Q1) A critic is from Hawaii. He states that Hawaii has low voter turnout, but vo
ID: 3222551 • Letter: Q
Question
Q1) A critic is from Hawaii. He states that Hawaii has low voter turnout, but votes reliably Democratic. In 2004, Hawaii has 23.7% of workers in a union, 43.49% voter turnout, was 7.9% Hispanic and 2.2% black. Calculate the predicted probability of Bush winning Hawaii in 2004 (3 points).
Q2) How would you explain your result that, despite Hawaii’s low voter turnout, increased voter turnout predicts a decrease in the likelihood of Bush winning the state? That is, using what you know about statistics, explain how you can see a general trend of higher turnout predicting increased support for Bush, despite the fact that you don’t see this in Hawaii. (5 points)
Q3) Consider one variable omitted from this analysis. Explain what that variable is and why it should be added (4 points).
Q4) Consider one limitation of the linear probability model. Explain what that limitation is and why it may cause problems for your analysis (4 points)
Model Summary R Square Square .765 .586 5491 .329 a. Predictors: (Constant. Percent black (2004), Percent hispanic (2004), Percent workers who are union members (2004). Turnout in 2004 presidential election ANOVA Squares df Mean Square F 1.725 15.913 ,000 Regression 6,901 4.879 Residual a. Dependent Variable: Did Bush win electoral vote, 2004? b. Predictors: (Constant. Percent black (2004), Percent hispanic (2004), Percent workers who are union members (2004), Turnout in 2004 presidential election Coefficients Standardized Unstandardized Coefficients Coefficients Std, Error Beta Model (Constant) 5.880 2.997 510 Turno n 2004 008 027 384 3.373 presidential election Perce nt Workers who are 056 635 6.327 009 union members (2004) 007 005 136 1.300 Percent black (2004) 005 010 188 1.742 a. Dependent Variable: Did Bush win electoral vote, 2004? Sig 000 002 000 200 088Explanation / Answer
Q.1 The model is If bush win the Hawaii or not when there is
23.7% of workers in a union, 43.49% voter turnout, was 7.9% Hispanic and 2.2% black.
So the linear model is
yPr-Bush = 2.997 - .027 *x turout - 0.056 xunion- 0.007 x hispanic - 0.010 xblack
yPr-Bush= 2.997 - 0.027 * 43.49 - 0.056 * 23.7 - 0.007 * 7.9 - 0.010 * 2.2
= 0.41827
so 41.8% winning probability.
Q.2 here coefficient for voter turnout is negative in nature so a high voter turnout predicts decrease probability of bush winning the state. So, we can see a higher turnout predicting support for bush if there is very low percentage of union members among those. Because that is the most impactful variable in this regression. A small decrease in union member population percentage can increase support for bush even there is high voter turnout. Small factors like low Hispanic and black people percentage can also cause this result.
Q.3 The population of white Caucasian people can be added. they are majority there and even a small change in their population can create vast differences in poll results.
Q>4 The probability model dependent variable values are always limited to 0 to 1. so in this model if we keep decreasing voter turnout and union percentage (e.g. voter turnout = 40%, union percentage 14%), lets see what is the probability of bush winning .
yPr-Bush = 2.997 - .027 *x turout - 0.056 xunion- 0.007 x hispanic - 0.010 xblack
yPr-Bush= 2.997 - 0.027 * 40 - 0.056 * 14 - 0.007 * 7.9 - 0.010 * 2.2
= 1.0557 ( not feasible or practical)
Here we can see that the probability reached above 1 so it is not feasible and practical and there is not major changes in the model. Voter turnout kept very modest and just decreased union percentage. so, that's the biggest drawback of linear probability model.
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.