Your Chapter 4 has the tables. The DASS-21 means are in, the correlation matrix is in, the t-test came back significant, and the adviser’s comment on the draft reads “interpret, do not just describe”. The defense is in nine days. This is the moment a BS Psychology thesis stalls, because the tables were the part the group knew how to do and the discussion — the paragraphs that say what a correlation of .32 between perceived stress and academic burnout means for the theory in Chapter 2 — is the part nobody taught.
The cost of leaving it as description is specific. A psychology panel, which usually seats at least one registered psychometrician, grades the discussion as the evidence that the group understood its own instrument and its own theory. A results chapter with no discussion is a revise-and-resubmit, and a revise-and-resubmit two weeks before the end of the semester is a graduation moved to the next one. The fix below takes one evening per sub-problem.
Need the discussion drafted from your tables tonight?
Give Tesify your Chapter 4 tables, your sub-problems and the theory you named in Chapter 2, and it drafts the interpretation paragraphs in the results-then-discussion structure a Philippine psychology panel expects, with the statistics reported in APA style. Over 9,000 students have used it to write more than 15,000 chapters, and every word is still written by you.
What a psychology panel reads the discussion for
Four things, in the order they will ask about them. Whether each scale score was interpreted against the scale’s own norms and cut-offs rather than against a home-made weighted-mean range. Whether each inferential result was reported with its effect size, not only its p-value. Whether the finding was connected back to the theory and the related studies in Chapter 2, agreeing or disagreeing with them by name. And whether the group said what the finding cannot show — the design limits, the sample limits, the instrument limits — before the panel says it for them.
The generic anatomy of a results chapter, table by table, is in the guide to writing Chapter 4 results and discussion. What follows is the psychology-specific layer that guide does not carry.
Fix 1: interpret scale scores against the scale, not against a weighted-mean range
The single most common fault in a Philippine psychology Chapter 4 is a table that reports a DASS-21 subscale as a weighted mean of 2.41 with the verbal interpretation “moderate”, taken from a five-band range the research manual supplies for researcher-made questionnaires. The DASS-21 is not scored that way. Its subscales are summed, doubled, and read against the severity categories in the manual — normal, mild, moderate, severe, extremely severe — and the Chapter 4 sentence reports the proportion of respondents in each category and the mean subscale score. The same is true of the Perceived Stress Scale, the PHQ-9, the GAD-7 and the Rosenberg Self-Esteem Scale, each with its own published scoring and cut-offs.
The rewrite is mechanical once you have the scoring rule: re-score, re-tabulate as category counts and means, and rewrite the sentence as “Forty-one per cent of respondents scored in the moderate to extremely severe range for stress, with a mean stress subscale score of 18.6 (SD = 8.2)”. Where the scoring rule and the permission terms for each scale come from, and how the instrument section of Chapter 3 should already have stated them, is set out in the guide to choosing and adapting a standardized scale for a psychology thesis. If Chapter 3 did not state the rule, fix Chapter 3 in the same evening; a panel checks that the two chapters agree.
Two other scale-reading errors travel with this one. Reverse-scored items that were never reversed before summing, which flattens every subscale toward the middle and is visible to a psychometrician from the table alone; and a translated or adapted scale reported against the original norms, which the discussion has to flag as a limitation because the norms were not established on the adapted version. Check both before rewriting a single sentence.
Fix 2: report the effect size beside every p-value
Psychology panels expect the APA reporting convention: the statistic, the degrees of freedom, the p-value, and an effect size with a confidence interval where the software gives one. For a correlation the coefficient is the effect size, and the sentence reads “perceived stress was moderately and positively correlated with academic burnout, r(148) = .32, p < .001, 95% CI [.17, .46]”. For a t-test report Cohen’s d; for an ANOVA report eta squared or partial eta squared; for a chi-square report Cramér’s V or phi. Then interpret the size in words — small, medium, large — using the conventions Cohen proposed, and say so.
The reason this matters to the discussion, not only to the table, is that a significant result with a small effect and a non-significant result with a medium effect in a small sample are two different findings, and the discussion has to say which one you have. A group that reports “there is a significant relationship” and stops has not interpreted anything.
Fix 3: discuss against the theory and the studies by name
The discussion paragraph for each sub-problem has a fixed shape. State the finding in one sentence. Say whether it agrees with the theory named in Chapter 2 and how — if the framework was Lazarus and Folkman’s transactional model of stress, a correlation between appraisal and burnout is the prediction the model makes, and the paragraph should say that in those words. Then name two or three related studies from Chapter 2 and say whether they found the same direction and a comparable size, in the same population or a different one. Then offer one explanation for any difference that is specific to your respondents: a Philippine private university sample at the end of a semester, a sample that is 78 per cent women, a scale administered online rather than in a classroom.
The verbs matter, because a psychology panel reads hedging as a sign of understanding. Findings are consistent with a theory; they do not prove it. A correlation suggests; it does not establish causation. A difference between groups was observed; it was not caused by group membership unless the design randomised it.
Fix 4: say what the finding cannot show before the panel does
Three limitations recur in undergraduate psychology theses and each has a one-sentence discussion form. The design is cross-sectional, so the direction of the relationship between stress and burnout cannot be determined from these data. The scales are self-report, so the scores reflect what respondents were willing to disclose in the setting where the questionnaire was administered. The sample is a convenience sample from one institution, so the findings describe that population and generalise only with caution. Writing those sentences into the discussion, attached to the specific finding each one qualifies, converts a weakness the panel would raise into evidence of competence.
If a test switched from parametric to non-parametric because a scale score was skewed — Spearman instead of Pearson, Mann-Whitney instead of the t-test — say so in the discussion as well as in the results, and say why; the wording for that switch is in the answer on what to do when thesis data is not normally distributed.
A filled-in discussion paragraph
For a sub-problem on the relationship between perceived stress and academic burnout among third-year students:
Perceived stress was moderately and positively correlated with academic burnout, r(148) = .32, p < .001, 95% CI [.17, .46], a medium effect by Cohen’s conventions. The direction is consistent with the transactional model of stress, in which sustained appraisal of academic demands as exceeding one’s resources is expected to produce exhaustion and cynicism over time. The size is comparable to the coefficients reported in two of the Philippine university samples reviewed in Chapter 2 and smaller than the coefficient reported in a sample of nursing students during clinical placement, a difference plausibly attributable to the higher demand context of that sample. Because the present design is cross-sectional, the data cannot establish that stress precedes burnout rather than the reverse, and because both scales are self-report, the association may be inflated by shared method variance.
That is one paragraph, five sentences, and it answers the four things the panel reads for. Multiply it by the number of sub-problems and Chapter 4 is finished. Chapter 5 then summarises it and draws the conclusions and recommendations, in the structure described in the guide to writing Chapter 5 of a thesis.
How Tesify gets you from the tables to the paragraphs
Give it the sub-problem, the table, the scale and the theory, and it drafts the five-sentence paragraph in the shape above with the statistic reported in APA style; you then replace its placeholder for the related studies with the authors from your own Chapter 2, check every number against your output, and revise the explanation sentence to say what only your group knows about the respondents. The draft is a scaffold for your interpretation, not a substitute for it, and the numbers, the theory and the judgement remain yours. If the adviser has already sent the whole chapter back with a rewrite note, the triage plan for that situation — what to fix first when Chapter 4 is due on Monday — is in the guide for groups whose adviser wants Chapter 4 rewritten by Monday; this guide is the psychology-specific layer on top of it.
Write the discussion tonight, defend it next week
Upload your tables, name your scales and your theory, and draft the interpretation paragraphs for every sub-problem in the wording a Philippine psychology panel expects. Revise with your adviser tomorrow.
Start your psychology thesis on Tesify · Tighten the discussion with the AI editor
Frequently asked questions
What is the difference between results and discussion in a psychology thesis?
Results report what the data showed — the scale scores, the statistics, the effect sizes — without interpretation. Discussion says what the results mean: whether they agree with the theory and the related studies, why they might differ, and what the design cannot show. Most Philippine manuals combine both in Chapter 4, table by table.
Can I use a weighted mean to interpret a standardized psychological scale?
No. Published scales such as the DASS-21, PSS-10, PHQ-9 and RSES have their own scoring rules and cut-offs. Score them as the authors instruct and report the proportion of respondents in each published category with the mean score.
Do I need to report effect sizes in an undergraduate thesis?
Yes. Psychology panels expect the APA convention of statistic, degrees of freedom, p-value and effect size. The correlation coefficient is its own effect size; report Cohen’s d for t-tests, eta squared for ANOVA and Cramér’s V for chi-square, and interpret the size in words.
How long should the discussion be for each sub-problem?
One paragraph of four to six sentences: the finding, its agreement with the theory, its agreement with two or three related studies, one explanation specific to your respondents, and one limitation. Longer discussions usually repeat the table.
How much does Tesify cost?
Current plans and prices in pesos are shown on the Tesify site at sign-up, and a thesis group can share one account for one manuscript. The draft and every revision remain yours to edit and own.
Is it allowed to use an AI tool to draft the discussion?
Policies are set by each school’s research manual or academic integrity policy, and many programs now ask for a declaration of AI assistance. Tesify drafts from your own tables, scales and theory, every sentence is reviewed and revised by you, and the interpretation is your judgement; declare the assistance where your school requires it.
Is respondent data safe on Tesify?
Your manuscript and data are stored on your account and are not shared with other users. Anonymise scale responses before uploading, which is also what your ethics review and your distress protocol require.