by guest on October 13, 2015ilr.sagepub.comDownloaded from
HUMAN RESOURCE MANAGEMENT PRACTICES IN THE MNC599The sizeof the company was measured in terms of the size of the worldwide company and of the subsidiary, each assessed in terms of numbers of employees and categorized to reflect the fact that several respondents could estimate size only in ranges.A firm’s HR system is important in its use of practices such as organiza-tional learning (McDonnell et al. 2010; Tregaskis et al. 2010). We include two indicators of the extent of global HR structures and policies. We expect positive effects on the use of HRM practices if choice is important.International HR committeemeasured whether the MNC had a formal commit-tee of senior managers responsible for developing HR policies across countries. This was dummy coded with 1 indicating presence of such a committee. We also had other measures such as whether the worldwide company used a human re-source information system. We examined these, but their results were similar to those for the most straightforward measure of the presence of a committee.HR philosophymeasured whether the MNC had a worldwide philosophy on the management of its workforce that governed all global operations. This was mea-sured on a 5-point scale of agreement as to how far a philosophy existed.Finally, we included three other control variables. The sector of operationof a firm was dichotomized as production and service firms. In so far as prac-tices such as teamwork and associated ideas such as lean production were strongly associated with manufacturing, we would expect take-up to be greater in the production sector. Trade union recognitionmeasured whether the subsidiary recognized trade unions. There has been considerable de-bate on whether unions promote or retard use of HRM practices. It may also be that the effects are greater in relation to the LOG than to managers.Workforce skill levelmeasured whether the largest occupations group of the subsidiary was one of technical/professional, sales staff, clerical/administrative staff, or manual operatives.Mode of Analysis and Data DescriptionWe analyzed the data in two ways. First, we classified each of the individual measures listed in Table 1 in terms of country of operation and the contrast by country of origin between U.S.-owned firms, firms from other LMEs, and all other (CME) firms. The results from this descriptive analysis (available on request) suggested some patterns that were interrogated further. All the descriptive statistics are based on weighted data for each country.Second, we aggregated the measures under each of the concepts of moti-vation, opportunity, and control to create six summary indices (that is, three for managers and three for the LOG). These indices count the number of practices under each heading. The results were consistent with the exami-nation of the individual measures. How did we account for missing data?