Hi, I'm trying to conduct a logistic regression following a model in a paper on auditor conservatism in going concern reporting following the onset of the GFC in the U.S. However, there is one part that I don't understand:

Panel B provides the results from the expanded model that tests for a differential Big 4 versus non-Big 4 GCO reporting effect after the onset of the GFC and includes the BIG4 variable and the interaction term, GFC*BIG4. As expected, the coefficient on the BIG4 variable is negative, indicating a lower likelihood of receiving a GCO from a Big 4 auditor; however, the strength of this association is not significant (p-value = 0.25) in our sample. In addition, the interaction term in the expanded model is not significant (p-value = 0.40), indicating that there is no differential effect, based on audit firm size, on GCO reporting after the start of the GFC. Consistent with these firmsize results, we also assess the overall Big 4 reporting effect by testing the combined BIG4 and GFC*BIG4 coefficients and find that they are not significantly different than 0 (p-value > 0.45). However, we find that the overall GFC reporting effect, assessed by examining the combined effect of the GFC and the GFC*BIG4 coefficients, is positive and significant (p-value=0.02), providing additional evidence that auditors were significantly more likely to issue GCOs prior to bankruptcy after the onset of the GFC.

(Source: Geiger, Raghunandan & Riccardi (2014) - The Global Financial Crisis: U.S. Bankruptcies and Going-Concern Audit Opinions).

What exactly do they mean/do when they examine the combined effect of the GFC and the GFC*BIG4? Is that an additional test/regression? And how do you perform this?