The Democratic presidential nomination is still a close battle between Hillary Clinton and Barack Obama, and the upcoming Pennsylvania primary will be an important one. Obviously there are many factors involved in the outcome of any election, but some are more critical than others. Up to this point, African American population and median white income have been strong predictors in Democratic primaries, accounting for over 60% of the results. This doesn’t give the whole picture, but it could say something about Pennsylvania.

This article makes a comparison between states that have already voted and finds that Ohio is the most similar to Pennsylvania in the demographics being observed. Then the method of least squares is used to construct a model of Ohio’s primary results for each county. Variables included are African American population and median white income, in addition to percentage of residents aged 20-24. This last factor is included due to Obama’s popularity among college students. The least squares regression assigns weights to each of the variables, and then the resulting equation is applied to Pennsylvania. The demographics of each county are plugged in separately and then averaged, weighted by the populations of the counties.

This method indicates that Clinton and Obama will see an outcome in Pennsylvania much like the one they saw in Ohio. Of course there are many other factors that will come into play and this analysis can’t be given too much credence. But it’s an interesting application of the method of least squares nonetheless, and different from the content of the assignments in this class.

Source: http://www.realclearpolitics.com/horseraceblog/2008/03/a_review_of_the_pennsylvania_p.html