VALIDITY OF THE MULLEN DIAGNOSTIC SURVEY FOR LEADERSHIP IMPROVEMENT

Abstract

Based upon Likert's theories (2, 3 ) , Mullen ( 4 ) in a study using a national sample of school systems developed the Diagnostic Survey for Leadership Improvement. The survey measures satisfaction perceptions of subordinate role participants as they relate to the performance of school leadership personnel in selected leadership processes. Factor analysis has confirmed the construct validity of the survey as measuring five processes corresponding to Liken's organization model: decision-making, communica. rion, control, interaction-influence, and confidence 'and trust in leadership. Reliability and validity of the 52-item full scale survey has been established by a variety of procedures ( 4 ) . A short form of the two most discriminating items from each of the five scale factors has .99 reliability with the 52-item form. On Form 04, teachers in each school indicate their perception of the leadership performance of their principal in the form of an and should answer from which an intensity score is derived by subtracting the squared response from the squared should response. lntensity scores range from 0 to 16. As scores move toward 0, satisfaction increases. Combining the 10 items gives an intensity score for total satisfaction with leadership. Variance in intensity between schools (principals) would be significantly greater than that within schools (perceptions of one person). The analysis was part of a larger study of perceptions of the principal's leadership by black and white teachers in large and small high schools in which the principal was either black or white ( 1 ) . In 24 of 48 high schools the principal was white. Half the schools were classified as lacge and half as small based on enrollment and number of teachers. Schools with black principals were identified and matched by size and geographic location with those having white principals. White teachers (N = 349) responded to black principals' leadership; 368 black teachers to black principals; 202 black teachers to white principals and 630 white teachers to white principals. The unit of analysis was the mean intensity score for the black teachers and the white teachers for each of the 48 principals; thus 96 scores were generated, two for each principal. The data were analyzed by a split-plot factorial analysis of variance which gave no significant differences in satisfaction for race of principal, size of school, or for any interaction. A significant difference (PI.., = 18.41, p < .01) was found for race of teacher. Mean for white teachers was 5.56, for black teachers, 4.26, indicating less satisfaction among white teachers. Significantly greater variability in a random-model analysis of variance was also obtained for satisfaction between schools ( 5 = 4.06) as compared to satisfaction within schools (s' = 2.17; F.4.u = 1.86, p < .05). This greater variability between principals provides evidence for the survey's validity.