Developing a knowledge management diagnostic tool : a pilot study
Thesis(MBA)--Stellenbosch University, 2001.
The objectives of this research were as follows: • To develop a knowledge management diagnostic tool that can be used to evaluate the status of knowledge management initiatives from both a technology and a people perspective, • To conduct a pilot study using a convenience sample of 35 students from the Afrikaans Modular Masters of Business Administration course at the University of Stellenbosch in order to test this diagnostic tool for measurement reliability and validity, and • To present and discuss the results of the pilot study. A knowledge management diagnostic tool based on Bukowitz and Williams (2000) Knowledge Management Diagnostic was developed. The diagnostic tool consisted of seven section scales, each evaluating a step in Bukowitz and Williams (2000) knowledge management process. Each section scale contained ten items and responses to items were scored using s seven-point Likert scale. Correlation calculations, factor analyses and item analyses were conducted for each scale in order to test for measurement reliability and validity. A comparison of means between respondent group variables was also conducted using the Kruskal-Wallis and Mann-Whitney tests. The Kruskal-Wallis and Mann-Whitney comparison of means revealed no significant differences in total scores on the knowledge management diagnostic tool between groups. However, the fact that a convenience sample rather than a random sample was used, and the sample size was limited (n=35) indicates that these results may not be conclusive. Notwithstanding the small, non-random sample used in this pilot study, the results of the correlation calculations, factor analyses and item analyses indicated that the scales had satisfactory internal consistency reliability and an examination of the items that loaded against each factor indicated that the scales displayed face validity. In the light of these results, recommendations for further research using a larger, random sample were therefore presented. In addition, items were recommended for deletion in order to improve internal consistency reliability