To determine the effect of union membership run a regression with income (RINCOME) as the dependent variable and age (AGE), education (EDUC), number of hours worked per week (HRS1), and union membership (UNION:1 = Respondent belongs, 2 = Spouse belongs, 3 = Both belong, 4 = Neither belong). Is there sufficient evidence that union membership for the respondent, spouse, or both affects income after removing the effects of age, education, and hours worked?