922 WordsMay 1, 20174 Pages

the “Lg10” function under the “Transform/Compute Variable” pull down menu feature of IBM’s Statistical Package for Social Sciences (SPSS).
Once the variables were transformed, we ran a correlation analysis between the operationalized independent variables, moderating variables, control variables and the dependent variable to check for collinearity and to begin to identify and evaluate non-causal associations and strengths of relationships between variables. We also applied Variance Inflation Factors (VIF) to detect any overlap or similarity between the independent or explanatory variables. The VIF results showed that multi-collinearity was not an issue. See Table 6 for the correlation and VIF results. Hypothesis Tests
To execute the*…show more content…*

An odds ratio (Exp(B)) greater than 1.0 indicates predictor variable is more likely to be resistant (1), whereas an odds ratio less than 1.0 indicates the predictor variable is less likely to be resistant.
Results
The Pearson Coefficient correlations summarized in Table 6 show six statistically significant non-causal associations between variables. In particular, there is a correlation of -.172 (significant at the 0.05 level) between Firm Performance Resistance and the independent variable Terrorism Exposure; this is counter to the direction of effect suggested in Hypothesis 1(a).
For a more robust test of the relationships, we conducted a series of stepped logit regressions, the first of which are summarized in Table 8, Analyses 1 and 2. Analysis 1 looks at the relationship between Terrorism Exposure and Firm Performance Resistance, with Time Since Last Attack as a moderating variable. Analysis 2 looks at the relationship between Terrorism Exposure and Firm Performance Resistance, with Business Continuity Plan as a moderating variable. The overall percentage shown in the model summary (Overall %) is the predicted percentage of firms that would have the outcome of resistance (coded as 1) to the focal terrorist attack – meaning that the stock price did not drop by 0.5% or more. Finally, the Nagelkerke R squared and the statistical significance (Model

Related

## Categorization Of The Global Terrorism Database ( 2015 )

1683 Words | 7 Pagesfive predictor variables in our conceptual model, and the generally accepted statistical power threshold of 0.8 (Cohen, 1988). Based on the availability of stock price data for measuring the dependent variable, we focused on Fortune 1000 MNEs since 1990, which narrowed down the possible incidents from 141,967 in the GTD to 158 terrorist incidents, representing 37 different MNEs. Variables Table 5 identifies and operationalizes the independent, moderating, control, and dependent variables used in the

### Categorization Of The Global Terrorism Database ( 2015 )

1683 Words | 7 Pages