Doctoral Dissertations

Date of Award

8-2000

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Education

Major Professor

Robert L. Williams

Committee Members

R. Steve McCallum, Donald J. Dessart, John Lounsbury

Abstract

The primary purpose of this study was to evaluate the psychometric efficacy of positively versus negatively worded items on a self-report measure of middle schoolers' work ethic. A secondary purpose was to determine the extent to which the three conceptual factors presumed to reflect the objective benefits of work, artificial benefits of work, and one's comfort level with work represent empirical factors. Finally, demographic comparisons were made. Four versions of a middle schoolers' work-ethic scale were developed to allow for both between- and within-subjects comparisons of the differential effectiveness of negatively and positively worded items. Each scale version consisted of 30 items, with each conceptual factor (Objective, Artificial, and Comfort) represented by an equal number of items. The P Scale consisted of 30 positively worded items, the N Scale of 30 negatively worded items matched in content with each of the positively worded items, the NPl Scale of 15 matched item pairs, and the NP2 Scale of the remaining 15 matched item pairs. A total of 225 male and 213 female middle schoolers (grades 6, 7, and 8) from 3 schools in rural Tennessee participated in the study. Each subject responded to 1 of 4 versions of a work-ethic scale. A school psychologist or classroom teacher presented each item on an overhead projector and read it twice before proceeding to the next item. The within- and between-subjects analyses revealed that scale total and subscale scores based on negatively worded items were higher than companion scores based on positively worded items (p < .05). At the individual item level, the means of negative items were consistently higher than the means of their positive counterparts. More than 1/3 of the 120 items across scale versions were answered in a prowork maimer by 80% of students. The between-subjects analysis indicated that twice as many negative items as positive items produced prowork endorsement at the 80% level. In contrast, the within-subjects comparison of item facility produced no significant differences in the extent to which the item types resulted in prowork endorsement. The internal consistencies for scale total scores were good, ranging from .84 to .91. Among the subscales, the Artificial subscale, regardless of item type, consistently achieved the lowest reliability coefficients, ranging from a low of .61 within the P Scale to a high of .78 within the NPl Scale. Between-subjects comparison of internal consistency based on item wording indicated that scale and subscale scores based on negative items consistently achieved higher alpha levels than scores based on positive items. Consistent differences in alpha levels based on item type were not obtained in the within-subjects analysis. The demographic comparison of alphas indicated that males consistently achieved higher alpha levels than females. The internal consistencies of scales and subscales did not vary as a consequence of grade level. Evidence of criterion validity stemmed from the correlations between work-ethic scores and school-related measures. Specifically, the scale total and subscale scores for work ethic were positively correlated to grades and negatively correlated to absenteeism and teacher notations of "needs improvement" in work habits. Between- and within-subjects comparisons of correlations based on item type revealed that scores on the negative items were as likely to be related to grades as scores on positive items. The Artificial subscale for either item type was more frequently and widely related to grades than either the Objective subscale or Comfort subscale, indicating that students who earn high grades may be motivated by extrinsic benefits rather than the inherent value of working hard or the affective satisfaction gained from working hard. Maximum-likelihood factor analysis with the request of a 3-factor structure provided some empirical support for factors addressing the artificial benefits of work and one's comfort level with work, but not for the objective benefits of work. A general work-ethic factor was also indicated. However, the 3-factor model accounted for a relatively low percentage of the variance within the work-ethic scales. Results of this study support previous research that suggests item wording influences subjects' responses. Unlike previous research on work-ethic scales, this study found that negative items tended to evoke stronger prowork responses than positive items. Additionally, results of this study suggest that subjects are not necessarily more confused by negative items than positive items, as internal consistencies for these items were as high or higher than their positive counterparts. Results of factor analysis suggest that factors addressing the artificial benefits of work and one's comfort level with work are appropriate for inclusion in work-ethic scales.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS