TY - JOUR

T1 - Does self-control training improve self-control?

T2 - A meta-analysis

AU - Friese, Malte

AU - Frankenbach, J.

AU - Job, V.

AU - Loschelder, David D.

PY - 2017/11/1

Y1 - 2017/11/1

N2 - Self-control is positively associated with a host of beneficial outcomes. Therefore, psychological interventions that reliably improve self-control are of great societal value. A prominent idea suggests that training self-control by repeatedly overriding dominant responses should lead to broad improvements in self-control over time. Here, we conducted a random-effects meta-analysis based on robust variance estimation of the published and unpublished literature on self-control training effects. Results based on 33 studies and 158 effect sizes revealed a small-to-medium effect of g = 0.30, confidence interval (CI95) [0.17, 0.42]. Moderator analyses found that training effects tended to be larger for (a) self-control stamina rather than strength, (b) studies with inactive compared to active control groups, (c) males than females, and (d) when proponents of the strength model of self-control were (co)authors of a study. Bias-correction techniques suggested the presence of small-study effects and/or publication bias and arrived at smaller effect size estimates (range: gcorrected = .13 to .24). The mechanisms underlying the effect are poorly understood. There is not enough evidence to conclude that the repeated control of dominant responses is the critical element driving training effects.

AB - Self-control is positively associated with a host of beneficial outcomes. Therefore, psychological interventions that reliably improve self-control are of great societal value. A prominent idea suggests that training self-control by repeatedly overriding dominant responses should lead to broad improvements in self-control over time. Here, we conducted a random-effects meta-analysis based on robust variance estimation of the published and unpublished literature on self-control training effects. Results based on 33 studies and 158 effect sizes revealed a small-to-medium effect of g = 0.30, confidence interval (CI95) [0.17, 0.42]. Moderator analyses found that training effects tended to be larger for (a) self-control stamina rather than strength, (b) studies with inactive compared to active control groups, (c) males than females, and (d) when proponents of the strength model of self-control were (co)authors of a study. Bias-correction techniques suggested the presence of small-study effects and/or publication bias and arrived at smaller effect size estimates (range: gcorrected = .13 to .24). The mechanisms underlying the effect are poorly understood. There is not enough evidence to conclude that the repeated control of dominant responses is the critical element driving training effects.

KW - Psychology

KW - self-control training

KW - intervention

KW - meta-analysis

KW - publication bias

KW - robust variance estimation

UR - http://www.scopus.com/inward/record.url?scp=85034422554&partnerID=8YFLogxK

U2 - 10.1177/1745691617697076

DO - 10.1177/1745691617697076

M3 - Journal articles

C2 - 28846503

VL - 12

SP - 1077

EP - 1099

JO - Perspectives on Psychological Science

JF - Perspectives on Psychological Science

SN - 1745-6916

IS - 6

ER -