Academic English revision

Your Data Analysis Section Reads Like SPSS Output. That's the Problem.

The data analysis section's cardinal sin isn't bad writing — it's reporting what the software did without explaining why you did it. Your reader doesn't need a recipe; they need your reasoning.

Open your data analysis section. Is the first sentence 'A t-test was conducted to examine...'? Congratulations — you've written a software operation log, not an analysis. Your professor isn't tracking your intellectual logic; they're mentally watching you click buttons. Every verb in your paragraph is conducted, performed, obtained, yielded. Not a single word explains why. You've documented actions, not thought.

Here's the structural problem. Your mental model is 'what I did → what I got.' Your professor's mental model is 'why you did it → what it means → how it connects to your research question.' You've skipped the middle — the part that's actually worth points. A data analysis paragraph isn't a receipt for software usage. It's a reasoning chain: why this test and not that one, why controlling for this variable, what the result implies within your theoretical framework. If you only report the what, your reader has to guess the why — and they'll guess that you never thought it through.

❌ 'A multiple regression analysis was conducted to examine the relationship between self-efficacy and academic performance. Self-efficacy was entered as the independent variable and GPA as the dependent variable. The results indicated a significant positive relationship (β = .42, p < .001).' ✅ 'Because self-efficacy was theorized to predict performance independently of prior achievement, a hierarchical regression was used. Prior GPA was entered in Step 1 to establish a baseline; self-efficacy was added in Step 2 to test its incremental contribution. The significant increase in R² (ΔR² = .18, p < .001) suggests that self-efficacy explains variance in performance beyond what prior grades account for — supporting Bandura's claim that belief in one's capabilities operates as a distinct motivational resource.' The first version reads like a menu. The second reads like a judgment.

Second pair. Do you default to 'the results were significant' and move on? That's telling your reader 'I pressed run, the light turned green.' Significance isn't analysis. It's a fact. ❌ 'The ANOVA results showed a significant difference between the two groups (F(1, 148) = 12.34, p < .001). Post-hoc tests confirmed that the treatment group scored significantly higher than the control group (p = .002).' ✅ 'The significant group difference (F(1, 148) = 12.34, p < .001) was larger than the effect size predicted in our pilot (η² = .08 vs. expected .04), which initially surprised us. The post-hoc comparison (p = .002, Cohen's d = 0.56) indicates a moderate, practically meaningful gain — not just a statistically detectable one. This aligns with the dosage-effect pattern reported by Smith (2019), where effects scaled with intervention intensity.' Notice the final sentence: your result is in conversation with existing literature. That's what analysis looks like.

Third pair. Is your paragraph structured as 'First... Second... Third... Fourth...'? That's a lab report structure, not an analysis structure. An analysis paragraph works like detective reasoning: you find a clue, it points somewhere, it combines with other clues. ❌ 'First, descriptive statistics were computed. Second, normality was assessed. Third, a paired-samples t-test was performed. The results indicated a significant difference (t(59) = 4.21, p < .001). Fourth, a Cohen's d effect size was calculated (d = 0.54).' ✅ 'Normality was not a concern (Shapiro-Wilk p = .23), so a paired t-test was appropriate. The pre-post difference (t(59) = 4.21, p < .001, d = 0.54) confirms a meaningful shift. What's notable is that the mean gain (M = 3.12) exceeded the maximum score on pilot practice items, suggesting genuine learning rather than test-familiarity effects — a distinction that matters for generalizability.' That last sentence is the soul of the paragraph. Everything before it is setup.

Here's the test. Delete every statistical value from your paragraph. Does the remaining text still stand on its own? If the answer is no, your paragraph is a slave to the numbers, not their master. In a strong analysis section, statistics are evidence and reasoning is the skeleton — remove the numbers and the skeleton is still intact. Your professor isn't auditing your SPSS workflow. They're auditing your brain.

Practical move: after your first draft, cross out every method-description clause and keep only the 'why' and 'what it means' sentences. If you cross them out and the paragraph evaporates, you wrote a procedure log, not an analysis. I run my drafts through easydue before submission — it flags the conducted/performed/obtained mechanical tone and shows me exactly which sentences need to expand into reasoning. The time I save goes into sharpening my argument, which is worth a hundred times more than polishing phrasing. Your reader doesn't care which software you used. They care whether you actually thought.