Chapter 7 • Modeling Relationships of Multiple Variables with Linear Regression 174 Summary Table means that 77.8% (Adj R-square=.778) of the variation in the teenage birthrate can be attributed to these four variables This is an excellent model. In fact, it is rare to find a model for social behavior that has such a high explanatory power. In sum, the model suggests that states that have higher proportions of births attributed to teenage mothers tend to also have higher levels of poverty and extend lower supports to the poor. It also suggests that per capita educational expenditure has no observable effect on the proportion of births attributed to teenage mothers. We also found that the relationship between race
and teen births maybe spurious, and this apparent relationship disappears when poverty rates are taken into account. However, it must be emphasized our intent here is not to offer a comprehensive analysis of this
challenging research question, but rather to use the question to illustrate a statistical technique. Certainly, additional research is warranted to focus on a variety of related questions concerning causality (does poverty
cause increased teen births, or do teen births
influence the poverty rates, levels of analysis (should the analysis be on state level
comparisons as we do here,
or on individuals, and measurement (even if overall education funding has no
effect on stemming teen births, it may matter how funds are spent.
Share with your friends: