Skip to main content
. 2018 Jan 11;52(2):157–174. doi: 10.1093/abm/kax024

Table 4.

Study characteristics as predictors (moderators) of intervention effects

Behavioral intentions Behavior Anticipated regret Anticipated negative affect Anticipated positive affect
T1 T2 T1 T2 T1 T2
% Female gender −.0019 .0002 .0086* .0009 −.0005 .0019 0.0029 −0.0132
Age .0002 .1317 .0001 −.0006 −.0043 .0063 0.0077 −0.0013
Internet-based .1084 −.144 .0892 −0.0204
Conceptual framework
 Anticipated regret .0339 −.3199 −.0118 .2311 0.6210* −0.5294
 Anticipated affect/emotion −.0399 .0307 −.130 −.0278 −0.558*** 0.5152
 Emotion education .277 .4635 .348 .0842 0.0591
Behavioral context
 Health behaviors (e.g., exercise) −.0989 −.1764 −.1866 −.1347 .2311 0.2862 0.0936
 Cancer screening −.127 .242 −.0255 −.2019 .2311 −0.0936
 Blood/organ donation .0458 −.0945 .0842 −0.1464
Intervention
 Questionnaire as intervention −.0905 −.3149 −.0908 .0833 .0586 −.299 −0.1319 0.468
 “No content” control −.0881 .0953 .1952 −.2019 .2311 −0.2967 0.2802
Length of follow-up (# hours) n/a −.009 n/a .000 n/a .000 n/a n/a

Beta coefficients from univariate meta-regression models. Anticipated negative and positive affect are in relation to the targeted health behavior. “No content” controls refer to control groups for which there was no or minimal contact. Questionnaire as intervention studies were comprised entirely of anticipated regret items.

*p < .05; ***p < .001.