Research constructs are operationalized based on pretest, related literature reviews and various theories. For the
questionnaire, the multiple-item method was used and each item was based on 6 point Likert scale from ‘very low’ to
‘very high’. A multiple regression method is used to test the data.
The unit of analysis in this study is organization. For this study, we selected 900 organizations in Korea as a
convenience sample. At our request, these organizations identified the representatives in charge of the firm's KM project
or similar project. We focused on the middle managers that were emphasized in Nonaka and Takeuchi [31]. The mail
survey was performed in parallel with visits. We collected 227 questionnaires, but fourteen were not appropriate for our
study. In sum, 203 valid responses were used.
We performed reliability test for internal consistency. Items with cronbach's alpha values less than 0.6 were dropped
for further analysis. Since each variable was measured by the multi-item constructs, factor analysis with varimax was
conducted to check the unidimensionality among the items. Analyses were performed on the 21 items that measured the
components of KM processes, on the 43 items that measured the components of KM enablers, and on the 6 items for
nonfinancial performance and return on investment (ROI) for financial performance. Items with factor loading values
lower than 0.5 were abandoned for further analysis. Among them, 2 items related to KM process and enablers had
loading value less than 0.5 or had item-to total correlation less than 0.4 were dropped. There was one item with loading
value lower than 0.5 for organizational performance.