Here are the results of a regression of Car Deaths in the UK by month from Jan 1969 to Dec 1984 on a dummy variable: 0 = no seatbelt law, and 1 = seat belt law (the law was instituted in February 1983)

Coefficients:

Estimate Std. Error t value Pr(> | t |)

(Intercept) 125.870 1.849 68.082 < 2e-16 *

Seatbelts -25.609 5.342 -4.794 3.29e-06 *

R-squared = 0.11

Did the seat belt law make a difference? Is there a need to add more variables to the model? How would you justify your answer with numbers? What possible independent / predictor variables could you add to this model?