臺大管理論叢 NTU Management Review VOL.29 NO.1

Post-Disaster Grain Supply Chain Management with Supplier Hoarding and Regime Intervention 34 H2a. Regime intervention in grain supply markets strengthens the relationship between a grain supplier’s attitude toward weakeedned operational shock-wave and the supplier hoarding intention after a disaster. 4. Research Method To test the hypotheses of this study we used the following steps: (1) content validity; (2) data collection; (3) preliminary test; and (4) result. 4.1 Content Validity All the survey measurement items were established by related literature which had been reviewed previously, such as, social cognitive theory, theory of planned behavior (TPB), and applied psychology. To make these items appropriate for the purpose of this study, we summed up the context of the questionnaire in Appendix A. For the measurement of these items with a seven-point Likert-type scale, the range is from (1) “Strongly disagree” to (7) “Strongly agree”. Next, we test the appositeness of contexts from the questionnaire survey to establish content validity. In accordance to the methods of Rubio, Berg-Weger, Tebb, Lee, and Rauch (2003), we used logical analysis to verify these contents. A pre-test was implemented by a specialist panel which consisted of five professionals in related areas, who were supposed to express their opinions using the seven-point Likert-scale. We determined the content validity of the questionnaire by computing a variety of descriptive indices of inter-rater agreement (Haynes, Richard, and Kubany, 1995). All 16 items had acceptable inter-rater reliability in which inter-rater agreement index values were bigger than the recommended threshold of 0.8 (Rubio et al., 2003). Then, we examined the extent of agreement by content specialists to see if the contents are suitable for a relatively homogeneous collection. All 16 items had excellent content validity where the average content validity index was 0.83, also bigger than 0.8. Finally, we calculated the factorial validity index to assess whether the manifest variables were equivalent to the questionnaire items. All the item results show an acceptable factorial validity index value of 0.85, beating 0.8. According to the analytical results, we can apply the questionnaire items to this study and have excellent content validity.