دانلود رایگان مقاله اغراق و درک نتایج تعامل در تحقیقات کسب و کار بین المللی

ABSTRACT

Approximately one-third of international business (IB) articles include conditional hypotheses, yet the vast majority risk errors in testing or interpreting the results. Scholars typically restrict their empirical analysis to the coefficient of the interaction term in the regression, exposing themselves to the hazard of overstating or understating results. To mitigate the risk of misstating, we advocate that IB scholars also evaluate the statistical significance of the marginal effect of the primary independent variable over the range of values of the moderating variable. We demonstrate that overstating results can occur when the interaction term coefficient is statistically significant but the marginal effect is not significantly different from zero for some value(s) of the moderating variable. Understating can occur when the interaction term coefficient is not statistically significant, but the marginal effect is statistically different from zero for some value(s) of the moderating variable. In this article, we describe, using simulated data, these two possibilities associated with testing conditional hypotheses, and offer practical guidance for IB scholars.

5. Discussion

 recommendations In this article, we aim to contribute to a broad, constructive conversation on conditional hypotheses in the IB literature and, specifically, how such models are commonly tested using regression models containing multiplicative interaction terms. We have analyzed the most common issues in testing and interpreting two-way interaction models. While we do not explicitly address issues related to nonlinearity (e.g. Zelner, 2009), nested or multi-level data (e.g. Hox, 2002; Kreft & de Leeuw, 1998; Luke, 2004; Peterson, Arregle, & Martin, 2012), threeway interactions (e.g. Andersson, Cuervo-Zazurra, & Nielson, 2014; Dawson, 2014), or nonmonotonicity (Schoonhoven, 1981), we note that our discussion, and guidance offered, is broadly applicable to these more complex cases. Additionally, we refer scholars to discussions about levels of measurement (Allison, 1977), symmetric interactions (Berry et al., 2012), false positive rates (Esarey & Sumner, 2016), and flexible estimation strategies (Hainmueller, Mummolo, & Xu, 2016). The key point we make is that researchers run the risk of overstating or understating results if they rely solely on interpreting the significance of the interaction coefficient to assess moderation a risk shown by our review of the extantIB literature, simulated regressions, and a Monte Carlo exercise. Overstating occurs when the researcher proposes a conditional hypothesis and then assumes, based upon finding a statistically significant interaction coefficient, that the marginal effect for all the values of the moderating variable are statistically different from zero. This assumption, it turns out, is not always true. A second hazard understating occurs when a researcher with a conditional hypothesis misses evidence of one or more nonzero marginal effects based solely upon the finding of a statistically insignificant interaction coefficient.