GSS Project- You Pick The Variable

Analyzing Survey Data: Guidelines

Before you begin this exercise, which is worth 100 points, you should have read all the materials about survey research and official statistics within the folder called Part 3, Strategies of Social Research -- Survey Research," and have completed quiz 3.

The General Social Survey (GSS)

You will be conducting analysis of data from the GSS. Follow the steps outlined bellow:

1. Find out what the General Social Survey (GSS) is, and figure out why it is such an important source of data for understanding American society. Write a brief paragraph that describes the GSS. Here is a link that provides you with background about the GSS. Do NOT use this site for analysis of the data.

·

· http://www.norc.org/projects/general+social+survey.htm

2. To analyze GSS data, we will be using a site at University of California, Berkeley. This site is maintained as part of the Survey Documentation and Analysis project that brings together and archives publicly accessible data from a variety of studies. SDA also provides online statistical analysis software to conduct simple and complex analysis of data. We will use their software to analysis GSS data for this assignment. Here's the URL:

http://sda.berkeley.edu/cgi-bin/hsda?harcsda+gss06

3. Formulate three logically connected hypotheses. You may formulate hypotheses about whatever topic you wish. However, remember that GSS provides us with survey data, that is, it consists of responses that people in a sample of adult American population have given to sets of questions designed measure what American think about a variety of topics -- mostly political, social and religious. However, for specific years, modules that deal in some detail with special topics are available; for example, there are two modules on environment, one on social support networks, science and other topics.

Your hypotheses should be accompanied by a brief rationale, that is, your reasoning behind each of your hypotheses. Here's an example of the format that this section of your exercise should follow. Remember you may select any topic for which there are data from the GSS.

A SIMPLE ILLUSTRATION (Yours might be more complicated)

I will use a topic that has been researched thoroughly -- attitudes about abortion. We are learning about social research methods, so our task is to locate our topic in society. Following Durkheim's suggestions, we think level of education, racial identity, gender, wealth and other background considerations (variables) influence how individuals think about a given topic. In this case, the topic is attitude about abortion. The GSS has been measuring attitudes about abortion since 1977. Attitude about abortion is my dependent variable. I hypothesize that where a person is located in society (gender, age, social class, etc.) should predict how they think about abortion. My task, and yours, is to find variables in the GSS that test hypotheses. Here are my hypotheses:

Hypothesis One: Educational level should be associated with attitude toward abortion. Specifically, the more education a person has achieved, the more likely he or she will be accepting of a women's decision to have an abortion.

Rationale: Education expands a person's ability to understand complex situations and increases their analytic skills. Education also increases a person's ability to "take on the role of the other" and, hence, understand how abortion might be a part of a reasonable way to make sense out of being pregnant. Or, from the literature, I know that education has a "liberalizing effect in the US," and liberals are more likely to favor a women's right to a legal abortion than conservatives.

Hypothesis Two: Gender should be associated with attitude toward abortion. Women should be more accepting of abortion than men.

Rationale: Women also are more attentive to the politics of abortion than men because the issue may directly affect their lives. Their responses to questions about legal abortion should exhibit a distinctive pattern, that is, be different from how men respond.

Hypothesis Three: A person's religious beliefs should be associated with their attitude about abortion.

Rationale: Religion is a powerful social force in the lives of some members of society. They use their understanding of the teachings of their religion to make sense out of controversial subjects. The more religious a person, the more likely their responses will form a distinctive pattern.

Mitigating Circumstance (Control or Test): Some social facts are more powerful influences than others. In American society, given the long history of racial discrimination and economic exploitation of American Africans, racial identity is a very powerful social fact. Therefore, the above three hypotheses might be influenced by whether or not the person who responded to the question about abortion is black or white.

Now, the above hypotheses and the qualifying circumstance form a theory about attitudes toward abortion. It is not a very sophisticated theory and it should be better grounded in literature, that is, in fully developed research project, I would cite previously published findings and I would review existing theories about the relationships between social facts and attitudes. Not necessary for this exercise since we are two goals: 1) learn how to use the GSS, and 2) appreciate the value of survey research for finding patterns and testing theories.

STEP ONE:

Find variables (questions asked of people who were part of the GSS sample) that fit with our theoretical problem (hypotheses).

My dependent variable is attitude toward abortion:

The variable I found is called ABANY (this is code, also called a mnemonic) that stands for the question asked. ABANY is my dependent variable.

ABANY: ABORTION IF WOMAN WANTS FOR ANY REASON - 206. Please tell me whether or not you think it should be possible for a pregnant woman to obtain a legal abortion if: g. The woman wants it for any reason? 0 NAP 1 YES 2 NO 8 DK 9 NA

My independent variables are education, gender and religion. It turns out that GSS has several ways to measure a person's education. EDUC is coded in years of schooling for 0 to 21. If you use this variable, you will have to recode it into categories, for example HS, College and Graduate. (There are instructions on how to recode at the SDA analysis site). It is helpful to remember that there are different levels of measurement: nominal or categorical and continuous (forget about rank order for this exercise). If you try to cerate a table using EDUC and do not recode it, you will get a table with 22 columns and two rows. The table is unreadable. Therefore, you should recode into a three or four categories.

I will be using cross-tabulation as my analytic tool. DEGREE is another measure of education and it is already coded for cross tabulation. I will use DEGREE.

Gender is easy. The mnemonic is SEX and it is coded for cross tabulations as Male and Female (the variable is nominal or categorical).

The GSS has many questions about religion, and quite sophisticated analysis is possible. Let's keep it simple. People were asked if they thought of themselves and fundamentalists, moderate or liberals in their religious beliefs. I will use this variable for "religion." It is called FUND in the GSS.

RACE is the mnemonic for racial identity. It is coded Black, White and Other. When an interviewer enters a subject's home, the code the respondent as black, white or other. If they can't tell, they ask the respondent. Interestingly, the proportion of the sample that is Black is very close to US census data (12%).

Build a table of your variables. Here is my table:

Variables	Question
Independent Variables: DEGREE SEX FUND	Coded: Highest degree obtained by interviewee. LS than HS; HS, junior college, college, graduate). Coded: Male or Female Coded: Fundamentalist, Moderate and Liberal
Dependent Variable: ABANY	"Woman has right to legal abortion for any reason." Coded as "yes" or "no"
Control or Test Variables: RACE	White, Black, Other

Table One: Variables Used to Test Hypotheses (GSS 2010)

STEP TWO:

Discover the patterns for the variables you have selected. First make sure all the variables you have selected are available for the year you want to analyze. For example, in my example all the variables are available for 2010, which is the most recent data for the GSS. You can check which years your variable is available by creating a table with your variable as the row and "YEAR" as the column.

Create a descriptive Table for your variables: My looks like this because I am using nominal variables. If you use continuous variables, you will have means and standard deviations:

Variables	Descriptive Statistic (%)
DEGREE	LH = 15.1%; HS = 50.2%; JC = 7.2%; Bachelor = 17.5%; Graduate = 10%. n= 2,045.
SEX	Male = 45.2%; Female = 54.8%. n=2,045.
FUND	Fundamentalists = 24.6%; Moderate = 42.1; Liberal = 31.5%. n= 1,939
ABANY	Yes = 42.9%: no = 57.1%. n= 1,234.
RACE	White = 75.7%; Black = 14.6%; Other = 9.7%. n = 2,045.

Table Two: Descriptions for Variables Use to Test Theory

STEP THREE:

Now create tables (or charts, if you feel creative) for each first order association: ABANY by DEGREE, ABANY by FUND, and ABANY by SEX. From the outputs you get, create your own table or chart that shows the results. Here's my example:

DEGREE	Less than High School = 32.4% High School = 34.3% Junior College = 40.2% Bachelors degree = 52.7% Graduate degree = 57.6%
FUND	Fundamentalist = 30.8% Moderate = 35.8% Liberal = 60%
SEX	Male = 45.2% Female = 40.9%

TABLE THREE: Percentage Yes, to "Woman should have right to legal abortion for any reason" (GSS 2010, n= 1,234)

STEP FOUR:

Run cross tabulation again with RACE as a control variable. Create a table showing the results. This can be complicated. Remember that you are applying the Elaboration Model, so you will want to pay attention to how the relationships change with different racial identifications.

Figure One: Percent Yes for Blacks and Whites by Degree (GSS 2010)

Figure Two: Percent Yes for Blacks and Whites by Religious Belief (GSS, 2010)

Figure Three: Percent Yes for Blacks and Whites by Gender (GSS 2010)

STEP FIVE:

Now comes the interesting part. Interpret your results. In my little study, I found that education and religion are associated with attitude toward abortion. The more education a person has the more likely they are to agree that a woman should have a right to a legal abortion for any reason. And, if a person says their religious beliefs are liberal, they are much more likely to agree with the question. While males are more likely to say they agree, the difference between men and women is very small and probably not significant. These associations occur within ranges of percentages from 30% of fundamentalists agreeing to a high of nearly 60% of liberals. The range for the other two variables is less wide (32% to 58% for education and only 41% to 45% for gender). These are the first order or initial associations that I discovered.

What happens when a test or control variable is introduced (the elaboration model)? For DEGREE, I discovered the same association for blacks and Whites, but a stronger one for blacks than whites. For religious beliefs, I found a different pattern. Black fundamentalists are much more likely to agree with a legal abortion for a women for any reason, than white fundamentalists. For moderates and liberals, the association is similar. For gender, blacks and whites have pretty the same association, that is, black men and black women show a similar pattern and as do black and white men.

Partial Relations Compared to Original	TEST OR CONTROL VARIABLE Antecedent Intervening
Same Relations	Replication	Replication
Less or none	Explanation	Interpretation
Split	Specification	Specification

Following the elaboration model (above) we can reason that racial identity is an antecedent condition, and for degree we replicated the original association, for religious belief, we specified the original, and for gender we replicated a very weak association.

SUMMARY

Your exercise should follow the above STEPS. It should consist to

1. A brief paragraph or two about the GSS. What is it, what does it do, why is it important for social research, how will you use it?

2. Your theory, three hypotheses with accompanying rationales.

3. Your variables, descriptions with table.

4. Your original or first order findings with chart or table.

5. Your control variable and results (charts or tables).

6. Conclusions about what you find and how your control variable affected the results.

Here's a list of possible independent and dependent variables that you can use for your analysis. You may use other variables if your hypotheses are not covered by any of the variables listed below.

Independent Variables (information from the GSS that reflects the conditions of urbanization that Wirth outlines in his article):

FINRELA: Opinion about family income. Is your family's income far above average, above average, average, below average, far below average? This is a Likert scale. It correlates closely with actual income and can be used as a measure of inequality.

AGE: ranges from 18 to 89. This is a continuous variable which you will have to recode, e.g., age (r: 18 - 30 "young"; 31- 50 "adult"; 51 - 89 "Elderly") or whatever categories best reflect your theoretical reasoning about the effects of age.

RACE: This variable is coded as white, black and other. The interviewer codes the interviewee accordingly, asking only if not obvious. There are other more sensitive measures of race on the GSS, but this one is the simplest and most widely used. It also is consistent with the US census; for example, about 12% of the GSS sample is coded as "black," and that is the percentage of the US population that is black according to the census.

SEX: Dichotomous variable (male or female).

DEGREE: highest degree obtained by interviewee (LS than HS; HS, junior college, college, graduate).

SRCBELT: This is a code that stands for where the respondent lived at the time of the interview

SRC BELTCODE - 0 NOT ASSIGNED 1 12 LRGST SMSA'S 2 SMSA'S 13-100 3 SUBURB, 12 LRGST 4 SUBURB, 13-100 5 OTHER URBAN 6 OTHER RURAL

XNORCSIZ: This variable is a code that NORC used for the size of the place where the respondent lived at the time of the interview.

EXPANDED N.O.R.C. SIZE CODE - 0 NOT ASSIGNED 1 CITY GT 250000 2 CITY,50-250000 3 SUBURB, LRG CITY 4 SUBURB, MED CITY 5 UNINC,LRG CITY 6 UNINC,MED CITY 7 CITY,10-49999 8 TOWN GT 2500 9 SMALLER AREAS 10 OPEN COUNTRY

RES16: This variable gives a rough indication of where a person grew up.

TYPE OF PLACE LIVED IN WHEN 16 YRS OLD - 25. Which of the categories on this card comes closest to the type of place you were living in when you were 16 years old? 0 NAP 1 COUNTRY,NONFARM 2 FARM 3 TOWN LT 50000 4 50000 TO 250000 5 BIG-CITY SUBURB 6 CITY GT 250000 8 DK 9 NA

COMTYPE: TYPE OF COMMUNITY IN WHICH R LIVES. Would you describe the place where you live as... 0 NAP 1 BIG CITY 2 SUBURBS, OUTSKIRTS 3 SMALL TOWN 4 COUNTRY VILLAGE 5 FARM, CNTRY HOME 8 DK 9 NA

Dependent Variables: You will have to figure out which dependent variable you want to use based on how you interpret the Wirth hypothesis. Here are some that might be useful. Of course there are many questions on the GSS that should reflect urbanism as a way of life. When you find a variable you want to use, make sure you know the year that it was asked. Your independent and dependent variables must have been asked in the same year in order the get a cross tabulation.

ABANY: ABORTION IF WOMAN WANTS FOR ANY REASON - 206. Please tell me whether or not you think it should be possible for a pregnant woman to obtaina legal abortion if: g. The woman wants it for any reason? 0 NAP 1 YES 2 NO 8 DK 9 NA

GUNLAW: FAVOR OR OPPOSE GUN PERMITS - 86. Would you favor or oppose a law which would require a person to obtain a police permit before he or she could buy a gun? 0 NAP 1 FAVOR 2 OPPOSE 8 DK 9 NA

TEENSEX: SEX BEFORE MARRIAGE -- TEENS 14-16 - 217a. What if they are in their early teens, say 14 to 16 years old? In that case, do you think sex relations before marriage are always wrong, almost always wrong, wrong only sometimes, or not wrong at all? 0 NAP 1 ALWAYS WRONG 2 ALMST ALWAYS WRG 3 SOMETIMES WRONG 4 NOT WRONG AT ALL 5 OTHER 8 DK 9 NA

PREMARSX : SEX BEFORE MARRIAGE - 217. There's been a lot of discussion about the way morals and attitudes about sex are changing in this country. If a man and woman have sex relations before marriage, do you think it is always wrong, almost always wrong, wrong only sometimes, or not wrong at all? 0 NAP 1 ALWAYS WRONG 2 ALMST ALWAYS WRG 3 SOMETIMES WRONG 4 NOT WRONG AT ALL 5 OTHER 8 DK 9 NA

BIBLE: FEELINGS ABOUT THE BIBLE - 120a. Which of these statements comes closest to describing your feelings about the Bible? 1. The Bible is the actual word of God and is to be taken literally, word for word. 2. The Bible is the inspired word of God but not everything in it should be taken literally, word for word. 3. The Bible is an ancient book of fables, legends, history, and moral precepts recorded by men. 0 NAP 1 WORD OF GOD 2 INSPIRED WORD 3 BOOK OF FABLES 4 OTHER 8 DK 9 NA

You can search of variable in a number of different ways. There is a search engine on the SDA cite. It is literal; you can use a wildcard (*). There are folder labeled by topic which you can open and see the variable names.

White (n =952) Less Than HS High School Junior College College Graduate 35.1 39.6 41.9 51.4 54.8 Black (n =167) Less Than HS High School Junior College College Graduate 35 39.54 35.979999999999997 69.3 72.7

White (n = 917) Fundamentalist Moderate Liberal 26.5 36.799999999999997 61.3 Black (n= 158) Fundamentalist Moderate Liberal 41.9 38.4 55 White (n= 952) Male Female 45.6 41.3 Black (n =167) Male Female 47.6 41