Figure 3.
Size and statistical significance of trend reversal coefficients by political identification.
Note. Trend reversal coefficient represents the interaction of Day of Test and our indicator for whether or not the test was taken after March 7. Sample restricted to respondents who did not self-identify either East Asian or South Asian. Horizontal line indicates neutral bias; below the line indicates pro-Asian bias; and above the line indicates pro-White bias. All regression models adjusted for age, gender, educational attainment, U.S. citizenship, and weekday.
*p < .05. **p < .01. ***p < .001.
