This type of analysis is logical for studies looking at entry mode choice; entry mode choice can easily be coded as a dichotomous variable. Other statistical methods used were multinomial logit (in 20 studies), discriminant analysis (7), PROBIT (7), multiple regression (6), and exponential estimation (5). Most of these types of analyses test group membership and are also logical choices for scholars performing entry mode research. ANOVA (4), TOBIT (4), t-tests (4), structural equation modeling/LISREL (2), events analysis (1), and univariate analysis (1) were also used to test entry mode statistical models. A summary of the statistical methods used in entry mode research is presented in Table 6.
Dependent variables: The dependent variable (DV) most commonly used (in 43 studies) was a dichotomous variable classifying full and shared ownership (e.g., wholly owned subsidiary versus joint venture). The second most commonly used dependent variable was equity levels (19). There were also other less commonly used DVs with categorical classifications such as wholly owned versus independent (10), acquisition versus joint venture (5), export versus foreign direct investment (5), majority versus minority owned (4), greenfield versus joint venture (3), minority joint venture versus equity joint venture versus majority joint venture versus wholly owned subsidiary (3), contract versus equity joint venture (2), and management service contract versus franchise (2). Furthermore, from these findings we discovered that most entry mode research used dichotomous variables as DVs when comparing modes chosen by companies entering a foreign market. A list of these dependent variables is presented in Table 7. PAGE 276.
Independent variables: In terms of independent variables (IVs), we found that there were around 200 different variables used in various entry mode studies. Thus, we present the top 10 of these IVs in Table 7.
