PAPERBACK
\$43.00

### REGRESSION ANALYSIS

Raw correlations can be misleading, so we have also performed regression analysis to attempt to control for other factors that might influence one’s opinion of the United States In these regressions, the dependent variable is a person’s opinion of the United States, and the independent variables include a variable representing whether one has traveled to the United States and several control variables. The results of the ordinary least-squares (OLS) regressions in which USOPINION is the dependent variable are reported in Tables C-1 and C-2.

Robust t-statistics are stated in parentheses. The right-hand-side variables included in all regressions are dummy variables representing the effect of each country and controls for age, education, and whether the respondents have friends or family in the United States with whom they correspond regularly. “Extra Controls” are controls for the respondent’s opinion of the dominant country in his or her region and whether the respondent has traveled outside his or her own country in the last 5 years. An observation is excluded from the regression if the respondent answered “Don’t Know” or refused to give a response on any of the relevant questions.

Each numbered column represents a different regression. “N” represents the sample size in each regression (the number of people whose responses are included in the regression). “R-Squared,” which can take values from 0 to 1, represents the amount of the variation in the dependent variable that has been accounted for in the regression by the variation in the independent variables. When R-Squared is higher, more of the variation in the dependent variable has been accounted for by the variation in the independent variables.

#### The Effect of Traveling to the United States on One’s Opinion of the United States

The row labeled “USTRAVEL” displays the estimated effect in each regression (1, 2, and 3) of United States travel on a person’s opinion of the United States. There are three important components of each box in this row. The first number in each box represents the estimated effect of visiting the United States on one’s opinion of the United States.

In regression 1 (Table C-1), OLS regression analysis of responses to the Pew survey indicates that visiting the United States raises one’s opinion of the United States by 0.109, on the scale that runs from 0 to 3. The t-statistic is in parentheses below the first number and represents how precise our estimate is. In regression 1, the t-statistic is 4.00, which is rather high, indicating that our estimate is somewhat precise. Asterisks after the t-statistic indicate whether the estimated coefficient is significantly different

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001