Table 3.

CFA fit indices for the original and modified ABCD Questionnaire measurement models

Original 18-item ABCD In the original study of 2017, 18 items were entered into factor analysis. This CFA tests the fit of these original items to their structure using the larger Nottingham SPICES dataset.
CMIN	P value	CMIN (Chi Square Minimum)/DF (Degrees of Freedom)	TLI	CFI	RMSEA	RMR
714.941	0.000	5.416	0.826	0.850	0.097	0.049
Original 18-item ABCD with 5 smoking items added In the original study of 2017, items relating to smoking behaviours were developed but could not be included in the published scale due to insufficient data. In the Nottingham SPICES Study, sufficient observations were made to test these smoking items.
CMIN	P value	CMIN/DF	TLI	CFI	RMSEA	RMR
994.931	0.000	4.442	0.865	0.881	0.086	0.049
Edited 20-item ABCD with smoking subscale As discussed above, independent item analysis and exploratory factor analysis using the independent SPICES Nottingham dataset revealed issues with the continued inclusion of some of the original ‘perception of risk’ subscale items, and the allocation of an item relating to dietary behaviours in the physical activity behaviours subscale. The published ABCD Questionnaire was edited to remove or reassign the problematic items and retested using CFA.
CMIN	P value	CMIN/DF	TLI	CFI	RMSEA	RMR
638.973	0.000	3.896	0.881	0.897	0.079	0.052
Modified 20-item ABCD with smoking subscale The measurement model created for the CFA was modified so that items within each ABCD subscale were set to covary with one another.
CMIN	P value	CMIN/DF	TLI	CFI	RMSEA	RMR
385.312	0.000	2.439	0.941	0.951	0.056	0.046

ABCD, Attitudes and Beliefs about Cardiovascular Disease; CFA, confirmatory factor analysis; CFI, Comparative Fit Index; RMSEA, root mean square error of approximation; TLI, Tucker-Lewis Index.