Textbook - A Stata Companion to Political Analysis, 3rd Edition by Philip H. Pollock
Chapter 2 - Descriptive Statistics Exercises - Complete exercises 1, 2, 3, 5 & 6.
Q1. (Dataset: gss2012. Variables: femrole, [aw=gssw], [fw=gssw_rnd].) Two pundits are arguing about how the general public views the role of women in the home and in politics.
Pundit 1: "Our society has a minority of traditionally minded individuals who think that the proper 'place' for women is taking care of the home and caring for children. This small but vocal group of traditionalists aside, the typical adult supports the idea that women belong in work and in politics."
Pundit 2: "Poppycock! It's just the opposite. The extremist feminist crowd has distorted the overall picture. The typical view among most citizens is that women should be in the home, not in work and politics."
1. Dataset gss2012 contains femrole, an interval-level variable that measures respondents' attitudes toward women in society and politics. Scores can range from 0 (women belong in the home) to 9 (women belong in work and politics).
If Pundit 1 is correct, femrole will have (circle one)
a negative skew.
no skew.
a positive skew.
If Pundit 2 is correct, femrole will have (circle one)
a negative skew.
no skew.
a positive skew.
If Pundit 1 is correct, femrole's mean will be (circle one)
lower than its median.
the same as its median.
higher than its median.
If Pundit 2 is correct, femrole's mean will be (circle one)
lower than its median.
the same as its median.
higher than its median.
2. Run "tab femrole [aw=gssw]." Run "sum femrole [aw=gssw], detail." Fill in the table that follows.
3. Obtain a bar chart of femrole by running histogram with the "d percent" options. (Even though we are treating femrole as an interval variable, the "d" option works nicely for interval variables having fewer than thirty values.) Don't forget to specify "[fw=gssw_rnd]." Make the x-axis range from 0 to 9 in increments of 1. Request "valuelabel." Ask for a font size of "small" or "medsmall." Override the default bar fill color with a color of your choice. Print the bar chart.
4. Consider the evidence you obtained in parts B and C. Based on your analysis, whose assessment is more accurate? (circle one)
Pundit 1's
Pundit 2's
Citing specific evidence obtained in parts B and C, explain your reasoning.
Q2. (Dataset: gss2012. Variables: attend, [aw=gssw], [fw=gssw_rnd].) The General Social Survey provides a rich array of variables that permit scholars to study religiosity in the adult population. Dataset gss2012 contains attend, a 9-point ordinal scale that measures how often respondents attend religious services. Values can range from 1 ("Never attend") to 9 ("Attend more than once a week").
1. The shell of a bar chart is given below. The categories of attend appear along the horizontal axis. What would a bar chart of attend look like if this variable had maximum dispersion? Sketch inside the axes a bar chart that would depict maximum dispersion.
2. What would a bar chart of attend look like if this variable had no dispersion? Sketch inside the axes a bar chart that would depict no dispersion.
3. Perform a tabulate analysis of attend. Complete the following table.
|
Services
|
Freq.*
|
Percent
|
Cum.
|
|
Never
|
496.97
|
?
|
?
|
|
|
97.77
|
?
|
?
|
|
Once/yr
|
255.93
|
?
|
?
|
|
Sev times/yr
|
212.61
|
?
|
?
|
|
Once/mo
|
133.36
|
?
|
?
|
|
2-3 times/mo
|
174.08
|
?
|
?
|
|
Nrly every wk
|
79.34
|
?
|
?
|
|
Every wk
|
388.04
|
?
|
?
|
|
>Once/wk
|
127.90
|
?
|
100.00
|
|
Total
|
1,966
|
100.00
|
|
Weighted frequencies rounded to two decimal places.
4. Obtain and print a bar chart of attend. (Remember that attend is a discrete variable.) Be sure to specify "[fw=gssw_rnd]." Make the x-axis range from 1 to 9 in increments of 1. Request "valuelabel." Ask for a font size of "small" and an angle of 45 degrees. Override the default bar fill color.
5. Based on your examination of the frequency distribution,
the mode of attend is ______________.
the median of attend is _____________.
6. Based on your examination of the frequency distribution and bar chart, you would conclude that attend has (circle one)
low dispersion.
high dispersion.
Q3. (Data set: gss2012. Variables: science_quiz, [aw=gssw], [fw=gssw_rnd].) The late Carl Sagan once lamented, "We live in a society exquisitely dependent on science and technology, in which hardly anyone knows anything about science and technology." This is a rather pessimistic assessment of the scientific acumen of ordinary Americans. Sagan seemed to be suggesting that the average level of scientific knowledge is quite low and that most people would fail even the simplest test of scientific facts.
Dataset gss2012 contains science_quiz, which was created from 10 true-false questions testing respondents' knowledge of basic scientific facts. Values on science_quiz range from 0 (the respondent did not answer any of the questions correctly) to 10 (the respondent correctly answered all 10).
1. Obtain a frequency distribution of science_quiz. Fill in the table that follows.
|
science_quiz
|
Freq.*
|
Percent
|
Cum.
|
|
0
|
6.86
|
?
|
?
|
|
1
|
13.15
|
?
|
?
|
|
2
|
18.01
|
?
|
?
|
|
3
|
37.28
|
?
|
?
|
|
4
|
69.69
|
?
|
?
|
|
5
|
67.05
|
?
|
?
|
|
6
|
65.81
|
?
|
?
|
|
7
|
91.97
|
?
|
?
|
|
8
|
53.98
|
?
|
?
|
|
9
|
56.03
|
?
|
?
|
|
10
|
28.17
|
?
|
100.0
|
|
Total
|
508
|
100.0
|
|
Weighted frequencies rounded to two decimal places
2. Run summary with the detail option.
3. Use histogram to create a bar chart for science_quiz. Specify appropriate parameters for graphing options fcolor and xlabel. Print the bar chart.
4. Examine the frequency distribution, the summary statistics, and the bar chart. Based on your analysis, science_quiz has a mean equal to _________, a median equal to ______________, and a skewness equal to _______________.
5. Exercise your judgment. What would be the more accurate measure of science_quiz's central tendency: the mean or the median? (circle one)
mean
median
6. Briefly explain your choice in E.
7. According to conventional academic standards, any science_quiz score of 5 or lower would be an F, a failing grade. A score of 6 would be a grade of D, a 7 would be a C, an 8 a B, and scores of 9 or 10 would be an A. Based on these standards, about what percentage of people got passing grades on science_quiz? (circle one)
About 30 percent
About 40 percent
About 50 percent
About 60 percent
What percentage got a B or better? (circle one)
About 30 percent
About 40 percent
About 50 percent
About 60 percent
Q4. (Dataset: states. Variables: defexpen, state.) Here is the conventional political wisdom: Well-positioned members of Congress from a handful of states are successful in getting the federal government to spend revenue in their states-defense-related expenditures, for example. The typical state, by contrast, receives far fewer defense budget dollars.
1. Suppose you had a variable that measured the amount of defense-related expenditures in each state. The conventional wisdom says that, when you look at how all 50 states are distributed on this variable, a few states would have a high amount of defense spending. Most states, however, would have lower values on this variable.
If the conventional wisdom is correct, the distribution of defense-related expenditures will have (circle one)
a negative skew.
no skew.
a positive skew.
If the conventional wisdom is correct, the mean of defense-related expenditures will be (circle one)
lower than its median.
the same as its median.
higher than its median.
2. Dataset states contains the variable defexpen, defense expenditures per capita for each of the 50 states. Perform a summary analysis of defexpen (with detail).
3. Which is the better measure of central tendency? (circle one)
Mean
median
Briefly explain your answer.
4. Obtain a histogram of defexpen. Because defexpen is an interval-level (continuous) variable, specify only the percent option. Request x-axis labels that range from 0 to 4500 in increments of 500. Print the histogram.
5. Based on your analysis, would you say that the conventional wisdom is accurate or inaccurate? (check one)
The conventional wisdom is accurate.
The conventional wisdom is inaccurate.
6. Use the sort command and the list command to obtain a ranked list of states, from lowest per capita defense spending to highest per capita defense spending. The state with the lowest per capita defense spending is ____________, with $_____ per capita. The state with the highest per capita defense spending is _______________, with $_____ per capita.
Q5. (Dataset: states. Variables: blkpct10 hispanic10.) Two demographers are arguing over how best to describe the racial and ethnic composition of the "typical" state.
Demographer 1: "The typical state is 8.25 percent black and 8.20 percent Hispanic."
Demographer 2: "The typical state is 11.26 percent black and 10.61 percent Hispanic."
1. Run summary (with detail) for blkpct10 (the percentage of each state's population that is African American) and hispanic10 (the percentage of each state's population that is Hispanic). (Hint: Stata will permit you to name more than one variable after the sum command: "sum blkpct10 hispanic10, detail.") Record the appropriate statistics for each variable in the table that follows.
2. Based on your analysis, which demographer is more accurate? (circle one)
Demographer 1
Demographer 2
Write a few sentences explaining your reasoning
3. Run sort and list to obtain information on the percentage of Hispanics.
Which five states have the lowest percentages of Hispanics?
Which five states have the highest percentages of Hispanics?
Chapter 3 Transforming Variables Exercises - Complete all the exercises
For the exercises in this chapter, you will use gss2012.dta.
Q1. (Dataset: gss2012. Variables: polviews, [aw=gssw].) Dataset gss2012 contains polviews, which measures political ideology-the extent to which individuals "think of themselves as liberal or conservative." Here is how polviews is coded:
|
Numeric code
|
Value label
|
|
1
|
ExtrmLib
|
|
2
|
Liberal
|
|
3
|
SlghtLib
|
|
4
|
Moderate
|
|
5
|
SlghtCons
|
|
6
|
Conserv
|
|
7
|
ExtrmCons
|
1. Run tab on polviews, making sure to use the weight variable, gssw. Eyeball the percent column, and make some rough-and-ready estimates.
The percentage of respondents who are either "extremely liberal," "liberal," or "slightly liberal" is (circle one)
about 17 percent.
about 27 percent.
about 37 percent.
The percentage of respondents who are either "slightly conservative," "conservative," or "extremely conservative" is (circle one)
about 15 percent.
about 25 percent.
about 35 percent.
2. (i) Use recode to create a new variable named polview3. Collapse the three "liberal" codes into one category (coded 1 on polview3), put the "moderates" into their own category (coded 2 on polview3), and collapse the three "conservative" codes into one category (coded 3 on polview3). (Don't forget to recode missing values on polviews into missing values on polview3.) (ii) Using the label var command, give polview3 this label: "Ideology: 3 categories." (iii) Run tab on polview3.
The percentage of respondents who are coded 1 on polview3 is (circle one)
about 17 percent.
about 27 percent.
about 37 percent.
The percentage of respondents who are coded 3 on polview3 is (circle one)
about 15 percent.
about 25 percent.
about 35 percent.
3. (i) Run label define to create a label named "polview3_label." In the label define command, connect the following numeric codes and value labels: 1 "Lib" 2 "Mod" 3 "Cons." (ii) Run label values to label the values of polview3 using polview3_label. (iii) Run tab on polview3. Based on your findings, fill in this table.
Ideology: Three Categories
|
|
Frequency*
|
Percent
|
Cumulative Percent
|
|
Lib
|
|
|
|
|
Mod
|
|
|
|
|
Cons
|
|
|
100.0
|
|
Total
|
|
100.0
|
|
Round weighted frequencies to two decimal places.
2. (Dataset: gss2012. Variables: Variables: mslm_col2, mslm_lib2, mslm_spk2, [aw=gssw].) Dataset gss2012 contains three variables that gauge tolerance toward "anti-American Muslim clergymen"-whether they should be allowed to teach in college (mslm_col2), whether their books should be removed from the library (mslm_lib2), and whether they should be allowed to preach hatred of the United States (mslm_spk2). For each variable, a less-tolerant response is coded 0, and a more-tolerant response is coded 1.
1. Imagine creating an additive index from these three variables. The additive index would have scores that range between a score of ________________ and a score of ________________.
2. Suppose a respondent takes the more-tolerant position on two questions and the less-tolerant position on the third question. This respondent would have a score of
3. Use generate to create an additive index from mslm_col2, mslm_lib2, and mslm_spk2. Name the new variable muslim_tol. Run tab on muslim_tol. Referring to your output, fill in the table that follows.
Round weighted frequencies to two decimal places.
4. In this chapter, you learned to use generate with the recode function. Recall that you run this command by telling Stata, in the recode expression, the upper boundaries of the range of codes you want to combine. (i) Use generate with the recode function to create a new variable, muslim_tol3, from muslim_tol. Keep the least tolerant group in one category (upper boundary, 0), combine the two middle categories (upper boundary, 2), and keep the most tolerant group in one category (upper boundary, 3). (ii) Run tab on muslim_tol3 to make sure that the percentages check out.
5. (i) Label muslim_tol3 with the following label: "Tolerance twrd Muslim clergy." (ii) Run label define to create the label, muslimtol3_label, using this syntax: label define muslimtol3_label 0 "Low" 2 "Mid" 3 "High." (iii) Run label values to label the values of muslim_tol3 using muslimtol3_label. (iv) Run tab on muslim_tol3. Print the frequency table.
Q3. (Dataset: gss2012. Variables: rincom06, [aw=gssw].) In this chapter you learned to use the xtile command to collapse an nes2012 measure of income into three roughly equal ordinal categories. In this exercise you will use xtile to collapse a similar variable from gss2012, rincom06. You will collapse rincom06 into rincom06_3, a three-category ordinal measure of respondents' incomes.
1. For guidance, refer back to this chapter's "A Closer Look" box on pages 42-43. Run xtile, using rincom06 to create rincom06_3. Run tabulate on rincom06_3. Referring to the Results window, fill in the numbers next to the question marks:
|
3 quantiles of rincom06
|
Frequency*
|
Percent
|
Cumulative Percent
|
|
1
|
399.15
|
?
|
?
|
|
2
|
426.84
|
?
|
?
|
|
3
|
316.01
|
?
|
100.00
|
|
Total
|
1,142
|
100.00
|
|
Weighted frequencies are rounded to two decimal places.
2. (i) Run label define to create a label, rincom6_label. In the label define command, connect the following numeric codes and value labels: 1 "Low" 2 "Mid" 3 "High." (ii) Run label values to label of the values of rincom06_3 using rincom_label.
Q4. (Dataset: gss2012. Variables: pornlaw2, [aw=gssw].) In this chapter you learned to use tabulate (with the generate option) to create indicator variables. In this exercise, you will create indicator variables from pornlaw2, which measures individuals' opinions about pornography. Respondents thinking pornography should be "Illegal to all" are coded 1, and those saying "Not illegal to all" are coded 2. You will create an indicator variable coded 1 for individuals saying "Illegal to all," and coded 0 for any other response.
1. Run tab on pornlaw2. Make sure to include the weight variable, gssw. The percentage of respondents saying "Illegal to all" is __________________.
2. Run tabulate on pornlaw2 with the generate option. In the generate option, use "porn" as the name stem. Run tab on porn1. The percentage of respondents coded 1 on porn1 is _____________.
3. Use drop to delete porn2 from the dataset. Run aorder to alphabetize the variables in gss2012.
Before exiting Stata, be sure to save the dataset.
Assignment link - https://www.dropbox.com/s/pewqshz7oylv26w/Assignment.rar?dl=0.