In this article, we review research on group cohesion in military units, us
ing meta-analytic techniques that convert research findings to a common met
ric. After combining some of the 39 samples to eliminate possible overlap i
n participants, we conducted separate analyses for 7 different outcomes. Ef
fect sizes (correlation coefficients), weighted by the number of participan
ts, were .40 for cohesion and group performance,.20 for cohesion and indivi
dual performance, and .47 for cohesion and job/military satisfaction. Cohes
ion was also positively related to measures of retention (.22), well-being
(.24), and readiness (.30) and was inversely related to rates of indiscipli
ne (-.14), although these latter results are not definitive because of the
smaller number of samples involved. Effect sizes weighted by the number of
groups generally tended to decrease the number of cases, increase the effec
t sizes, and broaden the confidence intervals. We conclude that group cohes
ion results in desirable outcomes for the military and that the research ha
s implications for policy and training issues. We recommend that future coh
esion researchers include all basic data in their research reports, delinea
te their concepts explicitly, use valid and reliable measures, and give car
eful consideration to the level of analysis and other moderator variables.