📄 USING PREZI PRESENTATION AS INSTRUCTIONAL MATERIAL IN ENGLISH GRAMMAR CLASSROOM

Summary

The paper asks whether Prezi, a zooming presentation tool that shows one focused object at a time, helps Indonesian EFL university students learn grammar better than board-and-worksheet teaching. The authors frame grammar as a subject learners often find dull and argue that the visual and aesthetic side of instructional design gets little attention. They ran a pre-test/post-test control-group experiment at UIN Ar-Raniry on conditional sentences. One randomly chosen class unit was taught with Prezi and another conventionally. Both took identical tests, and only the Prezi group filled in a perception questionnaire. Analysis in SPSS covered Kolmogorov–Smirnov normality checks, removal of one outlier, homogeneity testing and an independent t-test. The Prezi group improved by about 17 points and the control group by about 12. Students reported strongly positive attitudes. The authors themselves raise novelty or placebo effects and possible nausea from heavy zooming. The groups also started far apart (pre-test 59.7 vs 35.8), which weakens the headline comparison of post-test means.

Key points

• The population was 298 second-semester English Education students (batch 2014) split into 10 units. Unit 10 served as control (n=12) and unit 2 as the experimental group (n=17).
• Each correct answer was worth 5 points. Control mean scores went from 35.8 to 47.5, and experimental means went from 59.7 to 76.7 (79.06 after dropping one outlier).
• The experimental post-test failed the Kolmogorov–Smirnov normality check (p=0.026). One outlier was removed and the test was rerun before the t-test.
• In the questionnaire, 100% agreed or strongly agreed that Prezi raised their interest and helped their learning. About half rated conditionals as hard.
• The authors caution that they cannot tell whether gains came from focus on the content or from the visual appeal itself.

Interesting ideas

• I1. The two groups differed sharply at baseline, so gains rather than raw post-test means are the fairer measure; a pre/post design only helps if the starting gap is actually used.
• I2. The authors openly note that novelty and presentation could explain the result, not the content itself, which points to a confound worth isolating.
• I3. Removing an outlier to pass a normality check changed the headline mean, showing how small-sample cleaning can shift conclusions.
• I4. Prezi's core design choice is to show a single focused object at a time, hiding the rest to cut distraction.
• I5. Conditionals are hard because the tense and mood of the two clauses must be compatible with each other.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

• I1 (learn): Always report improvement relative to each system's own baseline on the same items, and flag comparisons where baselines differ widely.
• I2 (learn): Add ablations that change only the surface form of an input while keeping the meaning fixed, as a negative control against gains from formatting alone.
• I3 (learn): Report evaluation numbers both with and without excluded items, so that repair rules or filtering cannot silently inflate scores.
• I4 (back): Explanations of a parse or meaning graph could zoom from the whole tree into one frame or node at a time instead of dumping the full structure at once.
• I5 (frames): Treat if-clause and main-clause as a linked pair with compatibility checks on tense and mood, so a hypothetical event is not asserted as a fact in the world of the text.