Brockmeier, l., leech, d., pate, j., & Gibson, n. (april 2011).  (Page 4/10)

Demographic information collected on the survey included gender, race or ethnicity, educational level, and school configuration. The number and percentage of principals responding to the survey by gender were 90 (34%) female principals and 171 (66%) male principals. By race or ethnicity, there were 200 (77%) Caucasian principals, 58 (22%) African American principals, and 3 (1%) Hispanic principals. The number and percentage of principals reported having a master’s degree were 47 (18%), whereas the number and percentage of principals reporting having an Educational Specialist’s degree and doctorate were 139 (53%) and 75 (28%), respectively. By school configuration, there were 99 (38%) elementary school principals, 66 (25%) middle school principals, 85 (33%) high school principals, and 11 (4%) other (i.e., combination school) principals.

Instrumentation

Hope, Brockmeier, Lutfi, and Sermon (2007) initially developed The Principal’s High Stakes Testing Survey to obtain information about the impact of high stakes testing on Florida’s principals across six hypothesized domains (i.e., curriculum, teaching, work satisfaction, stress, accountability, and students). Principals’ responses to each of the 48 items within six domains were measured on a five-point Likert scale ranging from 1 ( strongly disagree ) to 5 ( strongly agree ). Items comprising the survey were designed based upon a review of the literature, which presented positive and negative attributes of high stakes testing. Brockmeier, Pate, and Leech (2008) examined the psychometric characteristics (i.e., validity and reliability) of the instrument. Expert panel members suggested only a few very minor modifications to improve the instrument, while the exploratory factor analyses and confirmatory factor analyses yielded data to support the fit of the model and factor invariance of the model by gender and race or ethnicity. Cronbach’s alpha reliability for the 48-item instrument was .92; the subscale Cronbach’s alpha coefficients were .70 for curriculum, .85 for teaching, .73 for work satisfaction, .81 for stress, .84 for accountability, and .63 for students. Reliability was good for the total composite scores on the instrument and was good to adequate for scores on each of the subscales.

In the present study, Cronbach’s alpha reliability coefficient for the 48-item instrument was .90, whereas the subscale Cronbach’s alpha coefficients were .56 for curriculum, .82 for teaching, .74 for work satisfaction, .85 for stress, .84 for accountability, and .53 for students. When compared to the previous administration of the instrument, Cronbach’s alpha reliability coefficients were .14 lower for the curriculum subscale and .1 lower for the students subscale. However, there appeared to be less variation in principals’ responses in the present study than in the previous administration of the instrument and less response variation would reduce the value of the reported reliability coefficients. Reliability was good for the total composite score on the instrument and was good for four of six subscales. Analysis of the total scores on the curriculum subscale and students subscale should proceed with caution due to their respective reliability estimates on this administration of the instrument. Note that negatively worded items were reverse coded for the estimates of reliability and subsequent inferential statistical analyses.