Multiple Regression by David Brown

tiRed af

kale0498 — 2016-10-03 18:27:44 UTC

What determines the political ideology of a state?

A state’s ideology is determined by several factors, no one variable can accurately predict the ideology of a given state. By observing per capita income, religiosity, and urban population, we can more accurately determine the ideology of a state. A high per capita income for a state makes it more likely to be conservative. The desire to expand government programs is most likely low due to a lack of need for governmental assistance. Similarly, the closer to zero a state’s religiosity score is, the more likely it is to be conservative as well. A possible explanation is that those who have strong religious beliefs, particularly christian based ones, are more likely to advocate for the conservation of our government in such a way as to promote or protect their own beliefs. Contrarily, a larger urban population will increase the likelihood that a state is liberal. Urban areas have dense populations and are more likely to have better access to information than rural areas. There is great variety in the types of people in urban areas which can often lead to a more liberal way of thinking.

Bombastic With A Side of Awesome

2016-10-03 18:35:51 UTC

Hypothesis:
A state will be more liberal with a large population of urban and college attendance.

The pvalue is greater than .05 so we fail to reject our null, so are hypothesis must be accepted.

Stats out for R-ambe

2016-10-03 18:45:24 UTC

hypothesis: States in the South will be more conservative. Additional factors included are per capita income, religiosity, urban percentage and unemployment.

Model: The religiosity variable is the most statistically significant, and the south variable has no significance at all, thus the new hypothesis is that religion has more to do with a the conservatism of a state.

Model: In addition to religiosity, Variable factors include republican percentage, unemployment, urban percentage and gun rank of state. This shows that republican percentage and religiosity have the highest statistical significance. Urban percentage also has some significance, but the main factors that controls a state's conservatism are religion and republican political ideology.

Ryan, Aiden, Sophia

Team Riff Raff

2016-10-03 21:19:17 UTC

Hypothesis: In the interesting political system that the United States follows, there are well established conservative and liberal states, states that almost without exception either vote conservative or liberal and we desire to know the method behind the madness, what decides if a state votes red or blue? We decided upon several variables that we believed would determine how conservative a state is, including per capita income, level of education, population density, urban population, church attendance, religiosity and if the state is a southern state or not. Walking through the variables we figured that a higher per capita income, high church attendance, high religiosity and southern states would be far more likely to vote conservative. While on the other hand a high level of college education, high population density and large urban population will all lead to a more liberal state. In particular we think per capita income will have a large impact on deciding which way a state will vote and for that reason we decided to take a closer look at variable.
Conclusion: First we note that the R squared values of the two models were not very different meaning and therefore the variable (religiosity) we removed for the second nested model did not have much of an effect on the rest of the data. For per capita income there was the same coefficient between the two models, for every .0003 increase in dollars there is a one percent increase in the conservative measure of a state, meaning the measure is insignificant. When looking at the college variable for every about 3.2 percent decrease in college education there is a one percent increase in the conservative percentage for a state. For every .007 decrease in population per square mile there is a one percent increase in the conservative percentage of a state. For every .051 decrease in the percent of people living in an urban area in a state there is a one percent increase in the conservative percentile of that state. For every about .31 increase in the percentage of people going to church in a state there is a one percent increase in the conservatism percentile. For every change in the degree of religiosity from low to medium or medium to high there is a 1.1 percent increase in the conservative percentile.Examining the variable we determined that per capita income was not significant enough factor that affects the conservative percentage which is seen by the very small coefficient when looking at the other variables the percent of population with a college education and the percent of the population that regularly attends church seems to be more significant due to their higher coefficients.

Three MuskRteers (Emily, Marine, Andee)

2016-10-04 00:52:25 UTC

R we done yet

2016-10-05 15:34:34 UTC

Hypothesis: The more religious a state is, the more conservative the population is.

Theoretical Model: Religion or lack of religion influences political ideology because values often reflect their political beliefs or choices. If a person or population is highly religious they are going to have certain values that influences their preferences on politics. For example, if highly religious person might be against abortion and therefore vote for the more conservative politics. A religious person also might feel obligated to vote a certain way due to their religious beliefs and they might feel like they are doing something wrong or against their religion if they identify with more liberal political ideologies.

This model shows that our hypothesis is incorrect. We thought the most influential factor would be religiosity, but it appears to be population density. For every one person increase per square mile, the population that is conservative goes down 0.0063646%. For every one unit increase on the religious observance belief scale, the population that is conservative increases by 0.0319681%. While we were correct in assuming that more religious states are more conservative, population density still has a higher correlation with conservatism.

Stats R Us (Bella, Maya, Malia)

2016-10-05 17:37:51 UTC

Thesis:

We think that the variance in political ideology (measured by a state’s conservatism measure) will be best explained by religiosity of the state after controlling for state location (southern or not), per capita income of the state, college education of the state, and percent of urban population in the state. We think that these are the most important variables that will have an effect on a state’s measure of conservatism and thus to see religiosity true effect on political ideology we must control for location, per capita income, college education levels, and urbanity. We predict that religiosity and conservatism will have a positive relationship; as religiosity increases, the percent of conservative people in the state will go up.

First Regression Model:

Call:

lm(formula = conpct.m ~ religiosity + south + prcapinc + college +

urban, data = states)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-7.1753 -1.7036 0.1989 1.2606 7.2659

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 5.169e+01 3.694e+00 13.993 < 2e-16 ***

religiosity 7.723e-02 1.530e-02 5.047 8.25e-06 ***

southSouth -2.275e+00 1.381e+00 -1.647 0.1067

prcapinc 8.607e-05 1.933e-04 0.445 0.6583

college -2.498e-01 1.728e-01 -1.446 0.1554

urban -8.010e-02 3.606e-02 -2.221 0.0315 *

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.121 on 44 degrees of freedom

Multiple R-squared: 0.6935, Adjusted R-squared: 0.6587

F-statistic: 19.91 on 5 and 44 DF, p-value: 2.519e-10

Conservatism= 51.69 + .07723(religiosity) - 2.275(south) + .00008607(prcapinc) - .2498(college) - .08010(urban)

According to this model, religiosity is the is most significantly associated with political ideology and is positively correlated with conservatism. It has 3 stars which means it is almost 100 percent true to the relationship in the full population; it has a p-value of about .00000825. This model explains 66 percent of the variation in political ideology, and confirms our hypothesis.

A Revised Regression Model:

Call:

lm(formula = conpct.m ~ religiosity + south + prcapinc + college +

urban + over64 + blkpct08, data = states)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-7.0346 -1.4937 -0.2675 1.6395 6.5366

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 60.7112542 5.0982859 11.908 4.76e-15 ***

religiosity 0.0765297 0.0154466 4.954 1.23e-05 ***

southSouth -2.3541528 1.5677875 -1.502 0.1407

prcapinc 0.0001707 0.0001950 0.876 0.3862

college -0.3457134 0.1696551 -2.038 0.0479 *

urban -0.0970896 0.0355026 -2.735 0.0091 **

over64 -0.6418922 0.2592236 -2.476 0.0174 *

blkpct08 -0.0020226 0.0724300 -0.028 0.9779

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 2.982 on 42 degrees of freedom

Multiple R-squared: 0.7329, Adjusted R-squared: 0.6884

F-statistic: 16.46 on 7 and 42 DF, p-value: 3.25e-10

Adding the variables over64 (percent of population 65 and over) and blkpct08 (percent of black population in 2008) has changed our model. Religiosity is still the most significant with three stars, but college, urban, and over64 also are shown to be significant indicators of political ideology. It didn’t really add more percentage to the R-squared; this model only explains another 3 percent of the variation (from 66 to 69).

We think this model explains variation the best because it has the highest R-squared and the most significant variables. Holding all else equal, as religiosity increases by one percentage point, conservatism increases .077 percentage points. As you move from a southern to a non southern state, conservatism decreases by 2.354 percentage points. As Per Capita Income increases one dollar, conservatism increases by .00017 percentage points (for every 1000 dollar increase, conservatism increases by .17 percentage points). As percent of the population that graduated college increases by one percent, conservatism decreases by .346 percent. As the percent of the urban population increases by one percentage point, conservatism decreases by .097 percentage points. As percent of the population over 65 increases by one percent, conservatism decreases .64 percentage points. As the black population increases by one percentage point, conservatism decreases by .002 percentage points.

We R Buffs

2016-10-05 17:56:51 UTC

Hypothesis: States in the South tend to represent more conservative views

Save R Ship

2016-10-05 18:06:24 UTC

The R Club

2016-10-05 18:15:14 UTC

badgalRR

2016-10-05 18:20:56 UTC

badgalRR

2016-10-05 18:20:56 UTC

Using multiple regression, we intend to prove that the political ideology of a state is determined by college and church attendance. We believe there will be an inverse/negative relationship between the amount of college people in a state have attended and how conservative a state is. As college attendance goes up the less conservative a state will be. Contrastly, as church attendance increases the conservative nature of the state will increase as well.

Call:

lm(formula = conpct.m ~ attend.pct + college, data = states)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-7.1975 -1.8017 -0.1975 1.9728 8.0422

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 32.72316 4.76824 6.863 1.32e-08 ***

attend.pct 0.31445 0.05845 5.380 2.30e-06 ***

college -0.42542 0.12100 -3.516 0.000983 ***

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.248 on 47 degrees of freedom

Multiple R-squared: 0.6455, Adjusted R-squared: 0.6304

F-statistic: 42.79 on 2 and 47 DF, p-value: 2.606e-11

From the data summary above, we see that there is a positive slope (.31445) between church attendance and conservatism, meaning that there is a positive relationship between college attendance and conservatism. It also shows that there is a negative correlation (-0.42542) between college education and conservatism. Both these findings support our hypothesis. Our R-squared shows 63 percent of variance, which means that college and church attendance explains about 63% of state ideology.

When running each variable separately, we saw that the relationships were the same. Our original model with both variables, however, has stronger statistical significance, a higher R-squared value, and the coefficients are larger.

Stat Globetrotters (Jay Gideon, Danny Flaherty. Jason A-Fuentes)

2016-10-05 18:24:47 UTC

Hypothesis: we believe that the variables of urban population, population density, and college education will distribute a more liberal view among a states political orientation. We believe that the more urban a population will have the strongest effect on a states political orientation, because a trend of liberal values is often seen in metropolitan areas.

Data:
In our model, we ran multiple regressions that showed the affect that Urban population, Population Density and College education had on conservatism. We then ran two more regressions in model 2 and 3, excluding college and urban population respectively. We can see that all variables have a negative relationship to conservatism in a state.

However, our hypothesis was right in that an urban population isn't as large a factor as college education. We were right that college education does promote a more liberal view, but we can see from model 1 that for every percent increase in urban population there is a -1.9 percent change in conservatism where college population only has -.55 change

We R Not Prepared

2016-10-05 18:25:03 UTC

I Thought This Was Speed Dating

2016-10-05 18:27:02 UTC

The Trio (Emily, Allie, Megan)

2016-10-05 18:29:14 UTC

Theoretical Model: There are many factors that determine the political ideology of a state. Factors such as religion, race, gender, per capita income, having a college degree, population, and even where the state is located, all play a role in determining a states’ political views. As a common idea, coastal states typically lean more towards the liberal side than that of central states, and southern states are usually more conservative than other parts of the country. But there’s still several exceptions to these inclinations. Another key factor in determining states’ political ideology is population density. In the 2010 election, out of the 15 largest cities in America, Obama took 11 of those. It is said that the main reason for these cities to be democratic is due to their dense population which sprouts new economic progress. We think that religion and density will have the highest impact on whether a state is conservative or not. We believe the more religious a state is the more conservative the population will be, while states with higher density will have populations that are less conservative.

assignment4.lm <- lm(conpct.m~ attend.pct + college + religiosity + density, data=states)

lm(assignment4.lm)

plot(assignment4.lm)

Call:

lm(formula = conpct.m ~ attend.pct + college + religiosity +

density, data = states)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-7.885 -1.329 -0.056 1.354 6.792

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 33.687152 8.946839 3.765 0.00048 ***

attend.pct 0.221152 0.160738 1.376 0.17567

college -0.197843 0.133976 -1.477 0.14672

religiosity 0.021073 0.030448 0.692 0.49243

density -0.006108 0.001945 -3.140 0.00298 **

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 2.976 on 45 degrees of freedom

Multiple R-squared: 0.7149, Adjusted R-squared: 0.6896

F-statistic: 28.21 on 4 and 45 DF, p-value: 9.318e-12

Dependent variable:

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

conpct.m

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

attend.pct 0.221

(0.161)

college -0.198

(0.134)

religiosity 0.021

(0.030)

density -0.006***

(0.002)

Constant 33.687***

(8.947)

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

Observations 50

R2 0.715

Adjusted R2 0.690

Residual Std. Error 2.976 (df = 45)

F Statistic 28.213*** (df = 4; 45)

===============================================

Note: *p<0.1; **p<0.05; ***p<0.01

The Deplorables

2016-10-05 18:29:21 UTC

Hypothesis:
We believe that that there is a positive correlation between religiosity and conservativeness. and a negative correlation between both per-capita income and density to conservativeness.

Call:
lm(formula = conpct.m ~ prcapinc + religiosity + density, data = states)

Residuals:
Percent mass public Conservative
Min 1Q Median 3Q Max
-7.7466 -1.6414 0.0252 1.4132 6.4896

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 4.113e+01 4.445e+00 9.253 4.53e-12 ***
prcapinc 2.895e-05 1.656e-04 0.175 0.86201
religiosity 6.903e-02 1.103e-02 6.258 1.19e-07 ***
density -7.118e-03 2.297e-03 -3.099 0.00331 **
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.056 on 46 degrees of freedom
Multiple R-squared: 0.6927, Adjusted R-squared: 0.6727
F-statistic: 34.57 on 3 and 46 DF, p-value: 7.532e-12

Conclusion:
Based off the table we have found that we were correct in two of our three hypotheses were correct. There is a positive relationship between religiosity and how conservative a state is (.069). There is also a negative correlation between density and conservativeness (-.0071). However, we incorrectly predicted a negative relationship between per-capita income and conservativeness. There is a very slight but positive relationship between the two.

The Rsonists

2016-10-05 18:30:33 UTC

The Rsonists

2016-10-05 18:30:33 UTC

The Rsonists (

2016-10-05 18:30:34 UTC

The Rsonists (Abigail and Brendan)

2016-10-05 18:30:36 UTC

What variables make a state more conservative? The idea of conservatism is abstract and as such its building blocks are complex, however, we have decided to analyze how the percentage of college education, the percentage of unionization, the percentage of the state over 65, and the percent of a state’s population that is urban. We believe all of these factors will be influential but particularly believe that states with an older population will be more conservative. The variables for college education, unionization and urban population will be controls.

In the first model we can see that our hypothesis was not entirely wrong however the percent over 64 variable does not appear to be the most statistically significant. Our hypothesis was also wrong in the sense that the variable over 64 has a negative relationship to the percent conservative in a state meaning that as the percent over 64 increases by 1% the percent of the state that is conservative decreases by 0.7128. All of our coefficients are negative but for our controls this was expected. The variable with the most statistical significance is the percent of college graduates. The stars represent how statistically significant a variable is, the percent population of 64 is significant to the 99th percentile, the variable of percent unionized was accurate to the 99.9 percentile and the percent college educated was closer to the 100th percentile. The coefficient for the percentage over 64 is actually higher than the coefficients for the other variables, however it also has the highest standard error which tells us that the variable is not as statistically significant. The multiple R squared value for this model is .5996 which is a good high number.

In the new model we added the variable of per capita income in a state. It is important to note that this is the only variable in the model not measured through a percentage. One difference between this model and the previous one are that the statistical significance of the percent college education decreased. Another interesting observation is that the coefficients are no longer represented as values less than one because they are no longer all percentages. The standard error values have also changed with the introduction of a new variable, the percent error for percentage over 64 was previously 0.298 and in this new model is 3.074. The multiple R squared value is higher in this model than the previous one however only by a very small amount.

RateRRRR

2016-10-05 18:31:07 UTC

(Jessica Martinez, Yoselin Martin, Megan White)

Why R We Here?

2016-10-05 18:31:09 UTC

We R Confused

2016-10-05 18:31:14 UTC

Hypothesis: Using multiple regression we will attempt to prove that a higher college education rate, higher per capita income, and higher abortion rate will show a correlation to states showing a less conservative population. We believe that the strongest correlation will be between college education rates and conservatism.

Conclusion: All of the variables that we selected showed negative correlation with conservative population rates. The strongest correlation is potentially the association between college rates and conservative population rates. For every percentage of population college educated, the state shows 1% less conservative population. The correlation between income per capita might seem fairly week, but one must take in to consideration that income is measured per dollar. If it were measured in 1,000 dollars, the correlation would appear to be much stronger.

the DowneR

2016-10-05 18:31:21 UTC

Hypothesis - When it comes to political policies there are many social variables that directly relate one to their political affiliation. We believe that two of the more prominent variables that would attribute to one’s political ideology would be abortion laws and gun laws because of the social division between states and individuals. When it comes to abortion laws and gun laws there is a strict divide between party affiliation. Conservatives maintain on a stature of having very few gun laws and having very strict regulations or no laws on abortions while liberals tend to be the opposite. Although, are huge factors contributing to political ideologies, we believe that religiosity will have the greatest influence on political ideology because of the Republican's strong religious based system.

Rnold Statssenegger

2016-10-05 18:31:36 UTC

Luke Albers
Antonia Wright
Logan Dingwall

The StandaRd Deviants

2016-10-05 18:31:39 UTC

Mike, Sahar, John

Scranton Stranglers

2016-10-05 18:31:41 UTC

R Squad?

2016-10-05 18:33:23 UTC

Multiple Regression

Brett, Connor, Evan

R Squad

2016-10-05 18:33:46 UTC

2016-10-05 18:33:51 UTC

2016-10-05 18:34:02 UTC

2016-10-05 18:34:08 UTC

VaRsity Stats

2016-10-05 18:34:12 UTC

r2d2

2016-10-05 18:34:15 UTC

Les Tabarna

2016-10-05 18:34:28 UTC

Les Tabarnaks

2016-10-05 18:34:30 UTC

Les Tabarnaks

2016-10-05 18:34:36 UTC

Alexis

Les Tabarnaks

2016-10-05 18:34:40 UTC

Alexis and

Les Tabarnaks

2016-10-05 18:34:47 UTC

Alexis and Thomas

]

terence_weiller — 2016-10-05 18:34:52 UTC

ab(sin)the

terence_weiller — 2016-10-05 18:34:53 UTC

We hypothesize that while political positions are determined by the level of affiliation people have with several different social groups, some of these groups are significantly more impactful than others. The most impactful of these groups, namely, is the extent with which people observe a religious practice.

Team LKC

2016-10-05 18:35:19 UTC

In the data looking at political ideology, we predict the variables including, college, urban and religiosity will have an effect on whether states are conservative, independent or liberal. We predict that high college populations in states will be more liberal. More urban states depending on the geographical location of state will be more independent. And more religious states will be more conservative.

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-7.475 -1.800 0.377 1.596 6.890

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 50.89375 2.97165 17.126 < 2e-16 ***

college -0.23976 0.13602 -1.763 0.0846 .

urban -0.06817 0.03412 -1.998 0.0517 .

religiosity 0.05980 0.01074 5.566 1.29e-06 ***

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.146 on 46 degrees of freedom

Multiple R-squared: 0.6745, Adjusted R-squared: 0.6533

F-statistic: 31.78 on 3 and 46 DF, p-value: 2.8e-11

Quantitative Drinkers

2016-10-05 18:35:31 UTC

Rip

2016-10-05 18:35:41 UTC

RiphaRambe

2016-10-05 18:35:46 UTC

Rip haRambe

2016-10-05 18:35:47 UTC

We R Not Prepared

2016-10-05 18:35:49 UTC

Why R We Here?

2016-10-05 18:35:59 UTC

Radical Rhinoceri

2016-10-05 18:36:00 UTC

Hypothesis: We believe that the combination of religiosity, income, as well as geography will determine a state’s leaning towards their respective historical ideology.

MODEL 1 -

Call:

lm(formula = cook.index ~ prcapinc + religiosity + south, data = states)

Residuals:

Higher scores more Dem

Min 1Q Median 3Q Max

-20.158 -3.119 1.176 3.811 15.917

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) -3.243e+01 7.628e+00 -4.251 0.000103 ***

prcapinc 5.687e-04 2.789e-04 2.039 0.047196 *

religiosity -1.062e-01 3.103e-02 -3.421 0.001320 **

southSouth 4.457e+00 2.838e+00 1.571 0.123076

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 6.519 on 46 degrees of freedom

Multiple R-squared: 0.5025, Adjusted R-squared: 0.4701

F-statistic: 15.49 on 3 and 46 DF, p-value: 4.218e-07

MODEL 2 -

Call:

lm(formula = cook.index ~ prcapinc + college + religiosity +

south, data = states)

Residuals:

Higher scores more Dem

Min 1Q Median 3Q Max

-18.754 -3.621 1.311 3.985 15.211

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) -3.290e+01 7.658e+00 -4.296 9.17e-05 ***

prcapinc 3.281e-04 3.826e-04 0.858 0.3957

college 3.310e-01 3.597e-01 0.920 0.3624

religiosity -1.020e-01 3.141e-02 -3.248 0.0022 **

southSouth 4.485e+00 2.842e+00 1.578 0.1216

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 6.53 on 45 degrees of freedom

Multiple R-squared: 0.5117, Adjusted R-squared: 0.4683

F-statistic: 11.79 on 4 and 45 DF, p-value: 1.237e-06

MODEL 3 -

Call:

lm(formula = cook.index ~ prcapinc + density + religiosity +

south, data = states)

Residuals:

Higher scores more Dem

Min 1Q Median 3Q Max

-17.2914 -2.9624 0.7001 3.9065 15.5666

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) -1.848e+01 9.219e+00 -2.004 0.051077 .

prcapinc 6.204e-05 3.362e-04 0.185 0.854413

density 1.153e-02 4.712e-03 2.447 0.018367 *

religiosity -1.102e-01 2.952e-02 -3.732 0.000532 ***

southSouth 3.408e+00 2.729e+00 1.249 0.218136

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 6.192 on 45 degrees of freedom

Multiple R-squared: 0.561, Adjusted R-squared: 0.5219

F-statistic: 14.37 on 4 and 45 DF, p-value: 1.232e-07

MODLE 4 -

Call:

lm(formula = cook.index ~ prcapinc + urban + religiosity + south,

data = states)

Residuals:

Higher scores more Dem

Min 1Q Median 3Q Max

-19.0090 -2.5986 0.9915 4.0926 14.1907

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) -3.387e+01 7.641e+00 -4.433 5.91e-05 ***

prcapinc 3.614e-04 3.172e-04 1.139 0.260619

urban 9.918e-02 7.431e-02 1.335 0.188722

religiosity -1.137e-01 3.129e-02 -3.634 0.000713 ***

southSouth 5.147e+00 2.861e+00 1.799 0.078696 .

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 6.465 on 45 degrees of freedom

Multiple R-squared: 0.5215, Adjusted R-squared: 0.4789

F-statistic: 12.26 on 4 and 45 DF, p-value: 7.994e-07

MODEL 5 -

Call:

lm(formula = cook.index ~ prcapinc + attend.pct + religiosity +

south, data = states)

Residuals:

Higher scores more Dem

Min 1Q Median 3Q Max

-19.2188 -2.6250 0.8409 3.1618 17.0196

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) -4.554e+01 2.227e+01 -2.045 0.0467 *

prcapinc 5.180e-04 2.922e-04 1.773 0.0830 .

attend.pct 2.451e-01 3.907e-01 0.627 0.5336

religiosity -1.563e-01 8.581e-02 -1.821 0.0752 .

southSouth 5.385e+00 3.217e+00 1.674 0.1010

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 6.563 on 45 degrees of freedom

Multiple R-squared: 0.5068, Adjusted R-squared: 0.463

F-statistic: 11.56 on 4 and 45 DF, p-value: 1.533e-06

Conclusion: The most significant factor in determining state ideology is religiosity. This is a variable that remains significant throughout different regressions. Therefore we have concluded that ideology is not determined through a direct correlation between the three variables of prcapinc, religiosity, and south (geography).

Team 1

2016-10-05 18:36:04 UTC

Les Tabarnaks

2016-10-05 18:36:10 UTC

Alexis and Thomas

Why R We Here?

2016-10-05 18:36:13 UTC

Hypothesis: Political ideology (measured by how conservative a state is) will be best explained by the percentage of the population that attends church (attend.pct). States with higher church attendance will be more conservative. We will run a multiple regression analysis looking at the independent variables college, prcapinc, and attend.pct to determine what accounts for political ideology. We chose this variables because we believe they each have some sort of impact on political ideology and will help us determine which is most important.

Call:

lm(formula = conpct.m ~ college + prcapinc + attend.pct, data = states)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-7.7306 -1.5901 -0.0431 2.0177 8.0723

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 34.6975986 5.8601181 5.921 3.80e-07 ***

college -0.3491172 0.1780393 -1.961 0.056 .

prcapinc -0.0001105 0.0001880 -0.588 0.560

attend.pct 0.3037532 0.0616078 4.930 1.11e-05 ***

---

Signif. codes:

0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.27 on 46 degrees of freedom

Multiple R-squared: 0.6481, Adjusted R-squared: 0.6252

F-statistic: 28.24 on 3 and 46 DF, p-value: 1.65e-10

Conservatism = 34.69 – 0.349(college) -0.0001(prcapinc)+0.304(attend.pct)

Conclusion:

Based on our analysis, the model shows that church attendance is the only independent variable that is statistically significant to political ideology. It has three stars and a p-value of 1.11e-05. The model accounts for 63% of the variation in political ideology according to the r^2 value, which confirms our hypothesis that church attendance is important to determining political ideology.

2016-10-05 18:36:13 UTC

Team 1

SEJ Research

2016-10-05 18:36:16 UTC

Why R We Here?

2016-10-05 18:36:27 UTC

Call:

lm(formula = conpct.m ~ college + prcapinc + attend.pct, data = states)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-7.7306 -1.5901 -0.0431 2.0177 8.0723

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 34.6975986 5.8601181 5.921 3.80e-07 ***

college -0.3491172 0.1780393 -1.961 0.056 .

prcapinc -0.0001105 0.0001880 -0.588 0.560

attend.pct 0.3037532 0.0616078 4.930 1.11e-05 ***

---

Signif. codes:

0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.27 on 46 degrees of freedom

Multiple R-squared: 0.6481, Adjusted R-squared: 0.6252

F-statistic: 28.24 on 3 and 46 DF, p-value: 1.65e-10

Conservatism = 34.69 – 0.349(college) -0.0001(prcapinc)+0.304(attend.pct)

Conclusion:

Why R We Here?

2016-10-05 18:36:28 UTC

Call:

lm(formula = conpct.m ~ college + prcapinc + attend.pct, data = states)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-7.7306 -1.5901 -0.0431 2.0177 8.0723

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 34.6975986 5.8601181 5.921 3.80e-07 ***

college -0.3491172 0.1780393 -1.961 0.056 .

prcapinc -0.0001105 0.0001880 -0.588 0.560

attend.pct 0.3037532 0.0616078 4.930 1.11e-05 ***

---

Signif. codes:

0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.27 on 46 degrees of freedom

Multiple R-squared: 0.6481, Adjusted R-squared: 0.6252

F-statistic: 28.24 on 3 and 46 DF, p-value: 1.65e-10

Conservatism = 34.69 – 0.349(college) -0.0001(prcapinc)+0.304(attend.pct)

Conclusion:

Squad

2016-10-05 18:36:43 UTC

Katie Bowser

Nina Asher

Lauren Blake

Dylan Robinson-Ruet

In-Class Assignment 4

Hypothesis:

We believe that as religiosity increases and college attendance decreases, the amount of conservatism increases. States with a higher college attendance will be less conservative because college students tend to be more liberal. States with a higher religiosity will be more conservative because religious people tend to be more conservative.

Multiple Linear Regression:

Y = 48.723 - .344X + .059

Best fit model:

===============================================

Dependent variable:

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

conpct.m

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

religiosity 0.070***

(0.009)

urban -0.091***

(0.032)

Constant 47.408***

(2.267)

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

Observations 50

R2 0.653

Adjusted R2 0.638

Residual Std. Error 3.215 (df = 47)

F Statistic 44.133*** (df = 2; 47)

===============================================

Note: *p<0.1; **p<0.05; ***p<0.01

Adjusted R-squared for college = 41%

Adjusted R-Squared for religiosity = 58%

Adjusted R-Squared for college and religiosity = 63%

Adjusted R-Squared for every variable = 69%

SpectaculaR

2016-10-05 18:36:52 UTC

Theoretical Model:

We believe that level of college education, percentage black, and income per capita, as well as the level of gay support within a state affect the level of conservatism observed in a state. We feel that college education does have an effect on whether or not a state is conservative or not because education, or lack thereof, can effect how to evaluate political issues and party platforms. In regards to level of gay support and black people within a state, we feel that these may affect how conservative a state is because these are highly contested identities between party platforms. Lastly, we feel that per capita income within a state would be a determining factor for the level of conservatism within a state, because conservatives tend to want to allocate the nation’s money differently than democrats and have different economic values.

Multiple Linear Regression Model:

Stargazer Model:

Our Analysis & Conclusions:

Using the Stargazer table, we observe that as the number of black individuals within a state raises 1%, the level of conservatism raises by .065%, showing a positive correlation between black population and conservatism. As well, as college education raises one percent there is a decline of about .188% points in the conservatism of a state. We see similar behaviors as states move along the Billman’s Scale one point, their conservatism lowers .575% points correspondingly. Finally, as per capita income raises one dollar, the level of conservatism of that state raises by .001%. We have an adjusted R-squared of .7351, which means our linear model can predict the observed data with ~74% accuracy. Given this model, we can conclude that each of these factors play a role in making a state more conservative to a certain degree. However, the data shows us that per capita income is likely the number one contributor. With this in mind, however, we can see that the although per capita income seems to have quite an effect on conservatism (compared to the other variables) its P value (.699) makes the variable statistically insignificant in this case.

The Benchwarmers

evancook53 — 2016-10-05 18:36:52 UTC

We hypothesize that a state's political ideology will directly correlate to urban population, religiosity, per capita income, a state’s gun laws, and George W. Bush voter turnout. If our hypothesis is correct, a higher urban population will have a negative correlation with the number of conservatives in the state, because higher urban population generally means a more liberal society. Also, looking at the religiosity of each state will allow us to see a positive correlation between a more religious society and a more conservative state. The more religious a population, the more likely conservative views are held on policies like gun laws and abortions and thus more likely to be republican than democrat. There is a positive correlation between political ideology and per capita income as republicans tend to be more fiscally conservative. There is a positive correlation between political ideology and those who voted in favor of George Bush in 2000 as those states were largely conservative rather than Democratic. Finally, we believe that not only will urban population and religiosity correlate with the political ideology of each state but also with one another. A higher urban population should have a lower religiosity.

Call:

lm(formula = conpct.m ~ urban + religiosity + prcapinc + gunlaw.rank,

data = states)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-8.7066 -1.4481 0.5083 1.8004 5.0734

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 4.000e+01 5.713e+00 7.001 1.01e-08 ***

urban -5.311e-02 3.595e-02 -1.477 0.1465

religiosity 6.023e-02 1.097e-02 5.490 1.77e-06 ***

prcapinc 3.078e-05 1.651e-04 0.186 0.8530

gunlaw.rank 1.090e-01 4.689e-02 2.325 0.0246 *

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.072 on 45 degrees of freedom

Multiple R-squared: 0.6963, Adjusted R-squared: 0.6693

F-statistic: 25.79 on 4 and 45 DF, p-value: 3.779e-11

The nal model that best explains political ideology in the United States is religiosity. Religiosity by far amongst all the variables tests has the strongest correlation.

Harambe

2016-10-05 18:37:09 UTC

Political ideology is a tough subject in the United States in which a number of factors contribute to one’s own political stance. We believe that states with a higher proportion of its citizens with a college education will be less conservative. This is because young and more educated people tend to be more informed and as such be more liberal which less educated citizens tend to be less informed and thus more conservative. States that are more secular will also have less conservative citizens as religion and conservatism go hand in hand. Also, urban areas are generally more liberal as opposed to conservative rural areas. Next, a higher population of black people will also show less conservative support. Finally, states that are located in the South will have a larger portion of its population that is conservative.

Call:
lm(formula = college ~ secularism + urban + blkpct08 + south)

Residuals:
Percent of pop w/college or higher
Min 1Q Median 3Q Max
-10.0477 -2.2484 -0.1455 1.8349 6.3783

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 13.01894 2.77618 4.690 2.57e-05 ***
secularism 0.05768 0.01307 4.413 6.31e-05 ***
urban 0.08447 0.03554 2.377 0.0218 *
blkpct08 0.09662 0.07659 1.261 0.2136
southSouth 0.38787 1.75286 0.221 0.8259
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.339 on 45 degrees of freedom
Multiple R-squared: 0.5002, Adjusted R-squared: 0.4558
F-statistic: 11.26 on 4 and 45 DF, p-value: 2.046e-06

2016-10-05 18:37:32 UTC

Rip haRambe

2016-10-05 18:37:52 UTC

2016-10-05 18:38:25 UTC

2016-10-05 18:39:23 UTC

Code R

2016-10-05 18:39:56 UTC

The eastern south central, western south central, south atlantic and western northern central regions of the United States are highly (~35%-43%) conservative, the middle portions (mountain, eastern north central, mid atlantic, and pacific areas are all moderately conservative (35%-30%), and new england is the least conservative (30-23% conservative). States with higher percent rural population, lower population density, lower per capita income on average, less secularism, and low gunlaw score may best represent the variation of conservatism in the United States. Along the cook index scale the percent of a state's urbanized population increases with liberalism, and decreases with conservatism. Therefore, urban and density (which both communicate information about population concentration) could convey important information in understanding conservative variance in the states. These same areas of concentrated population also tend to result in higher per capita incomes on average; for any urban populous area an increase in per capita income would be witnessed, and a decrease in per capita income would be witnessed in more conservative states. The most conservative regions in the United States typically express sentiments of traditionalism tied to the religious fervor of the Second Great Awakening. This conservative "bible belt" has largely retained these values, so secularism should measure the variance observed between the largest percent conservative areas and the decreasing percent conservative areas. Political ideology is also tied to gun law states, with conservatives typically passing less restrictive gun laws than liberals.

Percent mass public Conservative

Min 1Q Median 3Q Max

-8.3116 -1.6985 0.1916 1.3791 5.7779

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 40.7942369 4.8698262 8.377 1.17e-10 ***

urban -0.0367852 0.0370423 -0.993 0.3261

density -0.0046334 0.0027466 -1.687 0.0987 .

secularism -0.0679743 0.0109628 -6.200 1.71e-07 ***

gunlaw.scale -0.0478012 0.0422661 -1.131 0.2642

prcapinc 0.0001305 0.0001735 0.752 0.4558

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.019 on 44 degrees of freedom

Multiple R-squared: 0.7133, Adjusted R-squared: 0.6807

F-statistic: 21.89 on 5 and 44 DF, p-value: 6.044e-11

The R-

2016-10-05 18:40:06 UTC

The R-gono

2016-10-05 18:40:08 UTC

The R-gonautes

2016-10-05 18:40:10 UTC

What Explains Political Variance

(Paul Budge, Jamie Loos, John Yvanovich)

Variance in political ideology throughout different states in the United States can be explained through varying cultural aspects of each particular society. We’re testing political ideology in the particular states by identifying the amount of conservative support. The degree to which the population in a particular state is college educated likely will suggest that the particular state will be more liberal. We also believe states with a higher degree of religiosity will be more conservative. In addition, presumably if a state is considered southern this will influence the population to be more conservative. We presume these particular variables will have an impact on the political ideology of a state.

=================================================================

Dependent variable:

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

conpct.m

(1) (2)

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

religiosity 0.032 0.058

(0.033) (0.043)

attend.pct 0.157 0.070

(0.174) (0.196)

college -0.372*** -0.338**

(0.133) (0.138)

South -1.523

(1.573)

Constant 40.567*** 46.098***

(9.473) (11.067)

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

Observations 50 50

R2 0.652 0.660

Adjusted R2 0.630 0.629

Residual Std. Error 3.250 (df = 46) 3.253 (df = 45)

F Statistic 28.784*** (df = 3; 46) 21.793*** (df = 4; 45)

=================================================================

Note: *p<0.1; **p<0.05; ***p<0.01

We found that the percentage of the population that is college educated has the biggest impact on the political ideology of the particular state. In accordance with our hypotheses for every percentage increase in college educated individuals their is a -0.372 decrease in individuals identifying as conservative.

Appropriate Names R Advised

2016-10-05 18:40:21 UTC

Shawn, Megan, Connor

A team

2016-10-05 18:40:43 UTC

Vitamin R

2016-10-05 18:41:01 UTC

Screen Shot 2016-10-05 at 12.40.2

In R We Trust - Sasha, Hunter, Liam Hypothesis: Variance in political ideology throughout the states can be explained in a state’s level of religiosity, urban density, geographical location (more southern or more northern). Level of religiosity in a state can be indicative of higher levels of conservatism in a state’s political ideology. Historically, the higher density of urban population has swung left in more liberal ideologies, for example in voting record. And lastly, geographical location, meaning whether or not a state is a southern state or not, can indicate a state’s level of conservatism or liberalism. For example, southern states are typically more conservative and conversely, the opposite for northern states. Original Model:Call:lm(formula = conpct.m ~ south + religiosity + density) Residuals:Min 1Q Median 3Q Max-8.0422 -1.2519 -0.1338 1.2932 6.7436 Coefficients:Estimate Std. Error t value Pr(>|t|)(Intercept) 42.861016 1.526706 28.074 < 2e-16 southSouth -1.057209 1.335816 -0.791 0.432754religiosity 0.074983 0.012473 6.012 2.78e-07 density -0.006508 0.001821 -3.574 0.000839 ---Signif. codes: 0 ‘’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 3.037 on 46 degrees of freedomMultiple R-squared: 0.6967, Adjusted R-squared: 0.6769F-statistic: 35.22 on 3 and 46 DF, p-value: 5.618e-12 Conclusion:-For every one percent increase in density per square kilometer measured corresponds with a .006 percent decrease in conservatism. The coefficient indicates statistical significance between the relationship of density and conservatism in a state. This is due to the fact that areas with higher densities tend to have a more liberal population, meaning an increase in density will lead to an decrease in conservatism. -For every unit increase in religious belief in a state, there is a percent increase of 0.075 in conservatism. Meaning, that the more a state as a whole values religion in their lives, the more conservative the state will be. Additionally, with a coefficient of 0.075, the relationship is statistically significant. This indicates a positive relationship between religiosity and conservatism in a state which thus, in turn, illustrates merit in our initial hypothesis regarding how religiosity might affect conservatism. -For every unit increase in whether a state is regarded as southern or not there is a 1.057 percent decrease in conservatism. You would expect a southern state to be more conservative however the data is not statistically significant and is disregarded. With an R squared value of 0.6769 we are explaining about 68% of the variance in the data.Final Model:Call:lm(formula = conpct.m ~ gunlaw.scale + attend.pct + density)Residuals: Min 1Q Median 3Q Max -7.7392 -1.2570 0.0975 1.3252 6.4749 Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) 22.468118 2.080880 10.797 3.35e-14 ***gunlaw.scale -0.074068 0.036643 -2.021 0.0491 * attend.pct 0.349230 0.047584 7.339 2.83e-09 density -0.004155 0.002508 -1.656 0.1044 ---Signif. codes: 0 ‘’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1Residual standard error: 2.936 on 46 degrees of freedomMultiple R-squared: 0.7164, Adjusted R-squared: 0.6979 F-statistic: 38.74 on 3 and 46 DF, p-value: 1.21e-12In comparison with the original model, we have found that the final model best associates with political ideology in states, and more specifically, levels of conservatism in a state. Measuring for severity of gun laws in a state and its influence on conservatism we found that for every unit increase of gun laws there was a -0.074 percent decrease in conservatism. Looking at church attendance we observe for every percent increase of a population attending church there is a 0.349 percent increase in conservatism. Lastly, with every percent increase in gun laws there is a -0.004 percent decrease in conservatism however the statistical significance of density is no longer noteworthy.

2016-10-05 18:41:23 UTC

Vitamin R

2016-10-05 18:42:05 UTC

Team CNA

2016-10-05 18:51:32 UTC

Ab(sin)the

2016-10-06 21:08:51 UTC

We hypothesize that while political positions are determined by the level of affiliation people have with several different social groups, some of which are significantly more impactful towards how conservative a population is than others. We believe that the religiosity is the most impactful factor towards how conservative a population is.

We decided to test religiosity compared to three other factors: gay.support, urban, and percapinc. We hypothesized that religiosity would have a bigger impact than all of the others. First, we tested religiosity compared to the amount of urban population in a state, and per capita income in a state. In this scenario, religiosity had the greatest impact towards how conservative a population is. For every percentage point religiosity went up, the conservative population variable went up by .064 percentage point. However, when we added gay.support to our regression, religiosity appeared to be significantly less impactful, and not as relevant to how conservative a state is.

Gay support took over in importance when we added it to our regression analysis. Urban population and religiosity loss most if not all of their significance. We figured that in our regression that this could be the case because the openness to lifestyle choices, when it comes to conservative population, is more important than urban population and religiosity. No matter what state you are in we believe that openness to different lifestyle choices is a major determinant on how conservative the population is - despite urban settings having a more liberal scene - and religiosity is often associated with a more closed minded attitude.

We R Great

2016-10-08 02:39:44 UTC

Political Ideology (conpct.m) = f(prcapinc, college, urban,` attend.pct)

We think that a state's political ideology is determined by many factors but the most important we think will be the ones listed in the theoretical model. We chose these four variables in order to limit the amount of variables in the model while having variables that impact the dependent variable (conpct.m) the most. Our expectation is that the church attendance will have the biggest impact on how conservative a state is. Our null hypothesis is political ideology is not statistically significant, but our alternative hypothesis is church attendance has an effect on political ideology. We think this because we know that church and religious affiliation tend to be more of a conservative trait amongst those in a state's population, so having more religious attendees in a state would increase it’s overall conservativeness.

Call:

lm(formula=conpct.m~prcapinc+college+urban+attend.pct)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-7.1835 -1.7916 0.1507 1.5029 7.5559

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 3.447e+01 5.636e+00 6.117 2.10e-07 ***

prcapinc 1.846e-05 1.901e-04 0.097 0.9231

college -3.106e-01 1.721e-01 -1.805 0.0778 .

urban -7.835e-02 3.594e-02 -2.180 0.0345 *

attend.pct 3.224e-01 5.986e-02 5.387 2.51e-06 ***

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.145 on 45 degrees of freedom

Multiple R-squared: 0.6818, Adjusted R-squared: 0.6535

F-statistic: 24.1 on 4 and 45 DF, p-value: 1.06e-10

For every one unit increase in a state’s population church attendance, there is a .3224 percentage point increase in the state’s overall conservativeness. Showing that we can reject our null hypothesis. And our alternative hypothesis is correct.

Call:

lm(formula = conpct.m ~ density + hispanic08 + college)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-9.584 -2.776 -0.280 2.704 7.240

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 50.960029 3.488391 14.608 < 2e-16 ***

density -0.005470 0.002491 -2.196 0.033160 *

hispanic08 -0.103706 0.056443 -1.837 0.072622 .

college -0.577642 0.142878 -4.043 0.000199 ***

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.844 on 46 degrees of freedom

Multiple R-squared: 0.5139, Adjusted R-squared: 0.4822

F-statistic: 16.21 on 3 and 46 DF, p-value: 2.503e-07

For every person that receives a college education, there is a -.5776 percentage point decrease in the conservativeness of that state.

Call:

lm(formula = conpct.m ~ college + prcapinc + attend.pct)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-7.7306 -1.5901 -0.0431 2.0177 8.0723

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 34.6975986 5.8601181 5.921 3.80e-07 ***

college -0.3491172 0.1780393 -1.961 0.056 .

prcapinc -0.0001105 0.0001880 -0.588 0.560

attend.pct 0.3037532 0.0616078 4.930 1.11e-05 ***

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 3.27 on 46 degrees of freedom

Multiple R-squared: 0.6481, Adjusted R-squared: 0.6252

F-statistic: 28.24 on 3 and 46 DF, p-value: 1.65e-10

This theory takes the two most influential variables from the theory. The data shows us that the variable with the biggest impact turns out to be a college education. From this output we see that for every percentage point of the population that receives a college education, there is a .349 percentage point decrease in that state’s overall conservativeness. Compare this to the .303 percentage point increase in conservativeness for every percentage point increase in a state’s populations church attendance.

Ladamit

2016-10-08 03:34:55 UTC

Assignment 4 : Team Ladamit

Jack Freiheit, Danielle Cole, Reid Blankinship

Hypothesis

The United States is an extremely diverse country and has been referred to as the “melting pot” throughout history. In order to find out what variables are most crucial when determining the political ideology of a state, we constructed a multiple regression model. Doing so allows us to understand why some state are more conservative or liberal. We predict that the Southern states will be more conservative than liberal. Additionally, we believe the states with high religious attendance will be more conservative. On the other hand, we predict that the states with mass amounts of college students will be very liberal. Lastly, we believe that the more populated urban areas will also contribute to a state leaning towards the liberal side.

Original Model

lm(formula = conpct.m ~ college + density. + urban + attend.pct +

religiosity + south)

Residuals:

Min 1Q Median 3Q Max

-7.580 -1.540 0.124 1.383 6.790

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 36.851884 11.129225 3.311 0.00189 **

college -0.156822 0.139294 -1.126 0.26648

density -0.004899 0.002253 -2.175 0.03520 *

urban -0.044581 0.036041 -1.237 0.22282

attend.pct 0.208229 0.186522 1.116 0.27046

religiosity 0.027010 0.040946 0.660 0.51300

southSouth -0.546520 1.552634 -0.352 0.72656

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 2.992 on 43 degrees of freedom

Multiple R-squared: 0.7248, Adjusted R-squared: 0.6864

F-statistic: 18.87 on 6 and 43 DF, p-value: 1.289e-10

For the most part, our predictions were accurate. In our first multiple regression we ended up with .72 as the multiple R-squared and .68 for the adjusted R-squared. We correctly predicted that religion would have a great impact on the conservative side. The attendance percentage ended up being .2 and the religiosity was .02. To our surprise, urban and population density did not do as well. Density did the worst with -.004 while urban had -.04. We predicted that population would have an effect because where there is densely populated areas a lot of people will go out to vote. That being said, we decided to take away the variable of density in our next regression model. Replacing this with republican percentage could give us a chance at raising the R-squared. Republicans are conservative in their political beliefs while Democrats are mainly liberal.

Final Model

lm(formula = conpct.m ~ college + reppct.m + urban + attend.pct +

religiosity + south)

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-6.2699 -1.4202 -0.2268 1.2774 6.6993

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 29.33936 10.00378 2.933 0.005366 **

college -0.17806 0.12166 -1.464 0.150576

reppct.m 0.29662 0.07197 4.121 0.000168 ***

urban -0.07668 0.02931 -2.616 0.012214 *

attend.pct 0.17189 0.16216 1.060 0.295079

religiosity 0.01523 0.03616 0.421 0.675724

southSouth 0.80279 1.43447 0.560 0.578628

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 2.669 on 43 degrees of freedom

Multiple R-squared: 0.781, Adjusted R-squared: 0.7505

F-statistic: 25.56 on 6 and 43 DF, p-value: 1.09e-12

Conclusion

After having initial success with our first multiple regression, we decided that it makes sense to eliminate our worst variable. Doing so helped make our multiple R-squared and adjusted R squared numbers rise. We took out density of population and changed it with republican percentage. This move made sense because this variable is directly related with being a conservative. Democrats generally lean to the liberal side wanting more government involvement with daily life, while republicans favor a more conservative government and lower taxes. In the end making this changed help us a lot in trying to find out how to determine state’s political ideology. Republican percentage ended at .3 while South had .8. The R-squared rose to .78 which was .6 higher than our first regression. The adjusted R squared ended up being .75, respectively rising by .7. While it is hard to determine what causes a state to be more liberal or more conservative our group can infer that the Southern states, religious states and states with high republican percentages will all be more conservative.

2016-10-10 17:02:54 UTC

RCH

Ryan, Carly, Harrison

Hypothesis: Our group believed that religiosity would be the most important variable in determining whether a state was more conservative or not. A state that is more religious is going to be more likely to have values that coincide with conservatism. Thus, we believe there is going to be a positive correlation between religiosity and conservative ideology.

Residuals:

Percent mass public Conservative

Min 1Q Median 3Q Max

-7.580 -1.540 0.124 1.383 6.790

Coefficients:

Estimate Std. Error t value Pr(>|t|)

(Intercept) 36.851884 11.129225 3.311 0.00189 **

college -0.156822 0.139294 -1.126 0.26648

urban -0.044581 0.036041 -1.237 0.22282

attend.pct 0.208229 0.186522 1.116 0.27046

density -0.004899 0.002253 -2.175 0.03520 *

religiosity 0.027010 0.040946 0.660 0.51300

southSouth -0.546520 1.552634 -0.352 0.72656

---

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 2.992 on 43 degrees of freedom

Multiple R-squared: 0.7248, Adjusted R-squared: 0.6864

F-statistic: 18.87 on 6 and 43 DF, p-value: 1.289e-10

Conclusion: After observing our multiple regressions we determined that church attendance had a more positive correlation with ideology than religiosity. For every 1% increase in percentage of a state’s population that attends church there is .21% increase in mass conservatism of a state. This ratio remained the strongest or got stronger after manipulating our variables in all of our models. Noting that for every 1-point increase on the religion observance belief scale there was only a .027% increase in mass public conservatism of a state.

We R Buffs

2016-10-12 18:43:53 UTC