people who admire United States science and technology and visit the United States improve their attitude toward the United States substantially.
We lack data on how people’s attitudes toward the United States changed when they visited the United States; we have only cross-sectional data on attitudes. It is therefore plausible to argue that despite the controls we have added to our regression, the “US travel” variable is still endogenous. The regressions may suffer from omitted-variable bias: Other variables that we have not controlled for, and that might not even be in our dataset, may affect one’s opinion of the United States and be correlated with whether one travels to the United States. Reverse causality may be at work: Having a good attitude toward the United States may cause travel to the United States, and not the other way around. Because the data do not contain a good instrument for travel to the United States, we have tried to address the issues in the ways discussed above, but our conclusions must still be viewed with caution.