In order to increase reliability, decrease measurement error, and improve the validity of the construct measurement the scale items were pretested (Dillman 2000). Three stages of pre testing were employed. A Q sort method was employed for the first two stages (Li et al. 2009). The first was conducted with a sample of18 academics and the second with a sample of 123 supply chain managers. Through the Q sort method, survey items were refined to measure the constructs of interest. At the end of the second pretest, 17 items were retained to measure MO, 13 items for SCO, and 5 items for FLX.
đang được dịch, vui lòng đợi..