Abstract
The replication by Landers et al. (1978) of the Martens’ (1969) social-facilitation study is evaluated. The atypical constant error (CE) performance curve obtained in the Martens’ study may diminish the external validity of the CE results, but not the internal validity. Furthermore, the validity of the AE results as well as the highly significant physiological corroborative data remain unscathed. Finally, the conclusions about social facilitation research reached by Landers et al. are questioned in light of their nonsignificant, noncorroborated, and statistically nonpowerful results.