<aside> 🙂

I use a grading bot for stats, see the output below. It shows me how the students responses differ from my solutions manual. To do this, I have a single copilot agent lite that I provide with the solutions manual for that assignment.

</aside>

You can see that could save me time, but I am still assigning points myself.

✅ Question 1: Gestation and Birth Weight

• Student Answer: r = 0.433, p = 0.0001 (called “positive and linear”).
• Solutions Manual: r(78) = 0.43, p < .001 (moderate positive correlation).
• Difference: Matches solution except missing df and interpretation should use “moderate” instead of “slight bell-shaped curve” prediction.
• What to do better: Include df and interpret effect size correctly (moderate correlation).

✅ Question 2: Smoking and Birth Weight

• Student Answer: t = 3.526, df = 36.6, p = 0.0012 (Welch’s t-test).
• Solutions Manual: Welch’s t(36.58) = 3.53, p = .001, Cohen’s d ≈ 0.90.
• Difference: Matches solution except missing effect size.
• What to do better: Report Cohen’s d for significant results.

✅ Question 3: Age Groups and Birth Weight

• Student Answer: F = 0.903, p = 0.4097 (ANOVA).
• Solutions Manual: F(2,77) = 0.90, p = .410 (NOT significant).
• Difference: Matches solution except missing df and effect size (η²).
• What to do better: Include df and state clearly “no significant difference.”

✅ Question 4: Predicting Birth Weight from Gestation

• Student Answer: slope = 0.58, R² = 0.188, p = 0.0001.
• Solutions Manual: slope = 0.58, t(78) = 4.24, p < .001, R² = .19.