| title: “Companion: Chapter 4” |
| output: html_document |
Q1
(Datasets: nes, nesD. Variables: fedspend_scale, pid_x, wt.) Here is a widely observed difference between Democrats and Republicans: Democrats are more likely than Republicans to favor government spending. Is this difference borne out by the data? Dataset nes contains the numeric variable fedspend_scale, which measures respondents’ spending opinions on a scale that runs from 0 (oppose government spending) to 16 (support government spending). This is the dependent variable. Use the 7-point party identification scale (pid_x) as the independent variable.
Q1, Part A
A : a higher mean than do Republicans on fedspend_scale
Q1, Part B
compmeans(nes$fedspend_scale, nes$pid_x, nes$wt)
## Warning in descr::compmeans(...): 231 rows with missing values dropped
## Mean value of "nes$fedspend_scale" according to "nes$pid_x"
## Mean N Std. Dev.
## StrDem 12.077649 1126 2.473178
## WkDem 10.890360 863 2.568933
## IndDem 11.233952 659 2.312933
## Ind 9.873966 790 3.349926
## IndRep 8.055172 689 3.409352
## WkRep 8.673046 713 3.156444
## StrRep 7.441923 830 3.229499
## Total 9.896405 5670 3.372351
Q1, Part C
Yes. The mean values of the dependent variable decline across the values of the independent variable, from 12.08 among Strong Democrats to 7.44 among Strong Republicans.
Q1, Part D
The mean value of fedspend_scale does not decline consistently from one pid_x value to the next.Independent Democrats have a higher mean score than Weak Democrats and Weak Republicans have a higher mean than Independent Republicans.
Q1, Part E
svyboxplot(fedspend_scale ~ pid_x, design=nesD,
ylab="Pro Spending Scale Score",
xlab="Party ID",col="lightgray", varwidth=T)
Q1, Part F
No. The claim about cohesiveness is incorrect. The interquartile range for Strong Republicans is from 11 to 14, a range of 3. The interquartile range for Strong Republicans ranges from 5 to 10, a range of 5.
Q1, Part G
plotmeansC(nes, ~fedspend_scale, ~pid_x, fedspend_scale ~pid_x, w=~wt,
ylab="Pro Spending Scale Score",xlab="Party ID")
Q2
(Dataset: nes. Variables: gay_marry, libcon3, wt.) Should gay couples be legally permitted to marry? This controversial issue has gained center stage in political debate. One can imagine several characteristics that divide people on this issue. For example, one could hypothesize that self-described liberals are more likely to favor gay marriage than are self-described moderates, who in turn are more likely to favor gay marriage than are self-described conservatives.
Q2, Parts A and B
xtp(nes, gay_marry, libcon3, wt, xlab="Ideology",
ylab="Opinion on Gay Marriage",
main="Gay Marriage Opinions, by Ideology")
## Cell Contents
## |-------------------------|
## | Count |
## | Column Percent |
## |-------------------------|
##
## ===========================================================
## Ideology
## Opinion on Gay Marriage Lib Mod Cons Total
## -----------------------------------------------------------
## No 611 617 2153 3381
## 32.64% 56.50% 78.06%
## -----------------------------------------------------------
## Yes 1261 475 605 2341
## 67.36% 43.50% 21.94%
## -----------------------------------------------------------
## Total 1872 1092 2758 5722
## 32.72% 19.08% 48.20%
## ===========================================================
Q2, Part C
Yes. According to the NES data, 67.4% of liberals support gay marriage, while only 43.5% of moderates do. Conservatives express the least support, 21.9%.
Q2, Part D
nes$gay_marry.yes = as.numeric(nes$gay_marry=="Yes")*100
plotmeansC(nes, ~gay_marry.yes, ~libcon3, gay_marry.yes ~libcon3, w=~wt,
ylab="Pecentage Supporting Gay Marriage",xlab="Ideology",
main="Gay Marriage Opinions, by Ideology")
Q3
(Dataset: gss. Variables: educ_4, mslm_spk, wtss.) Are people who have more education more likely than the less educated to be tolerant of radical Muslim clerics? The hypothesis: In a comparison of individuals, those having higher levels of education are more tolerant of radical Muslims than are those having lower levels of education.
Q3, Part A
xtp(gss, mslm_spk, educ_4, wtss,
ylab="Allow Radical Muslim Cleric Speaker?",
xlab="Education Level")
## Cell Contents
## |-------------------------|
## | Count |
## | Column Percent |
## |-------------------------|
##
## ====================================================================================
## Education Level
## Allow Radical Muslim Cleric Speaker? <HS HS Some Coll Coll+ Total
## ------------------------------------------------------------------------------------
## No 153 234 198 139 724
## 77.27% 65.73% 57.39% 36.58%
## ------------------------------------------------------------------------------------
## Yes 45 122 147 241 555
## 22.73% 34.27% 42.61% 63.42%
## ------------------------------------------------------------------------------------
## Total 198 356 345 380 1279
## 15.48% 27.83% 26.97% 29.71%
## ====================================================================================
Q3, Part B
printC(xtp(gss, mslm_spk, educ_4, wtss,
ylab="Allow Radical Muslim Cleric Speaker?",
xlab="Education Level"))
## Table appended to Table.Output.html in D:/ASUJ/MOD4
Q3, Part C
A supporting graph would look something like the following:
Q3, Part D
All of the evidence supports the hypothesis that individuals with higher levels of education have higher levels of tolerance toward radical Muslim clergy than do individuals with lower levels of education. Each time education goes up by one ordinal category—from 0-11 years to 12 years, from 12 years to 13-15 years, from 13-15 years to 16-plus years—the percentage supporting a radical Muslim cleric speaker increases: 22.7 percent, 34.3 percent, 42.6 percent, and 63.4 percent.
Q4
(Dataset: world. Variables: regime_type3, durable.) Three comparative politics scholars are trying to figure out what sort of institutional arrangement produces the longest lasting, most stable political system. Scholar 1: “Presidential democracies, like the United States, are going to be more stable than are any other type of system. In presidential democracies, the executive and the legislature have separate electoral constituencies and separate but overlapping domains of responsibility. The people’s political interests are represented both by the president’s national constituency and by legislators’ or parliament members’ more localized constituencies. If one branch does something that’s unpopular, it can be blocked by the other branch. The result: political stability.” Scholar 2: “Parliamentary democracies are by far more stable than presidential democracies. In presidential systems, the executive and legislature can be controlled by different political parties, a situation that produces deadlock. Since the leaders of the legislature can’t remove the president and install a more compliant or agreeable executive, they are liable to resort to a coup, toppling the whole system. Parliamentary democracies avoid these pitfalls. In parliamentary democracies, all legitimacy and accountability resides with the legislature. The parliament organizes the government and chooses the executive, the prime minister, from among its own leaders. The prime minister and members of parliament have strong incentives to cooperate and keep things running smoothly and efficiently. The result: political stability.” Scholar 3: “You two have made such compelling—if incorrect—arguments that I almost hesitate to point this out: Democracies of any species, presidential or parliamentary, are inherently unstable. Any system that permits the clamor of competing parties or dissident viewpoints is surely bound to fail. If it’s stability that you value above all else, then dictatorships will deliver. Strong executives, feckless or nonexistent legislatures, powerful armies, social control. The result: political stability.” The world dataset contains the variable durable, a numeric that measures the number of years since the last regime transition. The more years that have passed since the system last failed (higher values on durable), the more stable a country’s political system. The variable regime_type3 captures system type: dictatorship, parliamentary democracy, or presidential democracy. Run compmeans to analyze the relationship between durable and regime_type3. Perform a mean comparison analysis
Q4, Part A
Parliamentary democracies (most stable), presidential democracies, dictatorships (least stable)
Q4, Part B
printC(compmeans(world$durable, world$regime_type3))
## Warning in descr::compmeans(...): 38 rows with missing values dropped
## Table appended to Table.Output.html in D:/ASUJ/MOD4
Q4, Part C
stripchart(durable ~ regime_type3, data=world, vertical=T,
method="jitter")
svyboxplot(durable ~ regime_type3, worldD)
Q4, Part D
The strip chart shows two presidential democracies with unusually long durations, which exerts an upward pull on the mean value. The mean comparison table suggests this (standard deviation, 43.4), but the strip chart displays it visually. Student might also comment on the considerable overlap of dictatorships and parliamentary democracies. Among parliamentary democracies, there are two distinct clusters, one a very short durations and another at around 50 years.
Q4, Part E
The graphs were printed as part of the analysis to Part C.