Calculate the "intention-to-treat" (ITT) effect by running the reduced form regression. That is, using robust standard errors, regress complete on message. Based on your regression results, how much does receiving a message change the likelihood of completing the next week? Assign this number (the coefficient in front of variable message) to the variable l_change2_3 and ensure that its data type is float. (Round to four decimal places.) (1 pt)