How to Write Chapter 4 of Your Project, Step by Step

Chapter 4 is where your project stops being a plan and starts producing answers. It’s usually titled “Data Presentation, Analysis and Interpretation” (some departments say “Results” or “Results and Discussion”), and it’s the chapter supervisors read most closely and defence panels ask the most questions about.

The good news: Chapter 4 follows a predictable structure. Once you know it, the chapter almost writes itself around your data.

Before you start: check your department’s guidelines or a recent approved project. Some departments put the discussion of findings in Chapter 4; others save it for Chapter 5. The structure below covers both.

The standard structure of Chapter 4

4.1 Introduction

One short paragraph. Say what the chapter does and how it’s organised.

This chapter presents the analysis of data collected from 243 respondents. It begins with the response rate and the demographic characteristics of respondents, then answers the research questions and tests the hypotheses stated in Chapter One.

4.2 Response rate

Many students skip this, and supervisors always ask for it. State how many questionnaires you distributed, how many came back, and how many were usable.

Response rate = (usable questionnaires ÷ questionnaires distributed) × 100

If you distributed 260 and 243 were usable, your response rate is 93.5%. If some were rejected (incomplete, duplicated, or obviously careless), say how many and why. That shows you cleaned your data properly.

4.3 Demographic characteristics of respondents

Present your respondents’ profile, usually as a frequency table: gender, age group, level of study, marital status, and so on.

Table 4.1: Demographic characteristics of respondents (N = 243)

Variable	Category	Frequency	Percentage
Gender	Male	112	46.1
	Female	131	53.9
Level of study	100 level	58	23.9
	200 level	61	25.1
	300 level	63	25.9
	400 level	61	25.1

(Illustrative data.)

Then interpret it in a few sentences. Don’t repeat every number in the table; point out what matters:

Table 4.1 shows that slightly more of the respondents were female (53.9%), and respondents were spread fairly evenly across the four levels of study.

4.4 Answering the research questions

This is the heart of the chapter. Take your research questions in the same order as Chapter One, and answer each one with a table and an interpretation.

For questionnaire items on a Likert scale, the usual approach is to report the mean and standard deviation of each item, then compare each mean to a decision rule (criterion mean).

For a 4-point scale (Strongly Agree = 4, Agree = 3, Disagree = 2, Strongly Disagree = 1), the criterion mean is:

(4 + 3 + 2 + 1) ÷ 4 = 2.50

Items with a mean of 2.50 or above are taken as “agreed”; items below 2.50 are “disagreed”. For a 5-point scale, the criterion mean is 3.00.

Table 4.2: Mean responses on study habits of respondents (N = 243)

S/N	Item	Mean	SD	Decision
1	I study at a fixed time every day	2.31	0.94	Disagreed
2	I revise my notes after every lecture	2.68	0.88	Agreed
3	I prepare for tests only at the last minute	2.97	0.91	Agreed
	Grand mean	2.65		Agreed

(Illustrative data.)

Table 4.2 shows that respondents agreed that they revise their notes after lectures (x̄ = 2.68), but they also agreed that they prepare for tests only at the last minute (x̄ = 2.97). With a grand mean of 2.65, respondents reported moderately good study habits overall.

A detail supervisors check: negatively worded items (like item 3 above) must be reverse-coded before you calculate a grand mean or a scale total. Otherwise “bad” habits push your study-habits score up.

4.5 Testing the hypotheses

Restate each hypothesis (usually in its null form), name the test, show the table, and give the decision.

H₀₁: There is no significant difference in the academic performance of male and female students.

Then:

1. Name the test and why. “An independent samples t-test was used because academic performance (CGPA) is numerical and gender has two groups.” Not sure which test applies? Read Which statistical test should I use?.
2. Present the table: the key numbers only, in APA style, not a raw SPSS screenshot.
3. Give the decision: if p < .05, reject the null hypothesis; if p ≥ .05, fail to reject it.
4. Say what it means in plain words.

The result showed no significant difference in CGPA between male and female students, t(241) = 1.24, p = .217. The null hypothesis was therefore retained: in this sample, gender was not related to academic performance.

Notice the wording: we fail to reject or retain the null hypothesis. We never say we “accepted” it, and we never say a result “proved” anything.

4.6 Discussion of findings

If your department puts the discussion in Chapter 4, this is where you connect your results to the literature from Chapter Two. For each major finding:

• State the finding briefly.
• Compare it with previous studies. Name the studies from your Chapter Two that agree with it, and the ones that don’t.
• Explain why. Why might your result agree or differ? Different population, location, time, or method?
• Link back to your theory from Chapter Two, if you used one.

A good discussion doesn’t just list who agreed with you. It explains why your results look the way they do.

4.7 Summary of findings

End with a short numbered list of your key findings, one per research question or hypothesis. This makes Chapter 5 (summary, conclusion and recommendations) much easier to write.

Common Chapter 4 mistakes (and how to avoid them)

1. Pasting raw SPSS output. Rebuild every table in Word, in APA style, with a clear number and title.
2. Tables without interpretation. Every table needs a paragraph explaining what it shows.
3. Repeating every number in the text. Highlight what matters; the table holds the details.
4. Saying a hypothesis was “proved” or “accepted”. Use “rejected” or “retained / failed to reject”.
5. Ignoring non-significant results. A non-significant result is still a finding. Report it and discuss it.
6. Forgetting the response rate and data cleaning. Say how many responses you used and why.
7. Using the wrong test. Your test must match your data type and your question.
8. Not reverse-coding negative items. It quietly ruins your means, totals and reliability.
9. Answering research questions out of order. Follow the order in Chapter One exactly.

Before your defence

Your panel will ask questions like:

• Why did you use this test?
• What does this p-value mean?
• Why do you think your result differs from previous studies?

If you can answer those three questions for every table in your Chapter 4, you’re ready. If you can’t yet, that’s exactly what I help students with: not just running the analysis, but making sure you understand every number you present.