Our "comm" variable is nice and ordinal, but it's not really a ratio level variable and thus really shouldn't be used in a regression format in such as manner. Instead, we decide to use a dummy variable approach. The variables "urban," "rural," and "suburban" are each dichotomous variables that indicate whether the respondent is either in or not in those respective categories. We decide to use suburban (50%) as the reference category. In considering the following results, what more have we learned about the relationship between locale and support for abortion? Explain what our dummy variable approach tells us. Interpret the findings in terms of the null hypotheses and for the layperson.

. regress abort rural urban

Source |SSdfMSNumber of obs=305

-------------+----------------------------------F(2, 302)=2.18

Model |9.1027415224.55137076Prob > F=0.1145

Residual |629.6316853022.08487313R-squared=0.0143

-------------+----------------------------------Adj R-squared=0.0077

Total |638.7344263042.10110009Root MSE=1.4439

------------------------------------------------------------------------------

abort |Coef.Std. Err.tP>|t|[95% Conf. Interval]

-------------+----------------------------------------------------------------

rural |-.3973856.1953319-2.030.043-.7817695-.0130017

urban |-.0467271.2115281-0.220.825-.4629827.3695284

_cons |3.091503.116733126.480.0002.861793.321217

------------------------------------------------------------------------------