ANOVA table and the F statistic, with its p-value. This tells you whether the model as a whole is statistically significant. If the p-value is below 0.05, the model is doing better than chance at explaining your dependent variable. If it is above 0.05, the model is not significant, and you must say so plainly, rather than picking out individual coefficients as if the model worked.

Third, go to the coefficients table and look at each variable in turn. For each one, note the sign of the coefficient, positive or negative, and its size. The sign tells you the direction of the relationship. The size tells you how much the dependent variable changes when that independent variable rises by one unit, holding the others constant. Always translate the number back into your real variables. A coefficient of 0.64 on study hours does not mean much on its own. Saying "each additional hour of study per week was associated with an increase of 0.64 marks in the test score, holding other factors constant" is an interpretation your examiner can follow.

Fourth, check the p-value or significance column for each coefficient. A p-value below 0.05 means that variable's relationship with the dependent variable is statistically significant in your sample. A p-value above 0.05 means you do not have enough evidence to say the relationship exists, even if the coefficient looks large. Never describe a non-significant result as if it were a real effect. The correct sentence is "the relationship was not statistically significant", not "there was a small effect".

Fifth, look at the standardised coefficients, labelled Beta, if you want to compare which variable matters most. Because these are on a common scale, a Beta of 0.45 matters more than a Beta of 0.18 in the same model. This is how you answer the question every supervisor asks: of all your variables, which one had the strongest influence?

Here is a small example, so you can see how the pieces fit together. Imagine a study predicting student test scores from study hours and class attendance:

Read as plain English, this says: the model explains 52 percent of the variation in test scores. Both study hours and attendance are statistically significant predictors. Study hours has the stronger influence, with each extra hour per week associated with 0.64 extra marks, holding attendance constant. That short paragraph is all a reader needs. The table supports it, and the text explains it. If you want a fuller walkthrough of regression output, the guide on how to interpret regression output, coefficients, R-squared and p-values covers the same ground in more detail, and the ProjectNG guide on how to use SPSS for data analysis in your final year project will help you get the output right in the first place.

What a p-value does and does not tell you

The p-value causes more confusion than any other number in student research, so it deserves a short section of its own. A p-value below 0.05 means the result is unlikely to have happened by chance alone, if there were truly no relationship in the population. It does not mean the effect is large, important, or true in the real world. A tiny effect can be statistically significant if your sample is large, and a large effect can be non-significant if your sample is small. That is why you should always report the size of the effect alongside the p-value, and discuss whether the effect matters in practice, not just in statistics.

It also does not tell you that your hypothesis is proven. "Significant" in statistics means "unlikely to be due to chance", not "important" and not "certain". Write about your results with that caution, and your examiner will trust your judgement more, not less. If your sample was small, remember that your sample size affects what your p-values can show. The ProjectNG guide on how to determine sample size for your final year project explains why this matters before you collect data, and it is worth reading if you are planning a new study.

Correlation is not causation, and other traps

The most common interpretation error in student projects is treating a relationship as a cause. If students who attend more classes also score higher, that does not prove attendance causes higher scores. It could be that motivated students both attend more and study harder, or that another factor explains both. Your data can show that two things move together. Only a well designed experiment can show that one causes the other. So write "associated with" or "related to", not "caused" or "led to", unless your design genuinely supports a causal claim.

There are a few other traps worth avoiding. Do not cherry pick the results that fit your expectations and quietly ignore the ones that do not. Do not overstate a weak finding with strong language. Do not claim your results apply to all Nigerian universities when your sample came from one department in one university. And do not copy SPSS output tables into your chapter without explanation. Your examiner wants your words, not the software's.

When your results surprise you

Sometimes the data disagrees with your hypothesis, or with what the literature said. This feels like failure, but it is often the most interesting part of a project. A result that contradicts your expectations is still a result, and it still needs interpreting. Ask why it might have happened. Was your sample different from earlier studies? Was the setting different? Was there a limitation in how you measured something? Say so honestly, and compare your finding with the earlier work directly. "This finding differs from Adeyemi (2021), who reported a strong positive relationship, possibly because this study used a smaller sample from a single department, while Adeyemi sampled across three universities." That sentence is more impressive than a result that agrees with everyone, because it shows you are thinking, not just reporting.

Unexpected results also belong in your limitations section, and they often point to the best recommendations for further research. If your study could not answer a question properly, say what a future study should do differently. That is not a weakness. It is how research moves forward.

A simple sentence pattern you can adapt

If you are staring at a blank page, use this pattern for each major finding, and adapt the words to your own study:

The results show that [what you found, in plain words]. This answers research question [number], and it [supports / does not support] the hypothesis that [your hypothesis]. This agrees with [an earlier study], but differs from [another study], possibly because [a reason linked to your setting or method]. The implication is that [what this means in practice], although the finding should be read with caution because [a limitation].

That one paragraph contains description, meaning, comparison, implication, and limitation. Write one for each major finding, and your discussion will already be stronger than most. It also keeps you honest, because the pattern forces you to include the comparison and the caution, which are the two parts students most often leave out.

Before you submit your interpretation

Read your discussion chapter once through, with your research questions beside you, and check that every major finding has been explained in plain words, linked to a question, and compared with the literature. Check that every regression coefficient you mention has a sentence saying what it means in terms of your real variables. Check that you have not claimed causation where you only found association. And check that your limitations are stated, not hidden. If your method needs revisiting before you can interpret confidently, the ProjectNG guide on how to write a strong Chapter Three on methodology will help you confirm that your design supports the claims you want to make.

Interpretation is a skill, and like every other part of a final year project, it improves with practice. The first time you read a regression table it will look like a puzzle. The second time, you will know exactly where to look. And the third time, you will be explaining it to a classmate who is stuck where you were. Start with your questions, translate every number into a sentence about your real variables, compare honestly with what others found, and say what your findings mean in practice. Do that, and your results will stop being a set of tables and start being an argument your reader can follow.