This dataset, which was extracted from the 2000 General
Social Survey (GSS) (Smith et al., 2019), reports the responses of
adults in the United States to seven questions about legalized
abortion. The questions began, ``Please tell me whether or not you
think it should be possible for a pregnant woman to obtain a legal
abortion if...'' The abortion items were given to a random two-thirds
subsample of GSS participants, so about 33% of the values are NA
by design. Refusal to answer the question (a rare occurrence) was
also coded here as NA.
The data frame also includes variables on age, sex, race, Hispanic origin, education, religious and political affiliation. A second race item, which was modeled after the race question on the U.S. Census questionnaire, was given to a random half-sample.
In general, analyses of GSS data should account for the complex sample design. Sample weights, stratum and cluster indicators are included for this purpose.
abortion2000a data frame with 2,817 rows and 19 variables:
Agerespondent's age, a factor with levels "18-29",
"30-49", "50-64", and "65+"
Sexrespondent's sex, a factor with levels
"Female" and "Male"
Racerespondent's race, a factor with levels "White",
"Black", and "Other"; see NOTE below
CenRacerespondent's race, a factor with levels
"White", "Black", "Hisp" and "Other";
see NOTE below
Hisprespondent's Hispanic classification, a factor
with levels "nonHisp" and "Hisp"
Degreerespondent's education, a factor with levels
"<HS" (did not finish high school),
"HS" (high school diploma), "JunCol" (junior
college), "Bach" (Bacheor's degree), and "Grad"
(graduate degree)
Religrespondent's religious preference, a factor
with levels "Prot" (Protestant),
"Cath" (Roman Catholic), "Jewish", "None", and
"Other"
Partyrespondent's political party identification, a
factor with levels
"Dem" (Democrat), "Rep" (Republican), and
"Ind/Oth" (Independent or Other); see NOTE below
PolViewsrespondent's political views, a factor with
levels "Con" (Conservative), "Mod" (Moderate), and
"Lib" (Liberal)
Each of the next seven variables below is a factor with levels
"Yes", "No", and "DK" (don't know). The items
were prefixed by, ``Please tell me whether or not you think it
should be possible for a pregnant woman to obtain a legal abortion
if...''
AbDefect``...If there is a strong chance of serious defect in the baby?''
AbNoMore``...If she is married and does not want any more children?''
AbHealth``...If the woman's own health is seriously endangered by the pregnancy?''
AbPoor``...If the family has a very low income and cannot afford any more children?''
AbRape``...If she became pregnant as a result of rape?''
AbSingle``...If she is not married and does not want to marry the man''
AbAny``...The woman wants it for any reason?''
The three variables below may be used to compute estimates and standard errors that account for the survey's complex sample design:
WTSSALLnumeric sampling weight, inversely proportional to the individual's probability of being selected into the sample
VSTRATinteger code identifying the stratum for variance estimation
VPSUinteger code identifying the primary sampling unit (PSU) (i.e., the primary cluster) within stratum for variance estimation; see NOTE below