We used Zou et al.’s (1998) three-d

We used Zou et al.’s (1998) three-dimensional operationalization (financial,
strategic, and satisfaction with exports) of international performance. Given the
high correlations between the sub-dimensions (X0.66; averaging 0.68), we treated performance as a second-order construct with the three dimensions as indicators. The second-order international orientation construct consists of five orientations with medium strength inter-correlations (Table I). We view this construct as a
formative scale; hence, we do not report a-reliability for it.
Analysis approach. Given the sample size, the added complexity of other statistical techniques (e.g. structural equations modeling), and the need to assess simultaneous
relationships between dependent and independent variables, we used three multiple regression models with IM influence, second-order international (five) orientations, and international performance as dependent variables. Several steps were undertaken
to assure that the models meet established guidelines. The sample size meets Hair et al.’s (2010) guideline for a minimum ratio of observations/independent variables. Multicollinearity was assessed by examining inter-scale correlations, which were lower than Hair et al.’s (2010) recommendation (0.90). Additionally, all variance
inflation factors (p2), below the acceptable cutoff of 10 (Hair et al., 2010).

Analyses and findings
Model results
Antecedents and consequences of IM influence. The models for antecedents of IM
influence were significant ( pp0.01; Table II). IM-GM coordination was positively
related to all three facets of IM influence (b ¼ 0.20, 0.19, 0.34, respectively, pp0.05) in support of H1. Unexpectedly, IM-GM conflict was related positively with top management respect (b ¼ 0.30, pp0.01). H2 is refuted at least for top management
respect. Finally, international firm experience (control variable) was related positively with perceived influence (b ¼ 0.18, pp0.05).
Regression coefficients for IM – second-order international orientation appear in Table III. The model was significant ( pp0.01). The R2 of the second-order construct was higher than in analyses using isolated orientations except for market orientation (not reported here). Hence, the contribution of the second-order construct to the total R2 exceeds that of its indicators ( Johnson et al., 2011). Influence and top management respect, facets of IM influence, were related positively with second-order orientation (b ¼ 0.17, 0.46, respectively, pp0.05) in support of H3 relating IM influence with the overall orientation. Finally, while international experience was related negatively with overall orientation (b ¼ 0.17, pp0.05), their low correlation (r ¼ 0.08) indicates probable suppression (Cohen and Cohen, 1983).
Orientations, IM influence, and international performance. The last three columns in Table III report the results for models relating IM influence and international orientations with international performance. The models were significant ( pp0.01) and explained 26-37 percent of the variance. When only IM influence was included (fourth column, Table III) all independent variables were related with performance