Conduct a similar analysis using the NELS data. Try using F1S45A as the measure of time spent watching TV and a mean of the 10th-grade test scores (F1TxRStd, F1TxMStd, F1TxSStd, F1TxHStd) as the outcome. Because NELS did not include measures of ability, test the interaction of TV and Previous Achievement (ByTests). Also control for base year SES (BySES). Is the interaction statistically significant? Graph the interaction

(or the lack of an interaction). How do you account for the differences between these findings and those from Exercise 1?