Table 4. Imputed Hierarchical Multiple Linear Regressions predicting MOOC-end sentiment scores, adjusting for baseline sentiment scores (n = 1491)^a.

	MOOC-End Sentiment Scores
	Word Valence Score			Word Arousal Score			Word Dominance Score
Baseline Predictor Variables	Unstandardised B (SE)	B 95%CI	p	Unstandardised B (SE)	B 95%CI	p	Unstandardised B (SE)	B 95%CI	p
Initial Model Variables:
Baseline Sentiment score	.313 (.036)	.241 to .385	< .0005	.193 (.046)	.099 to .287	< .0005	.302 (.042)	.218 to .386	< .0005
Final Model Variables:
Baseline Sentiment score	.317 (.035)	.246 to .389	< .0005	.196 (.047)	.101 to .291	< .0005	.302 (.041)	.220 to .384	< .0005
Age	-.008 (.003)	-.015 to -.002	.015	.002 (.002)	-.002 to .005	.319	-.006 (.003)	-.012 to -.001	.024
Located in Australia (Yes)	-.171 (.138)	-.449 to .107	.233	.005 (.059)	-.113 to .123	.929	-.184 (.117)	-.420 to .052	.124
Health Professional (Yes)	.114 (.106)	-.100 to .329	.290	.011 (.051)	-.091 to .113	.832	.109 (.088)	-.069 to .286	.224
University Education (Yes)	.017 (.098)	-.181 to .214	.865	.080 (.050)	-.021 to .180	.118	-.025 (.081)	-.187 to .138	.761
Model Fit Statistics (Ranges in 20 imputed datasets) a:
Initial R2 range	.094 to .201			.013 to .068			.058 to .125
F range	153.23 to 373.22 (all p < .0005)			19.47 to 108.71 (all p < .0005)			91.11 to 211.27 (all p < .0005)
ΔR2 range	.005 to .028			.001 to .019			.003 to .030
ΔF range	2.17 to 12.26 (p = .000 to 007)			0.50 to 7.31 (p = .000 to .733)			1.20 to 12.92 (p = .000 to .311)
Final Model p range	All < .0005			All < .0005			All < .0005

Notes.

^a Pooled estimates of Model fit are not provided in SPSS. So we have reported the range of model fit statistics obtained in the 20 imputed datasets.