Comment on Zajonc's Social Facilitation Review Including a Reanalysis of Travis' (1925) Data

Abstract

Zajonc's review and theoretical integration of the social facilitation literature (1965) has stimulated a great deal of research and thought (e.g., Weiss 8: Miller, 1971). purpose of the present note is to comment o n three of the six studies to which he refers in the section of his review entitled Audience Effects and to present a reanalysis of the data from one of these studies. his review, Zajonc gives prominent consideration to Travis' (1925) smdy of pursuit-rotor performance alone and in the presence of an audience. Zajonc concludes, found a clear improvement in the performance when his subjects were confronted with an (p. 269) . Regarding this comparison, Travis reports, Statistically the P. E , of the differences is too large for these differences to mean much (p. 145) . Travls made two comparisons of means for his sample of 22 Ss, ( 1 ) the difference between the averages of S's 1 0 scores in the presence of an audience and of his highest 1 0 consecutive scores alone and ( 2 ) the difference between S's single best score in the presence of an audience and his single best score alone. Reanalysis of Travis' data (the data are reported in his article) by the writer indicates that the first difference is significant ( D = 4.66, J = 2.88, 21 d f , p < .01) as is the second (D = 4.18, t = 2.54, 21 d f , p < .02). Thus Zajonc is correct in his conclusion regarding Travis' in spite of Travis' disclaimer. Where did Travis go wrong? His method of analysis was that appropriate for independent rather than related measures. Concerning two studies, Zajonc (1965) indicates, In 1931 Husband found that the presence of spectators interferes with the learning of a finger maze, and in 1933 Pessin and Husband confirmed Husband's results (p. 270) . However, regarding this comparison Pessin and Husband (1933) concluded, Social stimuli such as are presented by silent observers (one or two) have no significant effect upon the efficiency of learning. N o statistically reliable differences were found between the stimulated and the control groups. fact, the differences do nor approach significance. This study involved between-subject rather than within-subject ,comparisons so there is no reason to suggest that improper statistics were used. Husband's (1931) study did not include the influence of an audience as a variable and no mention of the influence of spectators was made in the article. Zajonc was apparently misled by the statement in Pessin and Husband (1933) that The other writer ( H ) had previously noticed in the course of an extensive series of studies that another person present often acted as a disturbing element and interfered with the close concentration necessary to learn a fairly difficult maze (p. 149) . This statement is an interesting observation on Husband's earlier research but is hardly a finding as implied in Zajonc's review.