Abstract
This article analyzes shifts in public opinion on presidential approval surrounding certain salient international events. It estimates the relative size of approvers, disapprovers, and neutral respondents in Gallup polls who shifted to other responses in polls before and after the events. This is done by analyzing all possible nonnegative, integer solutions to the 3 × 3 tables representing the pre- to postevent change. By examining the relations among cell values in all the solutions, it is shown that most of the increase in support accruing to the president comes from those who had disapproved of his performance before and not from those who held no opinion.

This publication has 4 references indexed in Scilit: