5 Stata basics
5.1 Stata filetypes
Stat has a few types of files that we will work with:
a dta file is a Stata dataset, which can include numbers and labels for the same observation. So, for example, an observation could be coded both as the number 1 and with the label “support”, and another observation could be coded both as the number 0 and with the label “does not support”. The numbers can be used to calculate the percentage of participants who support the outcome, and the labels are useful for know what the numbers refer to.
a do file that can be used to save a set of commands and documentation indicating, for instance, what the commands are intended to do.
a log file that records the commands that are run and the output that occurs from the commands. For a log file, the smcl file type (Stata Markup and Control Language) is in Stata format, but – unless saved as a PDF or something equivalent – an smcl file cannot be read without access to Stata. The log type of log file is a text file that can be opened with a text editor that all or almost all computers should have.
5.2 Stata data editor for cross-sectional data
Many social science datasets have only one observation for each person, country, or other thing in the dataset, with this observation taken at or around the same time. In this case, the dataset is typically st up with one observation per row. Below is an example, from the dataset for the ANES 2024 Pilot Study, in which each row is a different participant. In the Stata data editor, the red cells indicate text only cells, the black cells indicate numeric only cells, and the blue cells indicate numeric cells that have a number and an associated text label. For example, the “Partial data” label has the number 3 associated with it.
Datasets get more complicated when there are multiple observations per each person, country, or other thing. But let’s worry about that later.
5.3 Useful Stata commands
| Command | Use |
|---|---|
| edit | Open the data editor |
| lookfor | Look for text in variable names or labels |
| codebook | Get information about a variable |
| tab | Tabulate results for a variable w/options such as mi and nol |
| tabstat | Tabulate summary statistics |
| sum | Get summary statistics |
| recode | Recode a variable |
| replace | Replace levels of a variable |
| gen | Generate a new variable |
| egen | Generate a new variable using a specialized function |
| clonevar | Clone an existing variable |
| label | Label a variable (see label define, label values) |
| set cformat | Change the number of decimal places reported in regressions |
| reg | OLS linear regression |
| margins | Predicted outcomes from a regression |
| marginsplot | Plot of predicted outcomes from the margins command |
| svyset | Set up survey weights |
| svy: | Prefix for survey commands |
| logit | Regression technique for dichotomous outcomes |
Stata tips by Todd Jones
5.4 Sample Stata commands to read in data
One way to read data into Stata is to select File > Open… in the top menu and to then select the Stata file that you want to read into Stata. Stata files end with a .dta extension. You can also read in CSV and other files, using File > Import in the top menu. Or you can type “edit” into the Stata command line and then paste the data from a spreadsheet into the Stata data editor.
Let’s read in the Stata dataset for the ANES 2020 Time Series Study:
5.5 Sample Stata commands to find data
Let’s find the measure of how respondents rated police, by using lookfor to search for “police” in the variable names and variable labels but not in the variable values:
lookfor police
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
storage display value
variable name type format label variable label
-------------------------------------------------------------------------------
V202171 double %12.0g V202171 POST: Feeling thermometer: police
V202351 double %12.0g V202351 POST: How often do police
officers use more force than
necessary
V202456 double %12.0g V202456 POST: During past 12 months, R or
any family members stopped by
police
V202491 double %12.0g V202491 POST: Do police treat blacks or
whites better
V202492 double %12.0g V202492 POST: How much better do police
treat blacks or whites
V202493x double %12.0g V202493x POST: SUMMARY: Police treat
blacks or whites better5.6 Sample Stata commands to describe data
Let’s get information about the V202171 measure of how respondents rated police on a 0-to-100 feeling thermometer. First, let’s use the describe command:
desc V202171
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
storage display value
variable name type format label variable label
-------------------------------------------------------------------------------
V202171 double %12.0g V202171 POST: Feeling thermometer: policeLet’s use the codebook command:
codebook V202171
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
-------------------------------------------------------------------------------
V202171 POST: Feeling thermometer: police
-------------------------------------------------------------------------------
type: numeric (double)
label: V202171, but 57 nonmissing values are not labeled
range: [-9,100] units: 1
unique values: 62 missing .: 0/8,280
examples: 40
60
85
90 Let’s use the tab tabulate command:
tab V202171
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
POST: Feeling thermometer: police | Freq. Percent Cum.
----------------------------------------+-----------------------------------
-9. Refused | 46 0.56 0.56
-7. No post-election data, deleted due | 77 0.93 1.49
-6. No post-election interview | 754 9.11 10.59
-5. Interview breakoff (sufficient part | 14 0.17 10.76
-4. Technical error | 1 0.01 10.77
0 | 190 2.29 13.07
1 | 6 0.07 13.14
2 | 2 0.02 13.16
3 | 3 0.04 13.20
4 | 2 0.02 13.22
5 | 22 0.27 13.49
7 | 3 0.04 13.53
8 | 3 0.04 13.56
10 | 17 0.21 13.77
12 | 1 0.01 13.78
15 | 150 1.81 15.59
17 | 1 0.01 15.60
20 | 26 0.31 15.92
24 | 1 0.01 15.93
25 | 19 0.23 16.16
30 | 244 2.95 19.11
33 | 1 0.01 19.12
35 | 21 0.25 19.37
40 | 449 5.42 24.79
41 | 1 0.01 24.81
43 | 1 0.01 24.82
45 | 39 0.47 25.29
49 | 5 0.06 25.35
50 | 513 6.20 31.55
51 | 3 0.04 31.58
54 | 1 0.01 31.59
55 | 33 0.40 31.99
60 | 755 9.12 41.11
61 | 1 0.01 41.12
62 | 1 0.01 41.14
63 | 1 0.01 41.15
65 | 67 0.81 41.96
66 | 1 0.01 41.97
67 | 2 0.02 41.99
68 | 2 0.02 42.02
70 | 1,125 13.59 55.60
72 | 1 0.01 55.62
73 | 1 0.01 55.63
75 | 154 1.86 57.49
78 | 3 0.04 57.52
79 | 2 0.02 57.55
80 | 168 2.03 59.58
83 | 1 0.01 59.59
84 | 1 0.01 59.60
85 | 1,593 19.24 78.84
86 | 12 0.14 78.99
87 | 2 0.02 79.01
88 | 4 0.05 79.06
89 | 2 0.02 79.08
90 | 241 2.91 81.99
93 | 1 0.01 82.00
94 | 1 0.01 82.02
95 | 109 1.32 83.33
97 | 1 0.01 83.35
98 | 9 0.11 83.45
99 | 20 0.24 83.70
100 | 1,350 16.30 100.00
----------------------------------------+-----------------------------------
Total | 8,280 100.00Let’s use the tab command, with the nol and mi options (for “no labels” and “missing”, respectively):
tab V202171, nol mi
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
POST: |
Feeling |
thermometer |
: police | Freq. Percent Cum.
------------+-----------------------------------
-9 | 46 0.56 0.56
-7 | 77 0.93 1.49
-6 | 754 9.11 10.59
-5 | 14 0.17 10.76
-4 | 1 0.01 10.77
0 | 190 2.29 13.07
1 | 6 0.07 13.14
2 | 2 0.02 13.16
3 | 3 0.04 13.20
4 | 2 0.02 13.22
5 | 22 0.27 13.49
7 | 3 0.04 13.53
8 | 3 0.04 13.56
10 | 17 0.21 13.77
12 | 1 0.01 13.78
15 | 150 1.81 15.59
17 | 1 0.01 15.60
20 | 26 0.31 15.92
24 | 1 0.01 15.93
25 | 19 0.23 16.16
30 | 244 2.95 19.11
33 | 1 0.01 19.12
35 | 21 0.25 19.37
40 | 449 5.42 24.79
41 | 1 0.01 24.81
43 | 1 0.01 24.82
45 | 39 0.47 25.29
49 | 5 0.06 25.35
50 | 513 6.20 31.55
51 | 3 0.04 31.58
54 | 1 0.01 31.59
55 | 33 0.40 31.99
60 | 755 9.12 41.11
61 | 1 0.01 41.12
62 | 1 0.01 41.14
63 | 1 0.01 41.15
65 | 67 0.81 41.96
66 | 1 0.01 41.97
67 | 2 0.02 41.99
68 | 2 0.02 42.02
70 | 1,125 13.59 55.60
72 | 1 0.01 55.62
73 | 1 0.01 55.63
75 | 154 1.86 57.49
78 | 3 0.04 57.52
79 | 2 0.02 57.55
80 | 168 2.03 59.58
83 | 1 0.01 59.59
84 | 1 0.01 59.60
85 | 1,593 19.24 78.84
86 | 12 0.14 78.99
87 | 2 0.02 79.01
88 | 4 0.05 79.06
89 | 2 0.02 79.08
90 | 241 2.91 81.99
93 | 1 0.01 82.00
94 | 1 0.01 82.02
95 | 109 1.32 83.33
97 | 1 0.01 83.35
98 | 9 0.11 83.45
99 | 20 0.24 83.70
100 | 1,350 16.30 100.00
------------+-----------------------------------
Total | 8,280 100.00Let’s use the sum summary command:
sum V202171
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Variable | Obs Mean Std. Dev. Min Max
-------------+---------------------------------------------------------
V202171 | 8,280 62.30145 33.62062 -9 100Let’s use the sum command, with the de “describe” option:
sum V202171, de
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
POST: Feeling thermometer: police
-------------------------------------------------------------
Percentiles Smallest
1% -7 -9
5% -6 -9
10% -6 -9 Obs 8,280
25% 45 -9 Sum of Wgt. 8,280
50% 70 Mean 62.30145
Largest Std. Dev. 33.62062
75% 85 100
90% 100 100 Variance 1130.346
95% 100 100 Skewness -.8707428
99% 100 100 Kurtosis 2.6402255.7 Sample Stata commands to manipulate data
Let’s use recode to code a new variable called FTPOLICE that is the measure of how respondents rated police but with missing values of the measure coded as missing data (which in Stata is a period):
recode V202171 (-100/-1 101/999 = .), gen(FTPOLICE)
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
(892 differences between V202171 and FTPOLICE)Let’s check the minimum value and maximum value of FTPOLICE, to make sure that FTPOLICE ranges from 0 to 100:
sum FTPOLICE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Variable | Obs Mean Std. Dev. Min Max
-------------+---------------------------------------------------------
FTPOLICE | 7,388 70.57485 25.12587 0 100Let’s find the measure of respondent age:
lookfor age
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
storage display value
variable name type format label variable label
-------------------------------------------------------------------------------
V201001 double %12.0g V201001 PRE: Select language
V201117 double %12.0g V201117 PRE: How outraged R feels about
how things are going in the
country
V201416 double %12.0g V201416 PRE: R position on gay marriage
V201507x double %12.0g V201507x PRE: SUMMARY: Respondent age
V201556 double %12.0g V201556 RESTRICTED: PRE: LATINO Rs:
country of Latino heritage
revised
V201556z double %12.0g V201556z RESTRICTED: PRE: LATINO Rs:
country of Latino heritage
revised - Other {SPECIFY}
V201562 double %12.0g V201562 PRE: LATINO Rs: language at home
V201567 double %12.0g V201567 PRE: How many children in HH age
0-17
V201633k double %12.0g V201633k PRE: Mention: Radio PROG - The
Savage Nation (Michael Savage)
V201658p double %12.0g V201658p PRE: IWR OBS: R reactions to
interview - R had difficulty
with selected language
V202001 double %12.0g V202001 POST: R select language
V202182 double %12.0g V202182 POST: Feeling thermometer:
Immigration and Customs
Enforcement (ICE) agency
V202317 double %12.0g V202317 POST: How much opportunity in
America for average person to
get ahead
V202377 double %12.0g V202377 POST: Should the minimum wage be
raised, kept the same, or
lowered
V202555 double %12.0g V202555 POST: Have world temperatuers
have risen on average or last
100 years or not
V202638 double %12.0g V202638 POST: IWR OBS: respondent's
estimated age
V203401 double %12.0g V203401 PRE: IWR DESCR: age
V203411 double %12.0g V203411 POST: IWR DESCR: ageLet’s codebook the measure of respondent age:
codebook V201507x
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
-------------------------------------------------------------------------------
V201507x PRE: SUMMARY: Respondent age
-------------------------------------------------------------------------------
type: numeric (double)
label: V201507x, but 62 nonmissing values are not labeled
range: [-9,80] units: 1
unique values: 64 missing .: 0/8,280
examples: 32
44
58
68 Let’s tab the measure of respondent age:
tab V201507x, mi
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
PRE: SUMMARY: |
Respondent age | Freq. Percent Cum.
--------------------+-----------------------------------
-9. Refused | 348 4.20 4.20
18 | 35 0.42 4.63
19 | 52 0.63 5.25
20 | 46 0.56 5.81
21 | 51 0.62 6.43
22 | 57 0.69 7.11
23 | 75 0.91 8.02
24 | 92 1.11 9.13
25 | 104 1.26 10.39
26 | 108 1.30 11.69
27 | 132 1.59 13.29
28 | 120 1.45 14.73
29 | 131 1.58 16.32
30 | 142 1.71 18.03
31 | 109 1.32 19.35
32 | 117 1.41 20.76
33 | 123 1.49 22.25
34 | 142 1.71 23.96
35 | 152 1.84 25.80
36 | 144 1.74 27.54
37 | 149 1.80 29.34
38 | 152 1.84 31.17
39 | 151 1.82 33.00
40 | 139 1.68 34.67
41 | 151 1.82 36.50
42 | 113 1.36 37.86
43 | 116 1.40 39.26
44 | 111 1.34 40.60
45 | 116 1.40 42.00
46 | 119 1.44 43.44
47 | 106 1.28 44.72
48 | 105 1.27 45.99
49 | 123 1.49 47.48
50 | 154 1.86 49.34
51 | 128 1.55 50.88
52 | 111 1.34 52.22
53 | 117 1.41 53.64
54 | 123 1.49 55.12
55 | 140 1.69 56.81
56 | 127 1.53 58.35
57 | 136 1.64 59.99
58 | 145 1.75 61.74
59 | 154 1.86 63.60
60 | 168 2.03 65.63
61 | 139 1.68 67.31
62 | 154 1.86 69.17
63 | 156 1.88 71.05
64 | 155 1.87 72.92
65 | 180 2.17 75.10
66 | 170 2.05 77.15
67 | 142 1.71 78.86
68 | 140 1.69 80.56
69 | 158 1.91 82.46
70 | 126 1.52 83.99
71 | 147 1.78 85.76
72 | 145 1.75 87.51
73 | 147 1.78 89.29
74 | 94 1.14 90.42
75 | 93 1.12 91.55
76 | 89 1.07 92.62
77 | 81 0.98 93.60
78 | 64 0.77 94.37
79 | 63 0.76 95.13
80. Age 80 or older | 403 4.87 100.00
--------------------+-----------------------------------
Total | 8,280 100.00Let’s use the replace command to code refusals as missing:
gen age = V201507x
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...replace age = . if age < 18
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
(348 real changes made, 348 to missing)sum age
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Variable | Obs Mean Std. Dev. Min Max
-------------+---------------------------------------------------------
age | 7,932 51.58522 17.20718 18 80We could also have combined the generate and logical restriction, such as:
gen ageALT = V201507x if V201507x >=18 & V201507x <=90
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
(348 missing values generated)sum age ageALT
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Variable | Obs Mean Std. Dev. Min Max
-------------+---------------------------------------------------------
age | 7,932 51.58522 17.20718 18 80
ageALT | 7,932 51.58522 17.20718 18 80Let’s change the variable name “age” to “AGE”. Stata is case sensitive, so Stata treats “age” and “AGE” as different variables. (It’s not necessarily a good idea to write your variable names in ALL CAPS, but I’ll do that so that it’s hopefully easier to read through the code and output).
rename age AGE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...Let’s find and recode the categorical measure of respondent race:
lookfor race
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
storage display value
variable name type format label variable label
-------------------------------------------------------------------------------
V201014c double %12.0g V201014c PRE: Senate race in state of
registration (all
registrations)
V201014d double %12.0g V201014d PRE: Governor race in state of
registration (all
registrations)
V201047x double %12.0g V201047x PRE SUMMARY: Senate and Governor
races
V201218 double %12.0g V201218 PRE: Will Presidential race be
close or will (winner) win by a
lot
V201220 double %12.0g V201220 PRE: Will Presidential race be
close in state
V201547a double %12.0g V201547a RESTRICTED: PRE: Race of R: White
[mention]
V201547b double %12.0g V201547b RESTRICTED: PRE: Race of R: Black
or African-American [mention]
V201547c double %12.0g V201547c RESTRICTED: PRE: Race of R: Asian
[mention]
V201547d double %12.0g V201547d RESTRICTED: PRE: Race of R:
Native Hawaiian or Pacific
Islander [mention]
V201547e double %12.0g V201547e RESTRICTED: PRE: Race of R:
Native American or Alaska
Native [mention]
V201547z double %12.0g V201547z RESTRICTED: PRE: Race of R: other
specify
V201549x double %12.0g V201549x PRE: SUMMARY: R self-identified
race/ethnicity
V201564a double %12.0g V201564a RESTRICTED: PRE: R spouse/partner
race: White [mention]
V201564b double %12.0g V201564b RESTRICTED: PRE: R spouse/partner
race: Black or African-American
[mention]
V201564c double %12.0g V201564c RESTRICTED: PRE: R spouse/partner
race: Asian [mention]
V201564d double %12.0g V201564d * RESTRICTED: PRE: R spouse/partner
race: Native Hawiian or Pacific
Islander [ment
V201564e double %12.0g V201564e * RESTRICTED: PRE: R spouse/partner
race: Native American or Alaska
Native [mentio
V201565x double %12.0g V201565x PRE: SUMMARY: R spouse or partner
race/ethnicity
V202055b double %12.0g V202055b POST: Senate race in state of
registration (pre nonvoter)
V202055c double %12.0g V202055c POST: Governor race in state of
registration (pre nonvoter)
V202087 double %12.0g V202087 POST: Did R vote for US Senate -
1 Senate race
V202153 double %12.0g V202153 POST: Feeling thermometer: SR.
SENATOR IN STATE WITHOUT RACE
V202154 double %12.0g V202154 POST: Feeling thermometer: JR.
SENATOR IN STATE WITHOUT RACE
V202155 double %12.0g V202155 POST: Feeling thermometer:
NONRUNNING SENATOR IN STATE
W/RACE
V202455 double %12.0g V202455 POST: How often R feels
protective of someone due to
race or ethnicity
V202537 double %12.0g V202537 POST: How much discrimination has
R faced personally because or
race/ethnicity
V203402 double %12.0g V203402 PRE: IWR DESCR: race/ethnicity
V203412 double %12.0g V203412 POST: IWR DESCR: race/ethnicity
V203500 double %12.0g V203500 CAND: Type of Senate race
V203501 str22 %22s CAND: Name of Senior Senator
(state without Senate race)
V203502 double %12.0g V203502 CAND: Gender of Senior Senator
(state without Senate race)
V203503 double %12.0g V203503 CAND: Party of Senior Senator
(state without Senate race)
V203504 str18 %18s CAND: Name of Junior Senator
(state without Senate race)
V203505 double %12.0g V203505 CAND: Gender of Junior Senator
(state without Senate race)
V203506 double %12.0g V203506 CAND: Party of Junior Senator
(state without Senate race)
V203508 str34 %34s CAND: Name of Democratic Senate
candidate (state with Senate
race)
V203509 double %12.0g V203509 CAND: Gender of Democratic Senate
candidate (state with Senate
race)
V203510 str28 %28s CAND: Name of Republican Senate
candidate (state with Senate
race)
V203511 double %12.0g V203511 CAND: Gender of Republican Senate
candidate (state with Senate
race)
V203512 str26 %26s CAND: Name of other Senate
candidate (state with Senate
race)
V203513 double %12.0g V203513 CAND: Gender of other Senate
candidate (state with Senate
race)
V203514 double %12.0g V203514 CAND: Party of other Senate
candidate (state with Senate
race)
V203515 double %12.0g V203515 CAND: Type of House race
V203523 double %12.0g V203523 CAND: Type of Gubernatorial racetab V201549x, mi
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
PRE: SUMMARY: R self-identified |
race/ethnicity | Freq. Percent Cum.
----------------------------------------+-----------------------------------
-9. Refused | 96 1.16 1.16
-8. Don't know | 6 0.07 1.23
1. White, non-Hispanic | 5,963 72.02 73.25
2. Black, non-Hispanic | 726 8.77 82.02
3. Hispanic | 762 9.20 91.22
4. Asian or Native Hawaiian/other Pacif | 284 3.43 94.65
5. Native American/Alaska Native or oth | 172 2.08 96.73
6. Multiple races, non-Hispanic | 271 3.27 100.00
----------------------------------------+-----------------------------------
Total | 8,280 100.00Below, the gen command “gen race = V201549x” generates a duplicate variable but without the value labels. The recode command tells Stata to take the “race” variable values of -9 through -8 and change them to be 9. For some operations, Stata requires category numbers to be whole numbers, so that’s a reason to change -9 and -8 to be 9. The label define command tells Stata what labels to apply to what variable values, and the label values command tells Stata to label the variable with those values.
gen RACE = V201549x
recode RACE (-9/-8 = 9)
label define RACELABEL 1 "White" 2 "Black" 3 "Hispanic" 4 "Asian" 5 "Native" 6 "Multiracial" 9 "DK/Refused"
label values RACE RACELABEL
tab RACE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
(RACE: 102 changes made)
RACE | Freq. Percent Cum.
------------+-----------------------------------
White | 5,963 72.02 72.02
Black | 726 8.77 80.79
Hispanic | 762 9.20 89.99
Asian | 284 3.43 93.42
Native | 172 2.08 95.50
Multiracial | 271 3.27 98.77
DK/Refused | 102 1.23 100.00
------------+-----------------------------------
Total | 8,280 100.00Another way to have labeled the race variable is to use the clonevar command, which clones the variable and keeps the value labels. We can then copy the existing labels for the race variable and then modify those existing labels:
clonevar RACE2 = V201549x
recode RACE2 (-9/-8 = 9)
label copy V201549x RACELABEL2
label define RACELABEL2 9 "9. DK/Refused", add
label values RACE2 RACELABEL2
tab RACE2
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
(RACE2: 102 changes made)
PRE: SUMMARY: R self-identified |
race/ethnicity | Freq. Percent Cum.
----------------------------------------+-----------------------------------
1. White, non-Hispanic | 5,963 72.02 72.02
2. Black, non-Hispanic | 726 8.77 80.79
3. Hispanic | 762 9.20 89.99
4. Asian or Native Hawaiian/other Pacif | 284 3.43 93.42
5. Native American/Alaska Native or oth | 172 2.08 95.50
6. Multiple races, non-Hispanic | 271 3.27 98.77
9. DK/Refused | 102 1.23 100.00
----------------------------------------+-----------------------------------
Total | 8,280 100.005.8 Sample Stata commands to analyze data
Let’s use age to predict FTPOLICE in a linear regression:
reg FTPOLICE AGE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Source | SS df MS Number of obs = 7,098
-------------+---------------------------------- F(1, 7096) = 733.15
Model | 422615.488 1 422615.488 Prob > F = 0.0000
Residual | 4090421.12 7,096 576.440405 R-squared = 0.0936
-------------+---------------------------------- Adj R-squared = 0.0935
Total | 4513036.6 7,097 635.907652 Root MSE = 24.009
------------------------------------------------------------------------------
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
AGE | 0.45 0.02 27.08 0.000 0.42 0.48
_cons | 47.16 0.91 51.97 0.000 45.38 48.94
------------------------------------------------------------------------------Let’s get predicted outcomes for FTPOLICE at selected ages. Note that the prediction labeled “_at 1” is not the prediction for AGE 1 but is instead the first prediction, for AGE 18. Similarly, the prediction labeled “_at 2” is the second prediction, for AGE 30.
reg FTPOLICE AGE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Source | SS df MS Number of obs = 7,098
-------------+---------------------------------- F(1, 7096) = 733.15
Model | 422615.488 1 422615.488 Prob > F = 0.0000
Residual | 4090421.12 7,096 576.440405 R-squared = 0.0936
-------------+---------------------------------- Adj R-squared = 0.0935
Total | 4513036.6 7,097 635.907652 Root MSE = 24.009
------------------------------------------------------------------------------
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
AGE | 0.45 0.02 27.08 0.000 0.42 0.48
_cons | 47.16 0.91 51.97 0.000 45.38 48.94
------------------------------------------------------------------------------margins, at(AGE=(18 30 45 60 80))
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Adjusted predictions Number of obs = 7,098
Model VCE : OLS
Expression : Linear prediction, predict()
1._at : AGE = 18
2._at : AGE = 30
3._at : AGE = 45
4._at : AGE = 60
5._at : AGE = 80
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 55.26 0.63 87.69 0.000 54.03 56.50
2 | 60.67 0.46 131.56 0.000 59.76 61.57
3 | 67.42 0.31 219.90 0.000 66.82 68.02
4 | 74.18 0.32 234.79 0.000 73.56 74.79
5 | 83.18 0.55 151.58 0.000 82.11 84.26
------------------------------------------------------------------------------Let’s get predicted outcomes for FTPOLICE at each level of AGE.
reg FTPOLICE AGE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Source | SS df MS Number of obs = 7,098
-------------+---------------------------------- F(1, 7096) = 733.15
Model | 422615.488 1 422615.488 Prob > F = 0.0000
Residual | 4090421.12 7,096 576.440405 R-squared = 0.0936
-------------+---------------------------------- Adj R-squared = 0.0935
Total | 4513036.6 7,097 635.907652 Root MSE = 24.009
------------------------------------------------------------------------------
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
AGE | 0.45 0.02 27.08 0.000 0.42 0.48
_cons | 47.16 0.91 51.97 0.000 45.38 48.94
------------------------------------------------------------------------------margins, at(AGE=(18(1)90))
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Adjusted predictions Number of obs = 7,098
Model VCE : OLS
Expression : Linear prediction, predict()
1._at : AGE = 18
2._at : AGE = 19
3._at : AGE = 20
4._at : AGE = 21
5._at : AGE = 22
6._at : AGE = 23
7._at : AGE = 24
8._at : AGE = 25
9._at : AGE = 26
10._at : AGE = 27
11._at : AGE = 28
12._at : AGE = 29
13._at : AGE = 30
14._at : AGE = 31
15._at : AGE = 32
16._at : AGE = 33
17._at : AGE = 34
18._at : AGE = 35
19._at : AGE = 36
20._at : AGE = 37
21._at : AGE = 38
22._at : AGE = 39
23._at : AGE = 40
24._at : AGE = 41
25._at : AGE = 42
26._at : AGE = 43
27._at : AGE = 44
28._at : AGE = 45
29._at : AGE = 46
30._at : AGE = 47
31._at : AGE = 48
32._at : AGE = 49
33._at : AGE = 50
34._at : AGE = 51
35._at : AGE = 52
36._at : AGE = 53
37._at : AGE = 54
38._at : AGE = 55
39._at : AGE = 56
40._at : AGE = 57
41._at : AGE = 58
42._at : AGE = 59
43._at : AGE = 60
44._at : AGE = 61
45._at : AGE = 62
46._at : AGE = 63
47._at : AGE = 64
48._at : AGE = 65
49._at : AGE = 66
50._at : AGE = 67
51._at : AGE = 68
52._at : AGE = 69
53._at : AGE = 70
54._at : AGE = 71
55._at : AGE = 72
56._at : AGE = 73
57._at : AGE = 74
58._at : AGE = 75
59._at : AGE = 76
60._at : AGE = 77
61._at : AGE = 78
62._at : AGE = 79
63._at : AGE = 80
64._at : AGE = 81
65._at : AGE = 82
66._at : AGE = 83
67._at : AGE = 84
68._at : AGE = 85
69._at : AGE = 86
70._at : AGE = 87
71._at : AGE = 88
72._at : AGE = 89
73._at : AGE = 90
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 55.26 0.63 87.69 0.000 54.03 56.50
2 | 55.71 0.62 90.53 0.000 54.51 56.92
3 | 56.16 0.60 93.49 0.000 54.99 57.34
4 | 56.61 0.59 96.59 0.000 55.47 57.76
5 | 57.06 0.57 99.82 0.000 55.94 58.19
6 | 57.51 0.56 103.20 0.000 56.42 58.61
7 | 57.97 0.54 106.73 0.000 56.90 59.03
8 | 58.42 0.53 110.42 0.000 57.38 59.45
9 | 58.87 0.52 114.28 0.000 57.86 59.88
10 | 59.32 0.50 118.32 0.000 58.33 60.30
11 | 59.77 0.49 122.54 0.000 58.81 60.72
12 | 60.22 0.47 126.95 0.000 59.29 61.15
13 | 60.67 0.46 131.56 0.000 59.76 61.57
14 | 61.12 0.45 136.36 0.000 60.24 62.00
15 | 61.57 0.44 141.38 0.000 60.71 62.42
16 | 62.02 0.42 146.60 0.000 61.19 62.85
17 | 62.47 0.41 152.02 0.000 61.66 63.27
18 | 62.92 0.40 157.65 0.000 62.14 63.70
19 | 63.37 0.39 163.48 0.000 62.61 64.13
20 | 63.82 0.38 169.48 0.000 63.08 64.56
21 | 64.27 0.37 175.64 0.000 63.55 64.99
22 | 64.72 0.36 181.94 0.000 64.02 65.42
23 | 65.17 0.35 188.34 0.000 64.49 65.85
24 | 65.62 0.34 194.80 0.000 64.96 66.28
25 | 66.07 0.33 201.25 0.000 65.43 66.71
26 | 66.52 0.32 207.64 0.000 65.89 67.15
27 | 66.97 0.31 213.89 0.000 66.36 67.58
28 | 67.42 0.31 219.90 0.000 66.82 68.02
29 | 67.87 0.30 225.59 0.000 67.28 68.46
30 | 68.32 0.30 230.86 0.000 67.74 68.90
31 | 68.77 0.29 235.60 0.000 68.20 69.34
32 | 69.22 0.29 239.72 0.000 68.66 69.79
33 | 69.67 0.29 243.15 0.000 69.11 70.23
34 | 70.12 0.29 245.80 0.000 69.56 70.68
35 | 70.57 0.28 247.63 0.000 70.01 71.13
36 | 71.02 0.29 248.62 0.000 70.46 71.58
37 | 71.47 0.29 248.77 0.000 70.91 72.04
38 | 71.92 0.29 248.10 0.000 71.36 72.49
39 | 72.37 0.29 246.67 0.000 71.80 72.95
40 | 72.82 0.30 244.54 0.000 72.24 73.41
41 | 73.28 0.30 241.79 0.000 72.68 73.87
42 | 73.73 0.31 238.51 0.000 73.12 74.33
43 | 74.18 0.32 234.79 0.000 73.56 74.79
44 | 74.63 0.32 230.72 0.000 73.99 75.26
45 | 75.08 0.33 226.38 0.000 74.43 75.73
46 | 75.53 0.34 221.85 0.000 74.86 76.19
47 | 75.98 0.35 217.19 0.000 75.29 76.66
48 | 76.43 0.36 212.47 0.000 75.72 77.13
49 | 76.88 0.37 207.72 0.000 76.15 77.60
50 | 77.33 0.38 203.00 0.000 76.58 78.07
51 | 77.78 0.39 198.34 0.000 77.01 78.55
52 | 78.23 0.40 193.75 0.000 77.44 79.02
53 | 78.68 0.42 189.27 0.000 77.86 79.49
54 | 79.13 0.43 184.90 0.000 78.29 79.97
55 | 79.58 0.44 180.65 0.000 78.72 80.44
56 | 80.03 0.45 176.54 0.000 79.14 80.92
57 | 80.48 0.47 172.57 0.000 79.57 81.39
58 | 80.93 0.48 168.73 0.000 79.99 81.87
59 | 81.38 0.49 165.03 0.000 80.41 82.35
60 | 81.83 0.51 161.47 0.000 80.84 82.82
61 | 82.28 0.52 158.05 0.000 81.26 83.30
62 | 82.73 0.53 154.75 0.000 81.68 83.78
63 | 83.18 0.55 151.58 0.000 82.11 84.26
64 | 83.63 0.56 148.54 0.000 82.53 84.74
65 | 84.08 0.58 145.61 0.000 82.95 85.21
66 | 84.53 0.59 142.80 0.000 83.37 85.69
67 | 84.98 0.61 140.10 0.000 83.79 86.17
68 | 85.43 0.62 137.50 0.000 84.21 86.65
69 | 85.88 0.64 135.01 0.000 84.64 87.13
70 | 86.33 0.65 132.61 0.000 85.06 87.61
71 | 86.78 0.67 130.30 0.000 85.48 88.09
72 | 87.23 0.68 128.08 0.000 85.90 88.57
73 | 87.68 0.70 125.94 0.000 86.32 89.05
------------------------------------------------------------------------------Let’s get a plot of these estimates:
marginsplot, xlabel(0(10)90) ylabel(0(10)100)
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Variables that uniquely identify margins: AGE
Let’s use race to predict FTPOLICE in a linear regression. The i. tells Stata to treat race as a categorical predictor:
reg FTPOLICE i.RACE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Source | SS df MS Number of obs = 7,388
-------------+---------------------------------- F(6, 7381) = 98.30
Model | 345065.388 6 57510.8979 Prob > F = 0.0000
Residual | 4318418.22 7,381 585.072242 R-squared = 0.0740
-------------+---------------------------------- Adj R-squared = 0.0732
Total | 4663483.61 7,387 631.309545 Root MSE = 24.188
------------------------------------------------------------------------------
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
RACE |
Black | -22.48 1.01 -22.24 0.000 -24.46 -20.50
Hispanic | -8.36 1.00 -8.35 0.000 -10.32 -6.40
Asian | -10.26 1.57 -6.52 0.000 -13.35 -7.17
Native | -8.09 1.98 -4.09 0.000 -11.97 -4.22
Multiracial | -12.17 1.61 -7.55 0.000 -15.33 -9.01
DK/Refused | -6.72 2.76 -2.44 0.015 -12.13 -1.32
|
_cons | 74.24 0.33 225.07 0.000 73.59 74.88
------------------------------------------------------------------------------The output makes no sense without the i. prefix, because the race variable doesn’t make sense as a continuous variable in which, for instance, Hispanics have higher levels of race than Blacks do. So don’t do this:
reg FTPOLICE RACE // Don't do this!
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Source | SS df MS Number of obs = 7,388
-------------+---------------------------------- F(1, 7386) = 150.63
Model | 93203.4406 1 93203.4406 Prob > F = 0.0000
Residual | 4570280.17 7,386 618.776085 R-squared = 0.0200
-------------+---------------------------------- Adj R-squared = 0.0199
Total | 4663483.61 7,387 631.309545 Root MSE = 24.875
------------------------------------------------------------------------------
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
RACE | -2.47 0.20 -12.27 0.000 -2.86 -2.07
_cons | 74.75 0.45 167.34 0.000 73.88 75.63
------------------------------------------------------------------------------Let’s re-run the correct regression and then get predicted values of FTPOLICE at different levels of RACE. Note that the seventh prediction (_at 7) is not for category 7 of the RACE variable, but is instead for the seventh prediction, which is for category 9 of the RACE variable.
reg FTPOLICE i.RACE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Source | SS df MS Number of obs = 7,388
-------------+---------------------------------- F(6, 7381) = 98.30
Model | 345065.388 6 57510.8979 Prob > F = 0.0000
Residual | 4318418.22 7,381 585.072242 R-squared = 0.0740
-------------+---------------------------------- Adj R-squared = 0.0732
Total | 4663483.61 7,387 631.309545 Root MSE = 24.188
------------------------------------------------------------------------------
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
RACE |
Black | -22.48 1.01 -22.24 0.000 -24.46 -20.50
Hispanic | -8.36 1.00 -8.35 0.000 -10.32 -6.40
Asian | -10.26 1.57 -6.52 0.000 -13.35 -7.17
Native | -8.09 1.98 -4.09 0.000 -11.97 -4.22
Multiracial | -12.17 1.61 -7.55 0.000 -15.33 -9.01
DK/Refused | -6.72 2.76 -2.44 0.015 -12.13 -1.32
|
_cons | 74.24 0.33 225.07 0.000 73.59 74.88
------------------------------------------------------------------------------margins, at(RACE=(1(1)6 9))
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Adjusted predictions Number of obs = 7,388
Model VCE : OLS
Expression : Linear prediction, predict()
1._at : RACE = 1
2._at : RACE = 2
3._at : RACE = 3
4._at : RACE = 4
5._at : RACE = 5
6._at : RACE = 6
7._at : RACE = 9
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 74.24 0.33 225.07 0.000 73.59 74.88
2 | 51.76 0.96 54.17 0.000 49.88 53.63
3 | 65.88 0.95 69.70 0.000 64.02 67.73
4 | 63.98 1.54 41.57 0.000 60.96 66.99
5 | 66.14 1.95 33.93 0.000 62.32 69.96
6 | 62.07 1.58 39.34 0.000 58.98 65.16
7 | 67.51 2.74 24.65 0.000 62.14 72.88
------------------------------------------------------------------------------Let’s plot the estimates of FTPOLICE by levels of race. For this, let’s run the regression quietly (abbreviated “qui”), so that the output doesn’t appear:
qui reg FTPOLICE i.RACE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...margins, at(RACE=(1(1)6 9))
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Adjusted predictions Number of obs = 7,388
Model VCE : OLS
Expression : Linear prediction, predict()
1._at : RACE = 1
2._at : RACE = 2
3._at : RACE = 3
4._at : RACE = 4
5._at : RACE = 5
6._at : RACE = 6
7._at : RACE = 9
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 74.24 0.33 225.07 0.000 73.59 74.88
2 | 51.76 0.96 54.17 0.000 49.88 53.63
3 | 65.88 0.95 69.70 0.000 64.02 67.73
4 | 63.98 1.54 41.57 0.000 60.96 66.99
5 | 66.14 1.95 33.93 0.000 62.32 69.96
6 | 62.07 1.58 39.34 0.000 58.98 65.16
7 | 67.51 2.74 24.65 0.000 62.14 72.88
------------------------------------------------------------------------------marginsplot, ylabel(0(10)100) xlabel(, angle(vertical)) recast(bar)
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Variables that uniquely identify margins: RACE
5.9 Controlling regression output
The command set cformat can control the number of digits that Stata reports. Let’s first use “set cformat” to get the default number of decimal places:
set cformat
reg FTPOLICE i.RACE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Source | SS df MS Number of obs = 7,388
-------------+---------------------------------- F(6, 7381) = 98.30
Model | 345065.388 6 57510.8979 Prob > F = 0.0000
Residual | 4318418.22 7,381 585.072242 R-squared = 0.0740
-------------+---------------------------------- Adj R-squared = 0.0732
Total | 4663483.61 7,387 631.309545 Root MSE = 24.188
------------------------------------------------------------------------------
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
RACE |
Black | -22.47952 1.010712 -22.24 0.000 -24.4608 -20.49823
Hispanic | -8.359811 1.001014 -8.35 0.000 -10.32209 -6.397537
Asian | -10.26044 1.574009 -6.52 0.000 -13.34595 -7.174932
Native | -8.09329 1.976856 -4.09 0.000 -11.96849 -4.218088
Multiracial | -12.16806 1.611974 -7.55 0.000 -15.32799 -9.008134
DK/Refused | -6.723327 2.758571 -2.44 0.015 -12.13091 -1.315739
|
_cons | 74.23615 .3298332 225.07 0.000 73.58958 74.88271
------------------------------------------------------------------------------Now let’s use “set cformat %9.2f” to report regression coefficients to two decimal places:
set cformat %9.2f
reg FTPOLICE i.RACE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Source | SS df MS Number of obs = 7,388
-------------+---------------------------------- F(6, 7381) = 98.30
Model | 345065.388 6 57510.8979 Prob > F = 0.0000
Residual | 4318418.22 7,381 585.072242 R-squared = 0.0740
-------------+---------------------------------- Adj R-squared = 0.0732
Total | 4663483.61 7,387 631.309545 Root MSE = 24.188
------------------------------------------------------------------------------
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
RACE |
Black | -22.48 1.01 -22.24 0.000 -24.46 -20.50
Hispanic | -8.36 1.00 -8.35 0.000 -10.32 -6.40
Asian | -10.26 1.57 -6.52 0.000 -13.35 -7.17
Native | -8.09 1.98 -4.09 0.000 -11.97 -4.22
Multiracial | -12.17 1.61 -7.55 0.000 -15.33 -9.01
DK/Refused | -6.72 2.76 -2.44 0.015 -12.13 -1.32
|
_cons | 74.24 0.33 225.07 0.000 73.59 74.88
------------------------------------------------------------------------------5.10 Survey weights in Stata
Survey weights can help account for the possibility that respondents in the sample are not representative of the population. Let’s set up the survey weights for the ANES 2020 Time Series Study post-election wave. See the corresponding ANES Codebook.
desc V200010b V200010d V200010c
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
storage display value
variable name type format label variable label
-------------------------------------------------------------------------------
V200010b double %12.0g Full sample post-election weight
V200010d double %12.0g Full sample variance stratum
V200010c double %12.0g Full sample variance unitsvyset [pweight=V200010b], strata(V200010d) psu(V200010c)
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
pweight: V200010b
VCE: linearized
Single unit: missing
Strata 1: V200010d
SU 1: V200010c
FPC 1: <zero>Let’s illustrate how survey weights can adjust estimates. Let’s start with the sample percentages, indicating that, for example, non-Hispanic White respondents are 72 percent of the sample:
prop RACE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Proportion estimation Number of obs = 8,280
_prop_7: RACE = DK/Refused
--------------------------------------------------------------
| Logit
| Proportion Std. Err. [95% Conf. Interval]
-------------+------------------------------------------------
RACE |
White | 0.72 0.00 0.71 0.73
Black | 0.09 0.00 0.08 0.09
Hispanic | 0.09 0.00 0.09 0.10
Asian | 0.03 0.00 0.03 0.04
Native | 0.02 0.00 0.02 0.02
Multiracial | 0.03 0.00 0.03 0.04
_prop_7 | 0.01 0.00 0.01 0.01
--------------------------------------------------------------Let’s compare to the population percentage estimates, in which, for example, the analysis adjusted non-Hispanic White respondents to be 65 percent of the population:
svy: prop RACE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
(running proportion on estimation sample)
Survey: Proportion estimation
Number of strata = 50 Number of obs = 7,453
Number of PSUs = 101 Population size = 7,453
Design df = 51
_prop_7: RACE = DK/Refused
--------------------------------------------------------------
| Linearized Logit
| Proportion Std. Err. [95% Conf. Interval]
-------------+------------------------------------------------
RACE |
White | 0.65 0.01 0.63 0.67
Black | 0.11 0.01 0.10 0.12
Hispanic | 0.13 0.01 0.12 0.15
Asian | 0.04 0.00 0.03 0.05
Native | 0.02 0.00 0.02 0.02
Multiracial | 0.04 0.00 0.03 0.05
_prop_7 | 0.01 0.00 0.01 0.01
--------------------------------------------------------------Let’s check the mean weights that were applied to the observations. Mean weights under 1 indicate that the group was oversampled relative to their percentage of the population, and mean weights over 1 indicate that the group was undersampled relative to their percentage of the population.
tabstat V200010b, by(RACE) stats(mean)
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Summary for variables: V200010b
by categories of: RACE
RACE | mean
------------+----------
White | .8954746
Black | 1.283733
Hispanic | 1.479558
Asian | 1.171886
Native | .940597
Multiracial | 1.191748
DK/Refused | .8243095
------------+----------
Total | 1
-----------------------Let’s predict FTPOLICE using RACE, without survey weights:
reg FTPOLICE i.RACE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Source | SS df MS Number of obs = 7,388
-------------+---------------------------------- F(6, 7381) = 98.30
Model | 345065.388 6 57510.8979 Prob > F = 0.0000
Residual | 4318418.22 7,381 585.072242 R-squared = 0.0740
-------------+---------------------------------- Adj R-squared = 0.0732
Total | 4663483.61 7,387 631.309545 Root MSE = 24.188
------------------------------------------------------------------------------
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
RACE |
Black | -22.48 1.01 -22.24 0.000 -24.46 -20.50
Hispanic | -8.36 1.00 -8.35 0.000 -10.32 -6.40
Asian | -10.26 1.57 -6.52 0.000 -13.35 -7.17
Native | -8.09 1.98 -4.09 0.000 -11.97 -4.22
Multiracial | -12.17 1.61 -7.55 0.000 -15.33 -9.01
DK/Refused | -6.72 2.76 -2.44 0.015 -12.13 -1.32
|
_cons | 74.24 0.33 225.07 0.000 73.59 74.88
------------------------------------------------------------------------------margins, atmeans at(RACE=(1(1)6 9))
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Adjusted predictions Number of obs = 7,388
Model VCE : OLS
Expression : Linear prediction, predict()
1._at : RACE = 1
2._at : RACE = 2
3._at : RACE = 3
4._at : RACE = 4
5._at : RACE = 5
6._at : RACE = 6
7._at : RACE = 9
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 74.24 0.33 225.07 0.000 73.59 74.88
2 | 51.76 0.96 54.17 0.000 49.88 53.63
3 | 65.88 0.95 69.70 0.000 64.02 67.73
4 | 63.98 1.54 41.57 0.000 60.96 66.99
5 | 66.14 1.95 33.93 0.000 62.32 69.96
6 | 62.07 1.58 39.34 0.000 58.98 65.16
7 | 67.51 2.74 24.65 0.000 62.14 72.88
------------------------------------------------------------------------------Let’s predict FTPOLICE using RACE, with survey weights:
svy: reg FTPOLICE i.RACE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
(running regress on estimation sample)
Survey: Linear regression
Number of strata = 50 Number of obs = 7,388
Number of PSUs = 101 Population size = 7,382.4092
Design df = 51
F( 6, 46) = 46.27
Prob > F = 0.0000
R-squared = 0.0871
------------------------------------------------------------------------------
| Linearized
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
RACE |
Black | -23.32 1.60 -14.57 0.000 -26.54 -20.11
Hispanic | -10.02 1.34 -7.46 0.000 -12.71 -7.32
Asian | -11.91 1.84 -6.49 0.000 -15.60 -8.23
Native | -5.11 3.07 -1.66 0.102 -11.28 1.06
Multiracial | -13.39 3.33 -4.02 0.000 -20.08 -6.69
DK/Refused | -9.50 3.65 -2.60 0.012 -16.83 -2.16
|
_cons | 73.99 0.52 143.52 0.000 72.96 75.03
------------------------------------------------------------------------------margins, atmeans at(RACE=(1(1)6 9))
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Adjusted predictions Number of obs = 7,388
Model VCE : Linearized
Expression : Linear prediction, predict()
1._at : RACE = 1
2._at : RACE = 2
3._at : RACE = 3
4._at : RACE = 4
5._at : RACE = 5
6._at : RACE = 6
7._at : RACE = 9
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 73.99 0.52 143.52 0.000 72.96 75.03
2 | 50.67 1.51 33.64 0.000 47.65 53.69
3 | 63.97 1.26 50.80 0.000 61.45 66.50
4 | 62.08 1.75 35.40 0.000 58.56 65.60
5 | 68.88 3.02 22.81 0.000 62.82 74.94
6 | 60.60 3.16 19.19 0.000 54.27 66.94
7 | 64.49 3.67 17.55 0.000 57.12 71.87
------------------------------------------------------------------------------5.11 Categorical predictors from continuous variables
Let’s plot FTPOLICE by respondent AGE:
svy: reg FTPOLICE AGE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
(running regress on estimation sample)
Survey: Linear regression
Number of strata = 50 Number of obs = 7,098
Number of PSUs = 101 Population size = 7,113.0638
Design df = 51
F( 1, 51) = 262.02
Prob > F = 0.0000
R-squared = 0.0889
------------------------------------------------------------------------------
| Linearized
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
AGE | 0.46 0.03 16.19 0.000 0.40 0.51
_cons | 47.24 1.59 29.71 0.000 44.05 50.43
------------------------------------------------------------------------------margins, atmeans at(AGE=(18(1)80)) noatlegend
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Adjusted predictions Number of obs = 7,098
Model VCE : Linearized
Expression : Linear prediction, predict()
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 55.49 1.10 50.42 0.000 53.28 57.70
2 | 55.95 1.07 52.10 0.000 53.79 58.10
3 | 56.40 1.05 53.85 0.000 54.30 58.51
4 | 56.86 1.02 55.70 0.000 54.81 58.91
5 | 57.32 0.99 57.64 0.000 55.32 59.32
6 | 57.78 0.97 59.68 0.000 55.83 59.72
7 | 58.24 0.94 61.82 0.000 56.35 60.13
8 | 58.69 0.92 64.08 0.000 56.86 60.53
9 | 59.15 0.89 66.45 0.000 57.37 60.94
10 | 59.61 0.86 68.96 0.000 57.88 61.35
11 | 60.07 0.84 71.61 0.000 58.38 61.75
12 | 60.53 0.81 74.41 0.000 58.89 62.16
13 | 60.98 0.79 77.37 0.000 59.40 62.57
14 | 61.44 0.76 80.50 0.000 59.91 62.98
15 | 61.90 0.74 83.82 0.000 60.42 63.38
16 | 62.36 0.71 87.33 0.000 60.93 63.79
17 | 62.82 0.69 91.06 0.000 61.43 64.20
18 | 63.28 0.67 95.01 0.000 61.94 64.61
19 | 63.73 0.64 99.19 0.000 62.44 65.02
20 | 64.19 0.62 103.63 0.000 62.95 65.44
21 | 64.65 0.60 108.33 0.000 63.45 65.85
22 | 65.11 0.57 113.30 0.000 63.95 66.26
23 | 65.57 0.55 118.55 0.000 64.46 66.68
24 | 66.02 0.53 124.08 0.000 64.96 67.09
25 | 66.48 0.51 129.88 0.000 65.45 67.51
26 | 66.94 0.49 135.94 0.000 65.95 67.93
27 | 67.40 0.47 142.23 0.000 66.45 68.35
28 | 67.86 0.46 148.70 0.000 66.94 68.77
29 | 68.31 0.44 155.29 0.000 67.43 69.20
30 | 68.77 0.42 161.92 0.000 67.92 69.63
31 | 69.23 0.41 168.46 0.000 68.41 70.06
32 | 69.69 0.40 174.79 0.000 68.89 70.49
33 | 70.15 0.39 180.73 0.000 69.37 70.93
34 | 70.61 0.38 186.10 0.000 69.84 71.37
35 | 71.06 0.37 190.72 0.000 70.32 71.81
36 | 71.52 0.37 194.42 0.000 70.78 72.26
37 | 71.98 0.37 197.06 0.000 71.25 72.71
38 | 72.44 0.36 198.54 0.000 71.71 73.17
39 | 72.90 0.37 198.83 0.000 72.16 73.63
40 | 73.35 0.37 197.96 0.000 72.61 74.10
41 | 73.81 0.38 196.00 0.000 73.06 74.57
42 | 74.27 0.38 193.11 0.000 73.50 75.04
43 | 74.73 0.39 189.43 0.000 73.94 75.52
44 | 75.19 0.41 185.13 0.000 74.37 76.00
45 | 75.64 0.42 180.39 0.000 74.80 76.49
46 | 76.10 0.43 175.35 0.000 75.23 76.97
47 | 76.56 0.45 170.15 0.000 75.66 77.46
48 | 77.02 0.47 164.89 0.000 76.08 77.96
49 | 77.48 0.49 159.65 0.000 76.50 78.45
50 | 77.94 0.50 154.51 0.000 76.92 78.95
51 | 78.39 0.52 149.51 0.000 77.34 79.45
52 | 78.85 0.55 144.67 0.000 77.76 79.95
53 | 79.31 0.57 140.03 0.000 78.17 80.45
54 | 79.77 0.59 135.59 0.000 78.59 80.95
55 | 80.23 0.61 131.35 0.000 79.00 81.45
56 | 80.68 0.63 127.32 0.000 79.41 81.96
57 | 81.14 0.66 123.50 0.000 79.82 82.46
58 | 81.60 0.68 119.86 0.000 80.23 82.97
59 | 82.06 0.70 116.42 0.000 80.64 83.47
60 | 82.52 0.73 113.16 0.000 81.05 83.98
61 | 82.97 0.75 110.07 0.000 81.46 84.49
62 | 83.43 0.78 107.14 0.000 81.87 85.00
63 | 83.89 0.80 104.36 0.000 82.28 85.50
------------------------------------------------------------------------------marginsplot, xlabel(10(10)80) ylabel(0(10)100)
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Variables that uniquely identify margins: AGE
Let’s re-run the analysis but with a categorical measure of respondent age:
desc V201507x
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
storage display value
variable name type format label variable label
-------------------------------------------------------------------------------
V201507x double %12.0g V201507x PRE: SUMMARY: Respondent agerecode V201507x (18/21=1 "18/21 [1]") (22/29=2 "22/29 [2]") (30/39=3 "30/39 [3]") (40/49=4 "40/49 [4]") (50/59=5 "50/59 [5]") (60/69=6 "60/69 [6]") (70/79=7 "70/79 [7]") (80=8 "80+ [8]") (-9=9 "Refused [9]"), gen(AGECAT)
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
(8280 differences between V201507x and AGECAT)tab AGECAT, mi
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
RECODE of |
V201507x |
(PRE: |
SUMMARY: |
Respondent |
age) | Freq. Percent Cum.
------------+-----------------------------------
18/21 [1] | 184 2.22 2.22
22/29 [2] | 819 9.89 12.11
30/39 [3] | 1,381 16.68 28.79
40/49 [4] | 1,199 14.48 43.27
50/59 [5] | 1,335 16.12 59.40
60/69 [6] | 1,562 18.86 78.26
70/79 [7] | 1,049 12.67 90.93
80+ [8] | 403 4.87 95.80
Refused [9] | 348 4.20 100.00
------------+-----------------------------------
Total | 8,280 100.00svy: reg FTPOLICE i.AGECAT
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
(running regress on estimation sample)
Survey: Linear regression
Number of strata = 50 Number of obs = 7,388
Number of PSUs = 101 Population size = 7,382.4092
Design df = 51
F( 8, 44) = 36.53
Prob > F = 0.0000
R-squared = 0.0971
------------------------------------------------------------------------------
| Linearized
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
AGECAT |
22/29 [2] | 0.63 3.41 0.18 0.855 -6.22 7.47
30/39 [3] | 9.88 3.23 3.06 0.004 3.40 16.37
40/49 [4] | 15.09 3.37 4.48 0.000 8.33 21.86
50/59 [5] | 21.88 3.10 7.06 0.000 15.66 28.10
60/69 [6] | 21.21 3.38 6.27 0.000 14.42 28.01
70/79 [7] | 23.60 3.05 7.73 0.000 17.47 29.73
80+ [8] | 25.67 3.42 7.51 0.000 18.80 32.53
Refused [9] | 15.93 3.24 4.92 0.000 9.43 22.43
|
_cons | 54.32 3.11 17.47 0.000 48.08 60.56
------------------------------------------------------------------------------margins, atmeans at(AGECAT=(1(1)8)) noatlegend
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Adjusted predictions Number of obs = 7,388
Model VCE : Linearized
Expression : Linear prediction, predict()
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 54.32 3.11 17.47 0.000 48.08 60.56
2 | 54.95 1.57 35.04 0.000 51.80 58.09
3 | 64.20 1.05 61.08 0.000 62.09 66.31
4 | 69.41 0.96 71.98 0.000 67.48 71.35
5 | 76.20 0.68 111.43 0.000 74.83 77.57
6 | 75.53 0.92 82.53 0.000 73.70 77.37
7 | 77.92 1.17 66.71 0.000 75.57 80.26
8 | 79.99 1.50 53.47 0.000 76.98 82.99
------------------------------------------------------------------------------marginsplot, ylabel(0(10)100)
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Variables that uniquely identify margins: AGECAT
Sometimes categorical predictors for continuous variables are preferable. One benefit of a categorical predictor is that the categorical predictor does not assume that the association between X and Y is linear. Another benefit of a categorical predictor is that the categorical predictor can retain missing data in a separate category, so that we do not lose those data.
The number of categories to make from a continuous predictor depends on the sample size that could fit into each category. Another factor to consider is important reference values for the measure. For example, for a feeling thermometer in which ratings range from 0 for very cold to 100 for very warm, it makes a lot of sense to have a separate category for the neutral response of 50. And it makes sense to have separate options for the coldest rating and for the warmest rating. But, for a particular target, if only 5 respondents select the coldest option, then it might not make sense to have the coldest option as a separate category. Instead, we might need to code a “cold” category of, say, ratings from 0 to 25.
Let’s check what the age predictor looks like if each age is coded as a separate category:
Let’s plot FTPOLICE by AGE:
svy: reg FTPOLICE i.AGE
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
(running regress on estimation sample)
Survey: Linear regression
Number of strata = 50 Number of obs = 7,098
Number of PSUs = 101 Population size = 7,113.0638
Design df = 51
F( 51, 1) = .
Prob > F = .
R-squared = 0.1111
------------------------------------------------------------------------------
| Linearized
FTPOLICE | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
AGE |
19 | -6.62 8.83 -0.75 0.457 -24.34 11.10
20 | -3.46 8.87 -0.39 0.698 -21.26 14.35
21 | -9.13 8.81 -1.04 0.305 -26.81 8.55
22 | -4.00 7.35 -0.54 0.589 -18.75 10.75
23 | -9.30 7.51 -1.24 0.221 -24.37 5.77
24 | -2.40 7.72 -0.31 0.757 -17.90 13.10
25 | -7.54 7.63 -0.99 0.327 -22.86 7.77
26 | -2.75 8.57 -0.32 0.749 -19.95 14.45
27 | -5.85 7.26 -0.81 0.424 -20.43 8.73
28 | 1.58 6.72 0.24 0.815 -11.91 15.08
29 | -4.69 6.58 -0.71 0.479 -17.91 8.53
30 | 5.35 7.19 0.74 0.460 -9.08 19.77
31 | 5.78 6.94 0.83 0.409 -8.16 19.72
32 | 3.82 6.75 0.57 0.574 -9.73 17.37
33 | 5.11 6.83 0.75 0.458 -8.59 18.81
34 | 2.54 7.12 0.36 0.723 -11.76 16.84
35 | 1.29 6.29 0.21 0.838 -11.33 13.92
36 | 7.79 7.11 1.10 0.278 -6.48 22.07
37 | 7.26 6.02 1.20 0.234 -4.84 19.35
38 | 8.16 6.51 1.25 0.215 -4.90 21.23
39 | 3.92 6.24 0.63 0.533 -8.61 16.45
40 | 7.71 6.88 1.12 0.267 -6.10 21.52
41 | 10.81 6.22 1.74 0.088 -1.67 23.28
42 | 7.17 7.34 0.98 0.333 -7.56 21.91
43 | 10.89 6.84 1.59 0.118 -2.85 24.62
44 | 8.96 5.71 1.57 0.123 -2.50 20.43
45 | 9.34 6.93 1.35 0.183 -4.56 23.25
46 | 11.71 7.62 1.54 0.131 -3.60 27.02
47 | 11.23 8.12 1.38 0.172 -5.06 27.52
48 | 13.47 7.23 1.86 0.068 -1.04 27.99
49 | 12.18 6.42 1.90 0.063 -0.71 25.08
50 | 13.70 6.33 2.17 0.035 1.00 26.40
51 | 15.72 6.46 2.43 0.019 2.74 28.69
52 | 22.86 6.13 3.73 0.000 10.56 35.16
53 | 13.97 6.83 2.05 0.046 0.26 27.68
54 | 17.08 6.44 2.65 0.011 4.14 30.01
55 | 17.35 6.60 2.63 0.011 4.10 30.59
56 | 17.34 6.56 2.65 0.011 4.18 30.50
57 | 21.36 6.90 3.10 0.003 7.51 35.21
58 | 16.29 6.81 2.39 0.020 2.62 29.97
59 | 14.89 6.63 2.25 0.029 1.58 28.19
60 | 13.86 6.34 2.19 0.033 1.14 26.59
61 | 18.61 6.11 3.05 0.004 6.35 30.87
62 | 19.43 6.28 3.09 0.003 6.81 32.04
63 | 9.23 8.27 1.12 0.270 -7.37 25.83
64 | 22.90 7.05 3.25 0.002 8.74 37.05
65 | 16.69 6.35 2.63 0.011 3.95 29.44
66 | 19.33 6.87 2.81 0.007 5.53 33.13
67 | 16.12 6.50 2.48 0.016 3.07 29.17
68 | 15.21 6.89 2.21 0.032 1.38 29.04
69 | 12.70 6.62 1.92 0.061 -0.59 25.99
70 | 9.71 7.24 1.34 0.186 -4.82 24.24
71 | 19.53 6.62 2.95 0.005 6.24 32.82
72 | 20.06 7.18 2.79 0.007 5.65 34.46
73 | 20.77 5.82 3.57 0.001 9.08 32.46
74 | 22.12 6.32 3.50 0.001 9.44 34.81
75 | 22.12 6.42 3.45 0.001 9.24 35.01
76 | 17.87 8.92 2.00 0.050 -0.03 35.78
77 | 21.34 7.21 2.96 0.005 6.86 35.81
78 | 16.99 7.45 2.28 0.027 2.03 31.94
79 | 19.06 7.04 2.71 0.009 4.94 33.19
80 | 20.85 6.41 3.25 0.002 7.97 33.72
|
_cons | 59.14 6.14 9.64 0.000 46.83 71.46
------------------------------------------------------------------------------margins, atmeans at(AGE=(18(1)80)) noatlegend
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Adjusted predictions Number of obs = 7,098
Model VCE : Linearized
Expression : Linear prediction, predict()
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 59.14 6.14 9.64 0.000 46.83 71.46
2 | 52.52 6.02 8.72 0.000 40.43 64.61
3 | 55.69 6.53 8.52 0.000 42.57 68.81
4 | 50.01 7.33 6.83 0.000 35.30 64.72
5 | 55.14 4.68 11.79 0.000 45.75 64.54
6 | 49.84 5.47 9.11 0.000 38.86 60.82
7 | 56.75 4.27 13.28 0.000 48.16 65.33
8 | 51.60 5.06 10.20 0.000 41.44 61.76
9 | 56.39 4.71 11.98 0.000 46.94 65.84
10 | 53.29 3.85 13.84 0.000 45.56 61.02
11 | 60.72 3.89 15.62 0.000 52.92 68.53
12 | 54.45 2.73 19.94 0.000 48.97 59.93
13 | 64.49 3.14 20.56 0.000 58.19 70.79
14 | 64.92 4.00 16.25 0.000 56.90 72.94
15 | 62.96 3.18 19.82 0.000 56.59 69.34
16 | 64.25 3.23 19.90 0.000 57.77 70.73
17 | 61.68 4.16 14.82 0.000 53.33 70.04
18 | 60.44 3.20 18.88 0.000 54.01 66.86
19 | 66.94 3.08 21.76 0.000 60.76 73.11
20 | 66.40 2.47 26.91 0.000 61.45 71.35
21 | 67.31 3.00 22.47 0.000 61.29 73.32
22 | 63.06 3.11 20.29 0.000 56.83 69.30
23 | 66.86 2.72 24.55 0.000 61.39 72.32
24 | 69.95 2.53 27.64 0.000 64.87 75.03
25 | 66.32 3.00 22.09 0.000 60.29 72.35
26 | 70.03 2.14 32.70 0.000 65.73 74.33
27 | 68.10 3.12 21.80 0.000 61.83 74.38
28 | 68.48 3.33 20.57 0.000 61.80 75.17
29 | 70.85 3.27 21.68 0.000 64.29 77.41
30 | 70.37 5.94 11.84 0.000 58.44 82.31
31 | 72.62 3.64 19.94 0.000 65.31 79.93
32 | 71.33 2.54 28.09 0.000 66.23 76.42
33 | 72.84 1.79 40.60 0.000 69.24 76.44
34 | 74.86 2.60 28.75 0.000 69.63 80.09
35 | 82.01 2.03 40.47 0.000 77.94 86.07
36 | 73.11 2.55 28.65 0.000 67.99 78.23
37 | 76.22 2.47 30.86 0.000 71.26 81.18
38 | 76.49 2.31 33.14 0.000 71.86 81.12
39 | 76.48 2.25 34.04 0.000 71.97 80.99
40 | 80.50 3.06 26.27 0.000 74.35 86.66
41 | 75.44 3.40 22.16 0.000 68.60 82.27
42 | 74.03 2.44 30.35 0.000 69.13 78.93
43 | 73.00 3.60 20.25 0.000 65.77 80.24
44 | 77.75 2.37 32.74 0.000 72.98 82.52
45 | 78.57 1.69 46.60 0.000 75.19 81.95
46 | 68.37 5.21 13.12 0.000 57.91 78.84
47 | 82.04 2.23 36.82 0.000 77.57 86.51
48 | 75.84 1.97 38.51 0.000 71.88 79.79
49 | 78.48 2.21 35.44 0.000 74.03 82.92
50 | 75.27 2.41 31.20 0.000 70.42 80.11
51 | 74.35 2.36 31.53 0.000 69.61 79.08
52 | 71.84 3.30 21.78 0.000 65.22 78.47
53 | 68.85 3.51 19.59 0.000 61.80 75.91
54 | 78.67 2.59 30.41 0.000 73.47 83.86
55 | 79.20 2.62 30.20 0.000 73.93 84.46
56 | 79.92 2.41 33.16 0.000 75.08 84.76
57 | 81.27 2.99 27.22 0.000 75.27 87.26
58 | 81.26 2.18 37.29 0.000 76.89 85.64
59 | 77.02 6.76 11.39 0.000 63.44 90.59
60 | 80.48 2.39 33.65 0.000 75.68 85.28
61 | 76.13 4.01 18.97 0.000 68.07 84.19
62 | 78.21 2.82 27.77 0.000 72.55 83.86
63 | 79.99 1.50 53.47 0.000 76.98 82.99
------------------------------------------------------------------------------marginsplot, xlabel(10(10)80) ylabel(0(10)100)
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
Variables that uniquely identify margins: AGE
Missing data reports
codebook V202110x
recode V202110x (min/-1 3/max = .) (1 = 0) (2 = 1), gen(VOTETB)
tab V202110x VOTETB, mi
codebook V201228
recode V201228 (min/0 5 = .), gen(PID3)
tab V201228 PID3, mi
tab PID3 VOTETB, mi
misstable summarize PID3 VOTETB
misstable patterns PID3 VOTETB
running D:\OneDrive - IL State University\2-teaching\R for teaching\POL497\prof
> ile.do ...
-------------------------------------------------------------------------------
V202110x PRE-POST: SUMMARY: 2020 Presidential vote
-------------------------------------------------------------------------------
type: numeric (double)
label: V202110x
range: [-9,5] units: 1
unique values: 8 missing .: 0/8,280
tabulation: Freq. Numeric Label
81 -9 -9. Refused
2 -8 -8. Don't know
1,920 -1 -1. Inapplicable
3,537 1 1. Joe Biden
2,582 2 2. Donald Trump
74 3 3. Jo Jorgensen
24 4 4. Howie Hawkins
60 5 5. Other candidate {SPECIFY}
(8280 differences between V202110x and VOTETB)
PRE-POST: SUMMARY: | RECODE of V202110x (PRE-POST:
2020 Presidential | SUMMARY: 2020 Presidential vote)
vote | 0 1 . | Total
----------------------+---------------------------------+----------
-9. Refused | 0 0 81 | 81
-8. Don't know | 0 0 2 | 2
-1. Inapplicable | 0 0 1,920 | 1,920
1. Joe Biden | 3,537 0 0 | 3,537
2. Donald Trump | 0 2,582 0 | 2,582
3. Jo Jorgensen | 0 0 74 | 74
4. Howie Hawkins | 0 0 24 | 24
5. Other candidate {S | 0 0 60 | 60
----------------------+---------------------------------+----------
Total | 3,537 2,582 2,161 | 8,280
-------------------------------------------------------------------------------
V201228 PRE: Party ID: Does R think of self as
Democrat, Republican, or Independent
-------------------------------------------------------------------------------
type: numeric (double)
label: V201228
range: [-9,5] units: 1
unique values: 8 missing .: 0/8,280
tabulation: Freq. Numeric Label
44 -9 -9. Refused
4 -8 -8. Don't know
1 -4 -4. Technical error
7 0 0. No preference {VOL -
video/phone only}
2,865 1 1. Democrat
2,563 2 2. Republican
2,527 3 3. Independent
269 5 5. Other party {SPECIFY}
(325 differences between V201228 and PID3)
PRE: Party ID: Does R |
think of self as | RECODE of V201228 (PRE: Party ID: Does R
Democrat, Republican, | think of self as Democrat, Republican,
or Independent | 1 2 3 . | Total
----------------------+--------------------------------------------+----------
-9. Refused | 0 0 0 44 | 44
-8. Don't know | 0 0 0 4 | 4
-4. Technical error | 0 0 0 1 | 1
0. No preference {VOL | 0 0 0 7 | 7
1. Democrat | 2,865 0 0 0 | 2,865
2. Republican | 0 2,563 0 0 | 2,563
3. Independent | 0 0 2,527 0 | 2,527
5. Other party {SPECI | 0 0 0 269 | 269
----------------------+--------------------------------------------+----------
Total | 2,865 2,563 2,527 325 | 8,280
RECODE of |
V201228 |
(PRE: |
Party ID: |
Does R |
think of |
self as |
Democrat, | RECODE of V202110x (PRE-POST:
Republican | SUMMARY: 2020 Presidential vote)
, | 0 1 . | Total
-----------+---------------------------------+----------
1 | 2,244 77 544 | 2,865
2 | 179 1,808 576 | 2,563
3 | 1,035 625 867 | 2,527
. | 79 72 174 | 325
-----------+---------------------------------+----------
Total | 3,537 2,582 2,161 | 8,280
Obs<.
+------------------------------
| | Unique
Variable | Obs=. Obs>. Obs<. | values Min Max
-------------+--------------------------------+------------------------------
PID3 | 325 7,955 | 3 1 3
VOTETB | 2,161 6,119 | 2 0 1
-----------------------------------------------------------------------------
Missing-value patterns
(1 means complete)
| Pattern
Percent | 1 2
------------+-------------
72% | 1 1
|
24 | 1 0
2 | 0 0
2 | 0 1
------------+-------------
100% |
Variables are (1) PID3 (2) VOTETBOdd Stata issue
In Stata, missing values are coded as a period, but, in logical operations, Stata treats a missing value as if the missing value were infinitely large. This can cause errors like in the analysis below, which erroneously suggests that two cards in the Primm et al. 2010 dataset of football card values had values above $500:
. tab NAME CARD_YEAR if PRICE_2007 > 500
| CARD_YEAR
NAME | 1965 1989 | Total
----------------------+----------------------+----------
namath, joe | 1 0 | 1
thomas, eric (316 A.. | 0 1 | 1
----------------------+----------------------+----------
Total | 1 1 | 2
. tab NAME if PRICE_2007==.
NAME | Freq. Percent Cum.
-----------------------------+-----------------------------------
thomas, eric (316 A&B)*** | 1 100.00 100.00
-----------------------------+-----------------------------------
Total | 1 100.00
The data row in Stata for the Eric Thomas card was missing a value for PRICE_2007. Nonetheless, the Eric Thomas card was misleadingly caught in the logical restriction of “PRICE_2007 > 500”.