Predicting Student On-Task Behaviour Using Multiple Regression Analysis

Laboratory Assignment: Multiple Regression

1. Hypotheses

Null hypothesis (H₀):
Daily minutes of moderate-to-vigorous physical activity (MVPA), prior-night sleep duration, and perceived teaching climate do not jointly or individually predict students’ on-task behaviour.

Alternative hypothesis (H₁):
Daily minutes of moderate-to-vigorous physical activity (MVPA), prior-night sleep duration, and perceived teaching climate jointly and/or individually significantly predict students’ on-task behaviour.

2. Assessment of Multiple Regression Assumptions

The assumptions for multiple regression were assessed as follows:

• Linearity: Scatterplots of each predictor (MVPA, Sleep, Teaching Climate) against on-task behaviour indicated approximately linear relationships.
• Independence of errors: The Durbin–Watson statistic was close to 2.0, suggesting residuals were independent.
• Homoscedasticity: The standardized residuals versus predicted values plot displayed a random scatter, confirming constant variance.
• Normality of residuals: The P–P plot of standardized residuals approximated a diagonal line, and the histogram showed an approximately normal distribution.
• Multicollinearity: Tolerance values were above 0.20 and VIFs were below 5.0, confirming no severe multicollinearity.

Conclusion: The data met all major regression assumptions; hence, the multiple regression results are considered valid and reliable.

3. Multiple Regression Results

The multiple linear was carried out to test the hypothesis that MVPA, duration of sleep, and teaching climate could predict on-task behaviour of learners in the early primary.

The total model had a significant value, F(3,146) = 686.12, p <.001 meaning that the total proportion of predictors together (as a block) significantly explained the variance in on-task behaviour (R2 =.934,Adjusted R2 =.932).

One by one all predictors made a significant contribution:

• MVPA: B = 0.261, SE = 0.012, t = 21.03, p < .001, 95% CI [0.237, 0.286]

• Sleep: B = 4.068, SE = 0.153, t = 26.61, p < .001, 95% CI [3.766, 4.370]

• Teaching Climate: B = 7.124, SE = 0.226, t = 31.49, p < .001, 95% CI [6.677, 7.571]

These results indicate that greater levels of physical activity, increased length of sleep and more favourable teaching climate in relation to improved on-task behaviour. The strongest predictor was sleep duration, after which there was teaching climate and MVPA.

Table 1: Multiple Regression Results

Table 1Multiple Regression Results

Model Summary: F(3,146) = 686.12, p < .001, R² = .934, Adjusted R² = .932

Interpretation

The multiple regression analysis also showed that a combination of physical activity, the duration of past-night sleep, and teaching climate explained 93.4% of variance in the on-task behaviour among children. It proves that the classroom setting factors (both lifestyle and physical activity) play an important role in determining the attentiveness and engagement in children. The strongest predictor turned out to be teaching climate, then sleep duration and MVPA, which showed the importance of supportive school climate and the adequate rest and physical activity.

MVPA, Sleep and Teaching Regression Coefficients of Teaching Bar chart.

Figure 1Bar chart for regression