statics

Part III BIVARIATE ANALYSIS Supplemental Material A

Should Abortion Be a Crime? What's Associated with Different Attitudes Toward Abortion? One of the most controversial issues about crime and justice in recent years is whether abortion should be a crime. Millions of women have had legal abortions in the past several decades, but many Americans feel that abortion should still be a criminal offense, as it once was. If abortion were "recriminalized," thousands of doctors and other health workers and millions of women would be affected. Partisans on both sides of this issue are often extremely vocal and demonstrative. As we saw in the supplemental reading for Part I, there appear to be three main positions on this issue among the GSS respondents. Just under half support a woman's right to have an abortion for any reason, about 11% oppose abortion under all circumstances, and the rest are opposed to unrestricted abortions but are willing to make exceptions when pregnancy results from rape, risks the mother's health, or is likely to result in serious birth defects. In this chapter, we're going to use your new analytic skills to begin exploring the correlates of these different points of view on abortion. Why are some people permissive and others not? Again, it would be helpful to review the rich social science and criminal justice literature on this topic, but for the moment we will begin our exploration inductively.

IIIa.1 Gender and Abortion As you think about possible correlates of different attitudes about abortion, the first one that probably comes to mind is gender, given that abortion affects women more directly than men. There is reason to believe that women would be more supportive of abortion than men. Let's see. Here is a table that summarizes attitudes toward abortion by gender. It is not an SPSS output, but we've created it from several SPSS tables. We simply looked at the SPSS output and wrote down the relevant percentages to create this table. Your task is to figure out how to get the SPSS output that would allow you to create this table.

Percentage of respondents approving of abortion under certain circumstances: Sex

Variable Circumstance Men Women

ABHLTH Woman's health is endangered 91 88

ABDEFECT Serious defect is likely 78 78

ABRAPE Pregnancy resulted from rape 78 81

ABPOOR Family is too poor for more children 53 50

ABSINGLE Woman is unmarried 53 46

ABNOMORE Family wants no more children 52 51

ABANY For any reason 50 47

Contrary to what we expected, women are not more supportive of abortion than men are. Actually, men are, in most cases, a little more supportive, by a margin of one to two percentage points. We used the individual items concerning abortion for this analysis because it was possible that men and women would differ on some but not on others. And this is shown to be the case. We might have expected women to be more supportive when the pregnancy is a result of rape. The results confirm this with about 81% of the women in the sample approving of abortion in the case of rape; and slightly fewer, 78% of men supporting abortion in the case of rape. As an alternative strategy for examining the sources of attitudes toward abortion, let's make use of the index ABORT we created in Chapter 6, combining responses to ABDEFECT and ABSINGLE.

Based on the summary table, what would you expect to find? We see essentially no difference between men and women in the case of a birth defect. But in the case of unmarried pregnant women, men are more approving of abortion. Create the following table:

Recall that a score of 2 on the index represents respondents who supported a woman's right to have an abortion if her health is threatened by the pregnancy and if she is unmarried. So we labeled this as a “pro-choice” attitude. Overall, 49% of the sample took that position. This table indicates some difference between men and women, with men more likely (52%) to score 2 on the index than women (46%). So far, then, we have learned that gender is not an explanation for all differences in attitudes toward abortion, but for some. Let's see if age has an impact.

IIIa.2 Age and Abortion

Abortion is somewhat more relevant to young people, as they are more likely to experience unwanted pregnancies than older people are. What would this lead you to expect in the way of a relationship between age and support for abortion? Think about that, and then use SPSS to run the tables that answer the question for you.

Recall that we recoded AGE into AGECAT in Chapter 4. We'll want to use AGECAT for our examination of the relationship between age and abortion attitudes. Now you can request the SPSS tables that relate age to abortion attitudes. Here's a summary of the tables you should have created.

Percentage of respondents approving of abortion under certain circumstances: Unde

Age Ove

Variable Circumstance 21 21-39 40-64 64 ABHLTH Woman's health

is endangered 85 92 87 94

ABDEFECT Serious defect is likely 62 79 76 87

ABRAPE Pregnancy resulted from rape 77 83 77 84

ABPOOR Family is too poor for more children 31 49 58 44

ABSINGLE Woman is unmarried 39 45 56 38

ABNOMORE Family wants no more children 25 43 61 50

ABANY For any reason 31 44 57 36

Take a minute to look over the data presented in this summary table. How do the analytic results square with your expectations? These data suggest that there is little relationship between age and attitudes toward abortion. This is confirmed when we use the ABORT index as our measure.

To continue our investigation of age effects, let's use the ABORT index.

Ultimately, the relationship between age and abortion attitudes is a bit more complex This analysis is limited and we will more once we are able to undertake multivariate analysis.

IIIa.3 Religion and Abortion If we began this analysis by asking you what variable you thought might account for attitudes toward abortion, there is a good chance that you would have guessed religion, given the unconditional and public opposition of the Roman Catholic Church. Your growing facility with SPSS and the GSS data makes it possible for you to test that expectation.

Let's start with the possible impact of religious affiliation. Here's a summary of what you should discover if you run the several abortion items by RELIGCAT.

Percentage of respondents approving of abortion under certain circumstance:

Religion

Variable Circumstance Prot. Cath. .

Jew Other None

ABHLTH Woman's health is endangered 91 85 100 96 89

ABDEFECT Serious defect is likely 77 72 100 71 90 ABRAPE Pregnancy resulted from

rape 80 73 100 76 100

ABPOOR Family is too poor for more children 51 36 80 55 73

ABSINGLE Woman is unmarried 46 37 80 62 68 ABNOMORE Family wants no more

children 54 36 80 57 66 ABANY For any reason 47 35 80 57 66

There are several observations you might make about these data. To begin, the expectation that Catholics would be the most opposed to abortion is confirmed. They are the least likely to approve of abortion under any of the conditions asked about. At the same time, the level of support for abortion among American Catholics is greatly at variance with the church's official position. Under the traumatic conditions summarized at the top of the table, 72% to 85% of the Catholics say they would approve abortion. Even under the less traumatic conditions, at least a third of the Catholics would support a woman's right to an abortion: 35% support it "for any reason," Finally, under the nontraumatic circumstances, Protestants and Catholics are not so different from one another, especially in contrast to the much greater support for abortion among Jews and those with no religion. To examine this relationship further, use the index ABORT to see if this pattern is reflected in scores on a composite measure. In our earlier examinations of religion, we've sometimes gone beyond affiliation to examine the measure of religiosity, or religiousness. How do you suppose church attendance would relate to abortion attitudes? To find out, let's use CHATT, the recoded variable created in Chapter 4.

If you run the appropriate tables, here's a summary of what you should have learned.

Percentage of respondents approving of abortion under certain circumstances:

Frequency of Church Attendance

Variable Circumstance Never Yearly Monthly Weekly

ABHLTH Woman's health is endangered 97 93 89 80

ABDEFECT Serious defect likely 95 90 67 58 ABRAPE Pregnancy resulted from

rape 95 90 80 58 ABPOOR Family is too poor for

more children 66 65 42 30 ABSINGLE Woman is unmarried 61 64 41 27 ABNOMORE Family wants no more

children 69 65 46 26 ABANY For any reason 67 64 38 23

What does this table tell you about religion and abortion attitudes? The overall relationship is pretty clear: Increased church attendance is related to decreased support for abortion. There is a substantial difference, however, between those who attend church about weekly and those who attend less often. Those who attend one to three times a month are much more likely to support unconditional abortion as those who attend church weekly. Even in the case of the traumatic conditions at the top of the table, only those who attend church weekly stand out in their relatively low level of support. The fact that the other three groups do not differ much from one another, by the way, is a result of what we call a ceiling effect. Whenever the overall percentage of people agreeing with something approaches 100%, it's not possible for there to be much variation among subgroups. In the extreme case, if everyone agreed on something, there would be no way for men and women to differ, as 100% of both would have to agree. Similarly, there could be no differences among age groups, religions, and so on. When the overall percentage approaches zero, a similar situation pertains, which we call a floor effect. The GSS data we've provided for your use permit you to explore this general topic further, if you wish. Why don't you check out the effect of POSTLIFE and PRAY on abortion attitudes? Before running the tables, however, take some time to reflect on what might logically be expected. How should beliefs about an afterlife affect support for or opposition to abortion? You may be surprised by what you learn. Then again, maybe you won't be surprised.

IIIa.4 Politics and Abortion There is a strong and consistent relationship between political philosophy and abortion attitude. What do you suppose that relationship is? Who would you expect to be the more supportive of abortion, liberals or conservatives? To carry out this investigation, you'll probably want to use the recoded variable POLREC.

Now you can examine the relationship between political philosophy and abortion attitude to see if your hunch is correct. Here's a summary of what you should discover.

Percentage of respondents

approving of abortion under certain

circumstances:

Political Philosophy

Variable Circumstance Liberal Moderate Conservative

ABHLTH Woman's health is endangered 94 89 86

ABDEFECT Serious defect is likely 86 83 64

ABRAPE Pregnancy resulted from rape

87 85 68

ABPOOR Family is too poor for more children 70 51 33

ABSINGLE Woman is unmarried 65 50 32 ABNOMORE Family wants no more

children 62 57 33 ABANY For any reason 63 51 31

How would you describe these results? Try your hand at writing a sentence or two to report on the impact of political philosophy on abortion attitudes. To pursue this relationship further, you may want to use the ABORT index. See if you can explain that too! Another direction you might want to follow in investigating the relationship between politics and abortion attitudes concerns political party identification. As you no doubt realize, the Democratic party has been generally more supportive of legal access to abortion than the Republican party has. How do you suppose the official differences separating the parties show up in the attitudes of the respondents in this sample? Among the general public, who do you suppose are the most supportive of abortion, Democrats or Republicans? You can now find out for yourself.

IIIa.5 Other Factors to Explore A number of other demographic factors can be examined for their effect on attitudes toward abortion. We'll suggest a few more for you to check out.

Education is strongly related to abortion attitudes. In which direction do you suppose that relationship goes? Why do you suppose that is? Race is another standard demographic variable that might be related to abortion attitudes. You may want to examine the individual abortion items because race has slightly different effects on different items. But you can also look at the relationship between race and the ABORT index variable. We suggest that you might like to explore the relationship between abortion attitudes and some family variables. How do you suppose abortion attitudes relate to respondents' views of the ideal number of children to have? The family variable that may surprise you in its relationship to abortion attitudes is MARITAL. You may recall that we also uncovered a surprising effect of marital status on political philosophy. Check this one out, and we'll take a more in-depth look once we begin our multivariate analyses. IIIa.6 Sexual Attitudes and Abortion Recalling that abortion attitudes are related to differences in political philosophy, it might occur to you that other philosophical differences might be relevant as well. The GSS data set contains two items dealing with sexual permissiveness or restrictiveness:

PREMARSX Attitudes toward premarital sex HOMOSEX Attitudes toward homosexual relations

Here's the relationship between attitude toward premarital sex and attitude toward abortion if a woman wants one for any reason (ABANY) from the SPSS printout. Notice the extent to which permissiveness on one issue is related to permissiveness on the other. Try creating a summary table of this information on your own.

Why don't you check to see if the same pattern holds attitude toward homosexuality?

IIIa.7 Capital Punishment and Abortion One of the most intriguing issues in the debate about abortion concerns the attitudes people hold about the value of human life. Would you hypothesize that a person who opposes capital punishment would also oppose abortion? Wouldn't that stance be consistent in opposing the deliberate ending of human life or potential human life? Or would you expect that people who oppose capital punishment might actually support the right to choose an abortion, arguing that capital punishment ends an independent human life whereas abortion eliminates a fetus, not an independent human being? These are powerful and dramatic issues, at the very center of contemporary debates about crime, justice, and freedom. With your new facility with SPSS and with data from the General Social Survey, you can test your hypotheses against a national representative sample.

IIIa.8 Summary In this supplement to the text, you've had an opportunity to search for understandings of the vast differences in people's feelings about abortion, one of the most contentious issues concerning crime and justice. We've found that religion and politics are powerful influences. We've also just seen that degree of permissiveness regarding abortion is strongly related to degree of per- missiveness on issues of sexual behavior. And we have hypothesized that attitudes about capital punishment may play a role in shaping attitudes about abortion. Thus far, we've only opened up the search for understanding, limiting ourselves to bivariate analyses of whether two variables are associated in some way. In the analyses to come, we'll dig ever deeper into the reasons for differences in the opinions people hold. Ultimately, you should gain a well-rounded understanding of the logic of criminal justice research as well as master some of the fundamental techniques for acting on that logic through SPSS.

Key Terms

ceiling effect

floor effect

Part III BIVARIATE ANALYSIS Supplemental Material B Suggestions for Further Bivariate Analysis By now, you've amassed a powerful set of analytic tools. In a world where people make casual assertions about crime and the criminal justice system, you're now in a position to determine the facts. You can determine how the U.S. population feels about a variety of topics dealing with crime and justice, and with your new bivariate skills you can begin to explain why they feel as they do. You can also examine the basic characteristics of the justice system, comparing national trends with those in your own state or local area.

In this supplemental material, we are going to suggest some additional analyses you might undertake. They will allow you to perfect your skills, and these suggestions open the possibility of your thinking more for yourself. What are you interested in? What would you like to learn more about? Here are some possibilities. Earlier in this supplement we suggested some topics you might pursue with the techniques of bivariate analysis. Let's start by returning to those topics.

IIIb.1 Attitudes About Sexual Behavior You might want to focus on the two sexual variables. What do you suppose would cause differences of opinion regarding premarital sexual relations and homosexuality? You have the ability and the tools to find out for yourself.

Consider all the demographic variables that may influence sexual attitudes as a way of beginning your examination: age, gender, race, religion, education, social class, and marital status, for example. Before examining each of these relationships, take some time to think about any links you might logically expect. Should men or women be more permissive about homosexuality? Should married, single, or divorced people be more supportive of premarital sex? How do you expect young and old people to differ? As you investigate these attitudes, be careful about assuming that the two items are just different dimensions of the same orientation. The kinds of people who are permissive about premarital sex are not necessarily the same ones who are permissive about homosexuality.

IIIb.2 Drug Use on Campus This study site contains the data set- "BINGE.SAV"- that will allow you to assess how much legal and illegal drug and alcohol use takes place on American college campuses. Using univariate analysis, you can first establish how many college students drink, binge drink, use marijuana, and smoke cigarettes. Next, you can examine patterns of use of all of these substances. Is it likely that students who binge drink will also use marijuana and cigarettes? Is there an association among the different patterns of substance use?

You can then move on to investigate whether factors that play a role in using one substance are also associated with using other substances. Does a student who thinks parties are an important or very important part of college life also tend to be one who binge drinks? Uses marijuana? Smokes cigarettes? What role do basic social identities play in these patterns of substance use? Are whites more or less likely to binge drink than blacks or Asian-Americans? What about students who live on campus compared with those who live at home? Is the image of the fraternity or sorority member who abuses alcohol more than non-Greek peers hopelessly outdated or an accurate image of a high-risk campus population? What other kinds of values play a role in shaping campus substance use or abuse?

IIIb.3 Date Rape and Sexual Assault on Campus Criminal justice research can be used to study why some groups of people are more likely to be the victims of serious crime than others. Often this information can point to ways of preventing victimization or at least lowering its risk. On college campuses today, many students are concerned about being the victims of sexual assault or date rape. The "BINGE.SAV" data set contains data about whether a student reports being the victim of sexual assault or date rape as the result of another student's alcohol use. These data open up some important analyses. How much does a student's own alcohol use or other substance use raise the likelihood of that student's being a victim of sexual assault or date rape? Are students who are members of Greek organizations more or less likely to be victimized? Does the risk of victimization increase or decrease with age, increased alcohol use, living on campus, or being an intercollegiate athlete?

IIIb.4 Crime and Justice Across the States The United States is a very diverse country, and its criminal justice systems reflect the great differences across its 50 states. Because most criminal justice activity is organized at the local or state level, understanding American criminal justice often means understanding local and state differences.

The data found in "JUSTICE.SAV" will allow you to begin investigating those social differences. How different are the rates of violent crime and crime against property across the 50 states? How are the rates of crime related to spending on criminal justice? Are the states with higher rates of crime more likely to spend more money on their criminal justice systems? Will high-crime states spend more on policing or on prisons? What regional patterns mark off the 50 states?

IIIb.5 Summary The suggestions in this supplement should keep you busy, but don't feel limited by them. The most fruitful guides to your analyses should be your own personal interests. Which of the topics most interest or concern you? Now you have a chance to learn something about them on your own. You don't have to settle for polemical statements about "the way things are." You now have the tools you need to find out for yourself

In examining these bivariate relationships, you may want to begin with Crosstabs because that technique gives you the most detailed view of the data. At the same time, you should use this exercise as an opportunity to experiment with the other bivariate techniques we've examined. For example, try chi-squares where appropriate, for a test of significance. You’ve also learned how to do t-tests and ANOVAs and Pearson’s r correlations. All of these allow you to conduct bivariate analysis on the issues that are most interesting to you.

What you've learned so far may be sufficient for most of your day-today curiosities. Now you can learn what public opinion really is on a given topic, and you can determine what kinds of people hold differing views on that topic. In the remaining chapters of this book, however, we are going to show you an approach to understanding that goes much deeper. As we introduce you to multivariate analysis, you're going to have an opportunity to sample a more complex mode of understanding than most people are even aware of.