Fall 2025 PEERS Goal Analysis

November Goal

On November 4, 2025, I submitted the following:

I will evaluate the effectiveness of transitioning from in-class, timed, weekly quizzes to at-home checks-for-understanding by conducting correlation analysis between students’ formal assessment scores (ie, Midterms & Final) and their check-for-understanding performance to determine the impact on content recall and mastery.
I’ll use Pearson’s correlation coefficient to see the relationship between students’ self-check scores and their score on the Final Exam

Classes had started in mid-August, and for the first two chapters I had gone with the plan that I developed over the summer, in which students would complete an in-class quick-write in Lockdown Browser, based on the assigned chapter. This eliminated the chance of students using generative AI, and maximized the probability that their score on these formative assessments were an accurate measure of what students know and are able to do. But based on… students’ feedback… after the first two chapters and in-class quick-writes, I changed to the at-home self-checks.

December Analysis

Regular coursework has largely concluded at the point of this analysis (Dec 10, 2025). Few students will be taking the Final exam next week seriously, and so the analysis here will compare students’ Chapter-based self-check quiz scores against their two Midterm exam scores. 

There are two things to keep in mind regarding this analysis. First, our Social Science courses are offered in both Honors and non-honors format. All instruction and in-class activities are undifferentiated; the differentiation is in the formative and summative assessments. Second, the formative assessment “self-checks” allow unlimited attempts. Generally speaking students score the maximum number of points after four- to six-attempts on each of the eight chapter-based self-checks. Though the self-checks are graded on accuracy because of the unlimited attempts allowed, this is functionally a completion-based assignment. This, for the analysis below, rather than comparing self-check scores to midterm scores, I will instead compare by completion-rate.

The research question thereby becomes: “Does consistent engagement with low-stakes practice predict performance on high-stakes exams?”

For each group (Honors/ non-honors), the following was calculated:

• Mean midterm scores for “completers” vs. “non-completers”
• Sample sizes and standard deviations
• An independent-samples t-test (Welch’s, which doesn’t assume equal variances)
• Cohen’s d effect size, which tells you the practical magnitude of the difference

Behavioral-Neuroscience

A Cohen’s d of ~0.4 means that completing the self-checks is associated with roughly a 0.4 standard deviation improvement in midterm scores. In practical terms, if the midterm SD is around 15 points, completers scored about 5-6 points higher on average (which translates to a course grade 5-6% higher). At half-a-letter grade for the course, that’s noticeable but not transformative.

This doesn’t mean self-checks are ineffective or that I should abandon them and return to the original in-class quick-writes. A few important caveats:

1. Correlation ≠ causation. Conscientious students both complete self-checks and study more effectively elsewhere. The self-checks might just be a proxy for general engagement rather than the cause of better performance.
2. The unlimited attempts issue. Students can click through repeatedly until they hit 100%. This decouples the self-check score from actual learning—a student who struggles, reviews, and retakes thoughtfully looks identical in your gradebook to one who just guesses repeatedly until they pass.
3. Ceiling compression. Most of my “completers” cluster near 100%, which limits the ability to detect a dose-response relationship between how much someone learned from self-checks and their exam scores.

NeuroPsych TL;DR

Students who consistently completed the self-checks scored higher on midterms than those who didn’t — and this held for both Honors and Regular sections. The relationship is statistically significant and in the expected direction.

Social-Cognitive Psychology

Something different and interesting is happening with the CogPsych students; they seem to be using the self-checks differently. 

• Honors (d = 0.89): A large effect. For these students, even moderate engagement with self-checks is strongly associated with exam performance. This is the clearest evidence that the self-checks “matter.”
• Regular (d = 0.05): Essentially no effect. Completion vs. non-completion at the 60% threshold tells me almost nothing about how Regular students performed on midterms.

Honors students may be engaging in genuine retrieval practice: reviewing mistakes, spacing attempts, treating it as study. Regular students may be more likely to click through for completion credit without the reflective component that drives learning.

CogPsych TL;DR

Self-check completion predicts midterm performance for Honors students (d = 0.89, a large effect) but not for Regular students (d = 0.05, essentially zero effect). Unlike in Behavioral Neuroscience, here the value of self-checks appears to depend entirely on how students engage with them, not just whether they complete them. The same tool, used differently, produces dramatically different outcomes.