Writing the discussion chapter

A simple pattern for each finding

First choose which findings you will discuss. Every table in the results chapter does not need a discussion. Take the main finding under each objective. If some result surprised you, add that one also. For each finding, go through these five moves.

1. Restate the finding in plain words

   Keep numbers out, or use only one or two. The reader already saw the statistics in the results chapter. Here you just remind them what you found.

2. Say what it means

   Now interpret the finding. Write what it is saying about the people you studied. In a lab study it may be about the process or the material instead.

3. Put it next to earlier studies

   Find which studies got the same result and which ones got a different result. Name those studies, with the year.

4. Reason for any mismatch

   Sometimes your finding and an earlier study will not match. Then suggest a reason for that gap. Maybe the setting was different, or the sample, the period or the way you measured. In most cases you won’t have proof for it, and examiners know that. Use “may be” or “possibly” when you write it.

5. Say what follows from it

   Write one or two sentences on what this finding means for practice, policy or theory. Keep the bigger implications for the end of the chapter.

The moves will not get equal space every time. Say ten earlier studies already got the same finding as you. One sentence on that agreement is enough then, and the remaining space can go to what your study is adding. A surprising finding needs most of its space for move four.

One finding, taken through all five moves

Take an invented study on access to credit for women who run small garment units in Coimbatore. Say the results chapter found that women in a self-help group (SHG) were more likely to have bank loan than women outside any group. But SHG membership made no difference to how big the loan was.

A discussion paragraph on this finding could be written like this.

“Women in a self-help group were more likely to get a bank loan, but the loan amount was not higher for them. So the SHG seems to open the door to bank credit. Once the woman is inside, it does not change much how much the bank will lend. This first result is consistent with earlier work on SHG–bank linkage in Tamil Nadu (Devi & Kumar, 2020; Rajan, 2018). On loan size our result is not in line with Devi and Kumar, who found SHG members got larger loans. One possible reason is the type of enterprise. Their sample came mostly from agriculture and dairy units. In those sectors SHG loans are many times pooled, while a garment unit takes its loan individually against the machinery. For banks and the district SHG federation, this means linkage alone may not be enough for garment units which need working capital to grow.”

All five moves are there in that paragraph. See that the reason starts with “one possible reason”. We used those words on purpose. Nobody checked that reason in the study, it is only a guess. The study names above are made up for this example. In your own chapter, each comparison should carry a real citation that also appears in your literature review.

When your result does not agree with the literature

Many scholars get tense when a finding goes against earlier studies. They need not worry. If your thesis only says again what is already known, the examiner will ask what is new in it. If you explain a disagreement properly, many times it becomes the most interesting part of the discussion.

Check the possible reasons one by one. First see if the setting was different, like another district, sector or time period. Then check if the sample was different in size, age or background. Maybe you measured the variable with another scale or another definition. Or something changed in between, for example a new scheme, a policy change or the pandemic. Mostly one or more of these explains the gap.

Without any evidence, some scholars write that the earlier author made a mistake. Avoid that. A few others quietly drop their own result from the chapter, and that is also wrong. If you never tested a reason, you cannot write it as a fact. Write “this may be because” or “one reason may be”. Examiners accept these words much more easily than a confident guess.

Writing limitations honestly, without sounding defensive

All studies have limits, and examiners know it. What the examiner want to see is that you also know them, and that you have thought about how they affect your results.

A weak limitations section gives only general points, such as “time constraint”, “respondents may not have answered honestly” or “the study is limited to one district”. A better section names each limit, says how it could have affected the findings, and says what you did to reduce it. For example, “The data are self-reported, so loan amounts may be under-reported. To reduce this, respondents were asked for the sanction letter where available, and 61% produced one.”

Keep only the three to five limits that really matter. If you write one full page of apologies, the examiner starts doubting the whole study.

Implications for theory, practice and policy

Near the end of the discussion chapter, move away from the single findings. Ask what they mean when taken together.

Under theory, say if your results support the model or framework you took. Sometimes the results add something new to it, and sometimes they go against it. Say your conceptual model had six paths and two of them were not supported. That itself is a point to write under theory.

For practice and policy, be specific. A line like “the government should take steps” gives the reader nothing to work with. Compare it with “District Industries Centres could pair SHG linkage with a working-capital product for garment units”, which a reader can act on. In an engineering or science thesis, the practical implication may be a design choice, a material setting or a cost figure.

Some theses keep the implications in the conclusion chapter and not in the discussion. Both are accepted. Only thing is, the same list should not come again in the other chapter. How to split the work between these two chapters is given in our guide on conclusion and future scope.

Viva questions that come from the discussion chapter

Many viva questions are coming from the discussion chapter, because this is the chapter where you made your claims. Be ready for questions of this kind.

• “Your finding is different from the study you have cited. Why do you think so?”
• “You wrote this may be because of X. Did you actually test X?”
• “Which finding of yours is the most important one, and why?”
• “Suppose you start this study again today. What would you do differently?”

That last question is really about limitations. If you wrote the limitations section honestly, the answer is already there in your thesis. In our viva preparation sessions we ask questions like these on your own thesis, so you are not answering them for the first time in front of the board.

Faults examiners point out most often

The results, repeated. The scholar takes the tables from the results chapter and pastes them here again. The only change is “this shows that” in front of each number. Take out those numbers. In their place write what the numbers mean for your study.

No citations. Your result has to be kept next to earlier researches and compared. Without citations in the discussion, the examiner has no way to see that link.

New results. When an analysis appears for the first time in the discussion, the reader gets confused. If you need that analysis, put it in the results chapter.

Claims bigger than the data. A survey done in one district cannot tell what is true for the whole of India. Keep your claims to the level of your sample.

If you are stuck on the discussion chapter, our thesis writing support goes through it with you one finding at a time. The writing of what each finding means stays with you. We read it and ask the questions an examiner is likely to ask.