{21:73.15 . . v‘ . .of—Eli'l to); L... 3‘ .4 I. t 4 .25. A o) . . 1.314;: ‘ ‘ an. £5.21): :3... . , «y..):;s . :3..ny2 . I! ~. ‘I. Ruin”. , . ‘1'... Iv... .3!!! . 3.5.3.1}ilbl .. fife-F: ..}..§.i .. I. . a 9:27.. .4 Llr. ET... .3535: .1... «hole...» i... r .v! :1; ti...) in IJ. .. ; 4%.. 7% 5 IA: :12. its 1. vi... .awhi:l’ adm’ A a. Iii.» .tt sill-l O... Tit?!) {)v-.\. ad , .u! ..o.. . cl 355:0‘9 |-t 4. .1 . . 73-15518 IIIIHIUIHIlllllllllHHlUllllIlllllillllllllllHlHlHllll 31293 00904 9309 This is to certify that the dissertation entitled THE EFFECTS OF A SUPREME COURT RULING ON PEOPLE'S ATTITUDES TOWARD ABORTION ISSUES presented by CHULHO YANG has been accepted towards fulfillment of the requirements for Ph . D . SOCIOLOGY degree in DNA/4 AA AAA Major professor Donald W. Olmsted Date May 27, 1993 MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771 —_————- —«—-— v- 'H—fl' ~m-r-G ——_——. - q—va—w—14 LIBRARY Michigan State l. University PLACE lN RETURN BOX to remove thts checkout from your record. TO AVOID FINES return on or before date due. nity Institution owns-p. MSU Is An Affirmative Action/Equal Opportu THE EFFECTS OF A SUPREME COURT RULING ON PEOPLE’S ATTITUDES TOWARD ABORTION ISSUES By Chulho Yang A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Sociology 1993 ABSTRACT THE EFFECTS OF A SUPREME COURT RULING ON PEOPLE’S ATTITUDES TOWARD ABORTION ISSUES BY Chulho Yang This study focuses on attitude changes concerning abortion issues for a three month period after the Supreme Court ruling on Webster vs. Reproductive Health Services in July 1989. The research was guided by three major questions: 1) is there any effect of a Supreme Court ruling on people's attitudes toward abortion issues, 2) if there is any effect of the Supreme Court ruling on people’s attitudes toward abortion issues, how does that attitude change occur, 3) is there any race difference in attitudes toward abortion issues. The data were provided by the Inter-University Consortium for Political and Social Research (ICPSR) at Ann Arbor, Michigan. CBS News/New York Times (producer of the data) conducted two abortion polls on nationwide samples using direct telephone interviews in 1989. The first survey was conducted three weeks after the Supreme Court ruling. The second survey was conducted two and a half months after the Supreme Court ruling. The research was carried out on two levels -- descriptive auui analytical. (hi the descriptive level, overall characteristitms of the data.were examined. (M1 the analytical level, eight major hypotheses were tested to analyze the effect of the Suprene:Court ruling, the contextual effects on attitude change and race differences in abortion attitudes. The following is a summary of the most significant findings. There was a significant increase in support of restriction on abortion by the states among those who had heard of the Supreme Court ruling on Webster vs. Reproductive Health Services compare with those who had not heard. Those who are religious 'were significantly'.l§§§_ supportive of abortion than those who are not religious. Black men were significantly less supportive of abortion than.white men, and black women were mgr; supportive of abortion than white women. The results were discussed in terms of two perspectives on the contextual effects and abortion.attitudes. The results suggested that the contextual effects may take place much faster than previously assumed. The results also suggested that those who describe themselves as "religious“ may not be as homogeneous in opinions on abortion as previously believed. Copyright by C HULHO YANG 1993 To my parents and my brother, Jaiho AC KNOWLEDGMENTS I wish to express my thanks to my major professor, Donald w. Olmsted for his guidance and encouragements throughout my training in the Department of Sociology, Michigan State University and this study. My other committee members, professors Thomas L. Conner, Ann G. Olmsted and Christopher K. Vanderpool, have helped me during this study in various ways. I sincerely thank my committee members for their support and trust during the Spring of 1992. I appreciate the patience and support of my wife, Insill, and the continuous interruptions of my children, Heeyeon and Ikchang. Their interruptions forced me to develop a fast and effective way of working in a short time period which I believe will be a valuable asset in my future professional career . TABLE OF CONTENTS Page I INTRODUCTION ..................................... 1 II. LITERATURE REVIEW ................................ 5 1. ATTITUDE ...................................... S 2. ABORTION ...................................... 8 3. THE EFFECT OF SUPREME COURT RULING ON PEOPLE’S ATTITUDES ......................... 12 4. THEORIES OF ATTITUDE CHANGE: REFERENCE GROUP, CONTEXTUAL AND PURELY INDIVIDUALISTIC REACTION THEORY ............................... 18 III. RESEARCH METHODS AND HYPOTHESES .................. 24 1. RESEARCH DESIGN .............................. 24 2. DATA SOURCE, CONDITIONS, MEASURES AND CATEGORIES .................................... 26 A. Data Source and Conditions ............... 26 B. Measures Relevant To Abortion Attitudes and Categories ............................ 28 3. RESEARCH HYPOTHESES .......................... 31 4. STATISTICAL METHODS AND MEASURES ............. 42 IV. DATA ANALYSIS AND HYPOTHESES TESTING ............. 44 V} INTERPRETATION OF RESULTS .................... .... 93 1. SUMMARY OF DATA ANALYSIS ..................... 93 2. EFFECTS OF THE SUPREME COURT DECISION ON ABORTION ATTITUDES ........................ 96 3. CONTEXTUAL EFFECTS ON PEOPLE’S ATTITUDES TOWARD ABORTION .................... 101 4. RACE DIFFERENCES IN ABORTION ATTITUDES ....... 106 VI. CONCLUSION ....................................... 109 1. SUMMARY AND DISCUSSION ....................... 109 2. SUGGESTIONS FOR FUTURE STUDY ................. 112 vii BIBLIOGRAPHY .......................................... APPENDICES APPENDIX A: SELECTED ITEMS USED IN THIS FROM CBS NEWS/NEW YORK TIMES ABORTION POLL, JULY 1989 ........................ 122 APPENDIX C: CBS NEWS/NEW YORK TIMES POLL National Survey Methodology 1989 ...... 125 viii 10 LIST OF TABLES Mean Score on Restrictiveness, by Earlier/Later and Heard/Not Heard Conditions ................... Correlations among Restrictiveness, Age, Education, Gender, Race, Religion, and “Heard“ and “Not Heard“: Earlier Category ........ Results of Regression of Restrictiveness Index on Age, Education, Gender, Race, Religion, and “Heard“ and “Not Heard“: Earlier Category ........ Correlations among Restrictiveness, Age, Education, Gender, Race, Religion, and “Heard“ and “Not Heard“: Later Category .......... Results of Regression of Restrictiveness Index on Age, Education, Gender, Race, Religion, and “Heard“ and “Not Heard“: Later Category .......... Correlations among “Abortion is Murder“, Age, Education, Gender, Race, Religion, and “Heard“ and "Not Heard“: Earlier Category ........ Results of Regression of “Abortion is Murder“ on Age, Education, Gender, Race, Religion, and “Heard“ and “Not Heard“: Earlier Category ........ Correlations among “Abortion is Murder“, Age Education, Gender, Race, Religion, and “Heard“ and “Not Heard“: Later Category .......... Results of Regression of “Abortion is Murder" on Age, Education, Gender, Race, Religion, and ”Heard“ and “Not Heard“: Later Category .......... Restrictiveness Score by Category Among Religious Protestant Who had "Heard“ ............. ix 47 52 53 58 62 63 65 66 71 11 12 13 14 15 16 17 18 Restrictiveness Score by Category Among Religious Catholic Who had "Heard“ ............... Justification Score by Category Among Religious Protestant Who had "Heard“ ............. Justification Score by Category Among Religious Catholic Who had "Heard“ ............... Restrictiveness Score by Religiosity Among Protestant Who had heard in the Later Category Restrictiveness Score by Religiosity Among Catholic Who had heard in the Later Category Justification Score by Religiosity Among Protestant Who had heard in the Later Category Justification Score by Religiosity Among Catholic Who had heard in the Later Category Mean Score of Justification, Gender and Race 73 75 86 89 LIST OF FIGURES Diagram of Effects on Restrictiveness of Earlier/ Later and Heard/Not Heard Categories in Hypotheses l and 2 ............................................. 36 Predicted Variance of Measures of Restrictiveness and Justifications Among Highly Religious People Who Had Heard of the Supreme Court Ruling on Webster vs. Reproductive Health Services in the Earlier and the Later Category ................................... 39 Predicted Variance of Measures of Restrictiveness and Justifications Among Less and Highly Religious People Who Had Heard of the Supreme Court Ruling on Webster vs. Reproductive Health Services in the Later Category ............................................. 40 xi CHAPTER I INTRODUCTION In and of itself, the “abortion issue“ reflects certain basic societal problems and thus deserves the interest of the sociological scholar. In addition, it appears that as the controversies related to abortion have emerged and developed over time in the United States, they have come to symbolize and stand for an expanding series of public issues reaching beyond abortion and encompassing such broad questions as the place of various religious perspectives in societal life and “philosophical“ views about the status relations among various religious, ethnic, economic and political groupings in the society. We posit that only by being aware of these broader ramifications can one account for the intensity of emotional arousal that the “abortion question'I engenders. Almost every day, somewhere in the United States, whether it is in front of an abortion clinic, a hospital or on the streets, the "war“ between those who have come to be called either “pro-choice“ or “pro—life“ is fought. In the 4,000- year-old history extending from.the Greeks and Romans through the Middle ages and into the present, every civilization has taken abortion with utmost seriousness. According to Rosenblatt (1991), the United States seems to be the only civilization to have engaged in an emotional and intellectual civil war over the abortion issue because of the general lack of consensus in the United States since the mid-60's, which has promoted bitter divisions over many social issues, such.as race, crime, war and abortion (Rosenblatt, 1992). According to Dayton (1991), during the colonial American period, abortion before quickening (defined in the early modern period as the moment when the mother first felt the fetus move) was not viewed as criminal, yet women and.men in colonial America acted covertly to effect abortions. The sexual revolution, the women's movement, and the pro—life movement of the 60’s brought the abortion issue into the center of the political arena. The landmark Supreme Court ruling on Roe vs. Wade in 1973 which established a woman’s right to terminate an unwanted pregnancy did not end the dispute over the abortion issue. It has been asserted that the Supreme Court rulings increase public support for the Supreme Court's position (Marshall, 1987; Casey, 1974; Adamany, 1973; Murphy' and Tanenhaus, 1969). However, several studies of the effect of the Supreme Court ruling on Roe vs. Wade found that support for legal abortions increased slightly (Jehnson and Canon, 1984), or crystallized issue preferences further and led to greater homogeneity of within-group attitudes toward abortion (Franklin and Kosaki, 1989). The "contextual theory" which states that one's social environment affects individual attitude formation (MacKuen and Brown, 1987), has been utilized to explain the impact of the Supreme Court ruling on the abortion attitudes by FTanklin and Kosaki (1989). It seems that the nature of available data limit clear explanation of the effect of the Supreme Court ruling in 1973 (this point will be discussed in detail in Chapter II: Literature Review). Sixteen years after the Roe vs. Wade ruling, the Supreme Court made another landmark ruling on abortion: Webster vs. Reproductive Health Services, which modified and narrowed the Roe vs. Wade ruling and gave more power to individual states to restrict abortion. After the Webster vs. Reproductive Health Services ruling in early July 1989, CBS News/The New York Times conducted.two abortion surveys on.a national sample using the telephone survey method. The first survey was conducted about three weeks after the ruling and the second was conducted about two and a half months after the ruling. The nature of data makes ideal for the test of contextual theory and the effect of the Supreme Court ruling (this point will berdiscussed.in detail.in Chapter II: Literature Review). Since the Roe vs. Wade ruling abortion was legalized” however, because of the Webster vs. Reproductive Health Services ruling the right to abortion was restricted somewhat. Thus, it is necessary to examine the effects of this restriction on 4 abortion on people’s attitudes toward abortion. The difference in abortion rates which black women were twice as likely to have abortion as white women, has held constant during the last decade (Henshaw'and Silverman, 1988), yet, previous studies showed that in general, blacks are less supportive of legal abortion than whites (Secret, 1987; Hall and Ferree, 1986; Combs and Welch, 1982). Since the racial issues are so important in the functioning of this society, it is necessary to examine the race difference in abortion attitudes. This study was intended to provide answers to three broad questions: 1. Is there any effect of a Supreme Court ruling on people’s attitudes toward abortion issues? 2. If there is any effect of the Supreme Court ruling on people’s attitudes toward abortion issues, how does that attitude change occur? 3. Is there any race difference in attitudes toward abortion issues? CHAPTER II LITERATURE REVIEW 1. ATTITUDE The term, attitude, is originated frothatin aptus which meaning includes ’fitness' or 'adpatedness'. One of the meanings of aptitude, which is the by—form of aptus, is that a subjective or mental state of preparation for action (i.e., mental attitude). The other refers to the outward or visible posture of a figure in statuary or painting (i.e., motor attitude). In recent years, the term, attitude, appears without a qualifying adjective in nearly all cases, and implicitly retains both its original. meanings: a :mental aptness and. a motor set. Thus, attitude refers to a neuropsychic state of readiness for mental and physical activity (Allport, 1985). Four criteria for an attitude has suggested by Park (see Young, 1931) as follows: 1. It has definite orientation in the world of objects (or values), and in this respect differs from simple and conditioned reflexes. 2. It is not an altogether automatic and routine type of conduct, but displays some tension even when latent. 6 3. It varies in intensity, sometimes being regnant, sometimes relatively ineffective. 4. It is rooted in experience, and therefore is not simply a social instinct (Young, 1931). The followings are some examples of definition for attitude which are accepted as typical definition of attitude: ...a mental disposition of the human individual to act for or against a definite object (Droba, 1933). ...the specific mental disposition toward an incoming (or arising) experience, whereby that experience is modified; or, a condition of readiness for a certain type of activity (Warren, 1934). ...a mental and neural state of readiness, organized through experience, exerting a directive or dynamic influence upon the individual's response to all objects and situations with which it is related (Allport, 1935). Attitude became, as Allport (1935) described it, “the most distinctive and indispensable concept in contemporary social psychology", mainly because of the presumed ability of attitudes to direct (and thus allow prediction of) human behaviors (Petty and Cacioppo, 1981). As Petty and Cacioppo (1981) argued that attitude, also, has important function in everyday social interaction: people are usually eager to hear our attitudes, and.we are often eager to give them“ 'There are a couple of reasons why attitudes are such an important part of social interaction. First of all, attitudes serve as convenient summaries of our beliefs. Secondly, knowing our 7 attitudes presumably helps others to know what to expect from us. Smith, Bruner, and White (1956) argued.that attitudes:may also express some important aspects of an individual’s personality. Katz (1960) has suggested four functions that attitudes might serve for a person as follows: 1) Ego- defensive function which are attitudes that are held because they help people protect themselves from flattering truths about themselves or about others who are important to them. 2) Value—expressive function which occurs when holding a certain attitude allows the person to express an important value. 3) Knowledge function which attitudes allow people to better understand events and people around them.. 4) Utilitarian function which attitudes help people to gain rewards and avoid punishments. Thus, according to Katz’s functional View of attitudes, different people may hold the same attitudes, but the attitudes may serve very different purposes for them (Petty and Cacioppo, 1981). As Petty and Cacioppo (1981) stated in their book, Attitudes and Persuasion, over the last forty years, a large number of theories have developed to explain how and why peoples' attitudes change. Among those theories, purely individualistic reaction theory and contextual theory could represent two opposite extremes. These two theories will be discussed further in Section 4: Theories of Attitude Change. 2 . ABORTION In 1742, in the village of Pomfret, Connecticut, nineteen-year-old Sarah.Grosvenor and.twenty-seven- year-old Amasa Sessions became involved in a liaison that led to pregnancy, abortion, and death. When Sarah’s pregnancy’ was known to Amasa, in consultation with John Hallowell, a self-proclaimed “practitioner of Physick“, he coerced his lover into taking an herbal abortifacient. Within two months, Sarah fell ill. Unbeknownst to all but Amasa, Sarah, Sarah’s sister Zerviah, and her cousin Hannah, Hallowell made an attempt to “Remove her Conception“ by a "manual operation“. Two days later Sarah.miscarried, and.her two young relatives secretly buried the fetus in the woods. Over the next month, Sarah struggled against a “Malignant fever“ and was attended by several physicians, but on September 14, 1742, she died. In 1745, two prominent Windham. County’ magistrates opened an investigation. into Sarah's death. 131 Nevember 1745, the investigating magistrates offered their preliminary opinion that Hallowell, Amasa Sessions, Zerviah.Grosvenor, and Hannah Grosvenor were guilty of Sarah's murder, the last three as accessories. The case went next to Joseph Fowler, King’s attorney for Windham County. A year and a half passed. as Fowler’s first attempts to get convictions against Hallowell and Sessions failed either before grand juries or before the Superior Court on technical grounds. Finally, in March 1747, Fowler presented Hallowell and Sessions separately for the “highhanded IMisdemeanor' of attempting to destroy both Sarah.Grosvenor's health and “the fruit of her womb". A grand jury endorsed the bill against Hallowell but rejected a similarly worded presentment against Sessions (Superioerourt of Connecticut, Box 9, 113, 173, 175; Windham County Superior Court, Box 172). The incident and the trial records show us some intriguing legal aspect of abortion and attitudes toward abortion in colonial America. First of all, Connecticut legal officials 9 concentrated pg; on the act of abortion per se, but on the fact that an abortion attempt had led to a young woman's death. It was because abortion before quickening (defined in the early modern period as the moment when the mother first felt the fetus move) was p9£_viewed by the English or colonial courts as criminal. Secondly, Fowler, the King's attorney for Windham County, moved to bring separate indictments against Hallowell and Sessions for the “highhanded ndsdemeanor“ of endeavoring to destroy Sarah’s health and I'the fruit of her womb“. These charges indicate that the English common law designation of abortion as a misdemeanor, not a felony or capital crime. Finally, the incident indicates that even though abortion before quickening was legal, women and.men in colonial America acted covertly to effect abortions. This indicate that abortion was understood as blameworthy because it was an extreme action designed to hide a prior sin, sex outside of marriage. Dayton (1991) also suggested that even married.women:may have hidden their abortion attempts because the activity was associated with lewd or dissident women (Dayton, 1991). Since Sarah Grosvenor's death, the legal aspect and the attitudes toward abortion have been changed drastically due to a women's movement and the “sexual revolution". However, the issue of abortion has not been settled yet. According to Rosenblatt (1992), in the 4,000-year-old history extending from the Greeks and Romans through the Middle Ages and into 10 the present, every civilization.has taken.abortion.with.utmost seriousness. However, in the United States, an emotional and intellectual civil war over the abortion issue has been presented since the sexual revolution of the 60’s mainly due to the general lack of consensus over the abortion issue. Rosenblatt (1992) argued that there are several reasons that make abortion an explosive issue in this country as follows: 1) Religiosity - America is, and always has been a religious country and the separation of church and state is so rooted in the democracy. According to Rosenblatt (1992), that separation has created and intensified a hidden national feeling about faith and God which is often more deeply felt and volatile than that of countries with official or state religions. The catholic Church seems more steadily impassioned.about abortion.in America than.anywhere else. IFor example, in the United States, Catholic church has allied itself with evangelical church in the pro-life position, in contrast with the Catholic church's usual effort to keep themselves distinguished from 'Fundamentalists". 2) Individualisnl- The idea.of the independent individual in.this country has been used by both the pro-choice and the pro—life advocates. The pro-life advocates argue that an individual has a right to have his or her own stance on a Hatter of conscience, and they see the pregnant woman as two individuals, each with an equal claim to the riches that American individualism offers. However, the pro-choice 11 advocates argue that people have the right to privacy and individual choice. 3) Optimism — The pro—life advocates argue that'every American, born or unborn, is entitled to look forward to a state of indefinite hope and progress. The pro- choice advocates interpret the ideas of hope and progress as a consequence of one’s entitlement to free choice in all things including abortion. 4) Sexuality — In a country like America where middle—class values were to become the standards and the moral engine of the country, abortion is a sign of careless sexuality and contradicts middle—class values, which dictate the rules of sexual conduct (Rosenblatt, 1992). Probably the two most important events in the war over abortion in the U.S.A. are the Supreme Court decisions on Roe vs. Wade in 1973 and.Webster vs. Reproductive Health Services in 1989. The first, Roe vs. Wade, established a woman's right to terminate an unwanted pregnancy and triggered a war between the pro-choice and pro-life movement camps. The later ruling, Webster vs. Reproductive Health Service, modified and narrowed the Roe vs. Wade ruling and invited an avalanche of state legislation that would attempt to test the each state’s right to control abortion (those above mentioned cases will be discussed in detail in a later section of this proposal) (Gillespie, 1989; Anderson, 1989; Salholz, et al., 1989). It has been asserted that Supreme Court rulings increase public support for the Supreme Court’s position (Franklin and Kosaki, 1989; Marshall, 1987; Casey, 1974). However, that is 12 not always true for the cases of abortion issues rather Supreme Court rulings on abortion usually ignite a new battle among the public. Thus, it would be interesting to find out what and how is the effect of the Supreme Court’s ruling on abortion issues on public. 3. THE EFFECT OF SUPREME COURT RULING ON PEOPLE’S ATTITUDE The, ”practical politics" of law-making’ in. the ‘U.S. system. This topic requires a brief statement about the relationships between the state and federal governmental levels. In the U.S., legislative activity at all levels takes place within the context of lobbying and ”pressure" by all kinds of organized “special-interest“ groups, and coalitions of such groups. These pressures are particularly important with respect to highly visible, emotionally-loaded issues. One of the clearest examples is the abortion issue. What it often comes down to is that citizen organizations which seek to influence the thrust and wording and impact of laws-—such as the various right-to-life or pro-choice organizations, have a better chance of being effective if they work at the state level rather than the federal level. For example, it is generally believed to be more feasible to arouse the voters and those who contribute funds to election campaigns, of a state legislator than it is to do the same 13 thing with respect to a U.S. congress-person. This, in turn, comes down to the greater homogeneit , culturally and politically, of a state legislator's constituents (potential voters and contributors), compared to that of a federal Representative or Senator; Further, getting say a dozen state legislators “lined up“ so they will be favorable to a given organization’s position on an issue is considerably less daunting than ‘would. be a comparable achievement at the national level. Another factor that it is desirable to remember is that in the United States many legislative areas--such as most criminal law--function at the state level rather than the federal level. Hypothetically, the U.S. is a union or amalgamation of 50 sovereign states, so there are 50 state constitutions, plus the U.S. Constitution, which together make up the legitimacy basis for governmental activities at the various levels. Thus, there are 51 arenas for the working out of the legitimacy (constitutionality) of the basic legal framework for the functions of the society. And the U.S. Supreme Court is continually called upon to establish or modify the boundaries of the pulsating power zones that provide the energy for the functioning of the political system. According to Festinger and. Aronson (1960), when. an individual finds that an opinion advocated by a credible source is discrepant from one’s own opinion, one experiences l4 dissonance. Aronson, Turner and Carlsmith (1962) elaborated four different ways one might take to reduce this dissonance as follows: (1) one could change one’s own opinion to bring it closer to that of the source; (2) change the source's opinion to bring it closer to one’s own opinion; (3) seek support for one’s Opinion by finding other people who hold similar opinions; (4) derogate the source, i.e., make the opinion of the source nonapplicable to one’s own by discounting the ability of the source to have a valuable opinion on the topic. Previous studies (Petty and Cacioppo, 1981; Aronson, Turner and Carlsmith, 1962; Kelman and Hovland, 1953; Hovland and Weiss, 1952; Haiman, 1949; Kulp, 1934) identified a source credibility as a crucial variable which ‘would. maximize dissonance reduction through opinion or attitude change rather than derogation. According to Petty and Cacioppo (1981), a source credibility consists of two components: expertise and trustworthiness. The expertise refers to the extent to which the source is perceived by the audience to know the correct stand on the issue. The trustworthiness refers to the extent to which a source is perceived as being motivated to communicate his/her stand on the issue. For example, in Kelman and Hovland’s (1953) study, a juvenile court judge and a person who had recently been arrested for drug—peddling served as the sources of a message about the treatment of juvenile delinquent. The judge and drug peddler might differ in terms of credibility for a number of reasons. The judge 15 differs from the drug—peddler in terms of knowledge on the topic of treating delinquents (expertise), motive to speak honestly about the topic (trustworthiness) (Liska, 1978). Empirical results of the previous studies (Petty and Cacioppo, 1981; Aronson, Turner and Carlsmith, 1962; Kelman and Hovland, 1953; Hovland and Weiss, 1952; Haiman, 1949; Kulp, 1934) showed that high-credibility sources produce more attitude change than low-credibility sources immediately after a presentation of the message. The United States Supreme Court has been considered as a credible source, that is, in terms of expertise and trustworthiness, mainly because of the Supreme Court's mythic and symbolic prestige in American politics (Marshall, 1987). Lerner (1967) defined the role of the United States Supreme Court as a “republican schoolmaster“ which reflects the idea that the Court may give the populace an example of the way good republicans should behave through its explication of the law and its high.moral standing. According to Lerner (1967), this role was consciously recognized by the justices themselves during the early history' of the Court. In Federalist, no. 78, Hanultxn1 argued.that the judiciary lacked the power to enforce its decrees and that it was therefore the I'weakest branch,“ thus the supreme Court needs public support if its decisions are to be effective (Franklin and Kosaki, 1989). The relationship between Supreme Court decision and public opinion (or attitude change) has. received much 16 attention in the literature. .According to Franklin and.Kosaki (1989) , most studies dealing with public opinion and the Court deal with questions of regime support or the response of the Court to public opinion. In addition, the studies that deal with the question of the response of the public to Supreme Court decisions generally focus on elite groups. One of the most widely accepted explanations of the effect of the Supreme Court rulings on public opinion is the legitimation hypothesis which argues that when the Supreme Court rules, it confers legitimacy on the position it favors. The Supreme Court's prestige allows it to bestow legitimacy on others through its rulings. Thus, leads to increased public support for the position taken by the Supreme Court (Marshall, 1987; Casey, 1974; Adamany, 1973; Murphy and Tanenhaus, 1969; Kessel, 1966) . The legitimation hypothesis is one of the specific mechanisms by which the Supreme Court may obtain greater support for its position. However, regardless of the specific mechanism, most researchers agreed that Supreme Court decision increase the support for the Supreme Court's position (i.e., a “positive response'I hypothesis). Studies of the positive response hypothesis have focused on changes in the aggregate level of public support for the Supreme Court's position (Johnson and Canon, 1984; Uslaner and Weber, 1980; Blake, 1977) . Those studies adopt the positive response hypothesis: they usually compare the level of public support prior to a Supreme Court decision with the level of support 17 after the decision. While, some studies of the Supreme Court’s impact on public opinion produced support for the positive response hypothesis (Johnson and Canon, 1984; Uslaner and Weber, 1980), others failed to produce support for the positive response hypothesis (Marshall, 1987; Adamany, 1973). For example, Marshall’s (1987) study (HS 18 Supreme Court decisions found that the average shift from pre-Supreme Court decision opinion to post-decision was zero. After reviewing the studies of the effect of Supreme Court decision on public opinion, Franklin and Kosaki (1989) argued that increased aggregate support for the Supreme Court’s position is not the best indicator of Supreme Court’s impact on.public opinion. Instead, they suggested that in the wake of Supreme Court decision, some people may respond positively and some negatively; Thus, both groups are affected by the Supreme Court's ruling even if the net effect is no change in aggregate opinion. In order to test the effect of Supreme Court decision on public opinion, Franklin and Kosaki (1989) developed structural response hypothesis which focus on the structure of opinion concerning a Supreme Court ruling. In the other words, they were concerned with which groups support and oppose a position and how intensely. When aggregate support rises (or falls), the Supreme Court decision had an effect. Even if the net effect is zero, while the structure of group support and opposition changes, the Supreme Court decision still had an impact. 18 Murphy and Tanenhaus (1969) pointed out that one problem with the study of public response to the Supreme Court decision is that only a small portion of the public is aware of most Supreme Court decisions, and an even smaller proportion is accurately informed about them. Thus, an examination of Supreme Court impact on public opinion is more meaningful in cases where public attention is high such as desegregation or abortion than in cases where public attention is low such as antitrust or the commerce clause. 4. THEORIES OF ATTITUDE CHANGE: REFERENCE GROUP, CONTEXTUAL, AND PURELY INDIVIDUALISTIC REACTION THEORY According to Franklin and Kosaki (1989), when public decisions are made (by the courts, the president or Congress), the individual's psyche processes the information and comes to some conclusions regarding it. However, the process would not stop here because peOple are not atomistic. Individuals may react to certain event, but they must also consider the reactions of those about them. In this way they test their own interpretations of what has happened and modify their conclusions in light of the communal interpretations. Franklin and Kosaki (1989) argued that when the Supreme Court rules, initial reactions may be entirely individualistic, but within a short period of time the individual is exposed to the reactions of others in the social environment. Thus, these 19 environmental inputs may change or reinforce both the interpretation of the public decision and the individual’s reactions to it. The term, reference group, was first used by Hyman (1942), who elaborated the concept and explored some of its properties in his work, “The Psychology of Status.“ Later, Kelley (1968) distinguished the two functions of reference group, one as standards of comparison for self-appraisal (comparative), another as the source of the individual’s norms, attitudes, and values (normative). The normative reference group, further, could be divided into the positive and negative types which indicates that individuals may form their attitudes in opposition to the norm of a group as well as in accordance with them (Hyman and Singer, 1968). As discussed earlier (see page 14), one of the ways individual might take to reduce a dissonance was to seek support for one's opinion. by finding other’ people ‘who hold similar opinions. Thus, individual’s positive reference group could provide assurance for one’s position and reduce the dissonance. Sprague (1982) stated that the identification. of a systematic association between variation in behavior and variation in environmental properties goes back at least to the publication of Emile Durkheim/s “Suicide“ at the turn of the century and finds its first modern statement in research in political behavior in the writing of Herbert Tingsten 20 (1937). According to Sprague (1982), a large number of studies in contextual analysis have been done at the level of communities or small regions because it was occasionally presumed that survey measurements are individual as opposed to contextual or social. However, some common individual measures (e.g., occupation or church attendance) are obvious proxies for" ea complex individual history, and index theoretically relevant, current social embedding. Thus, some survey measurements which are commonly called as individual properties acquire theoretical power precisely because they are proxy measures for a rich contextual experience (Sprague, 1982). When a stimuli (Supreme Court decision or public officials' publicized decisions) is provided for the individual, the individual pays attention (or not) to this stimuli and engages (or not) in expressive response. The social system: in ‘which the individual is embedded then responds (or not) to the individual’s behavior. However, the stimuli reaches the individual filtered.not only by his or her perceptual screen (the reception bias of the individual) but also by the personally immediate social system within which the individual lives and works. It is assumed that members of micro-environments are positively affected.by the group norms which is likely to be informationally biased (Sprague, 1982). According to Sprague (1982), the natureeof the contextual effect depends upon the homogeneity of opinion within the 21 social environment. If the environment is consensual the individual is likely to encounter a stream of reinforcements consistently favoring the modal position. However, if the social environment is heterogeneous, there will not be consistent reinforcement of a single interpretation of events. This will reduce any contextual effects and.produce a response more dependent on the individual alone. Sprague also argued that between group (or micro-environment) variance also affects the nature of contextual effect. If between-group variance is high, it could be expected that group members would move in opposite directions, leading to greater polarization” However, if between.group variance is low, then more uniform shifts in the population are the likely result. In sum, contextual theory predicts that those who are more solidly embedded in an environment - who have more of their interactions within the group — are most likely to be affected. by the group norms“ Those peripheral to the environment should be much less affected. On the other hand, purely individualistic reaction.theoryflpredicts that increased interaction with others would have much less effect on individual’s attitudes. Recently, Franklin and Kosaki (1989) made use of the National Opinion Research Center’s General Social Survey (GSS) data in order to determine the effect of one of the most highly debated and criticized Supreme Court decisions, Roe vs. Wade. The GSS survey was conducted in February through April 22 of both 1972 and 1973. The Supreme Court's ruling on Roe vs. Wade ‘which established. a *woman's right to terminate an unwanted pregnancy, was handed down on January 21, 1973. Based on the results of their study, Franklin and Kosaki (1989) concluded that even though the aggregate support for the Supreme Court’s position was about same for the pre— and post decision groups, the Supreme Court decision crystallized the issue preferences further and led to greater homogeneity of within-group beliefs. In addition, they asserted that further analysis of data offered strong support for the contextual theory and little support for the purely individualistic reaction theory (Franklin and Kosaki, 1989). lemy'opinion, Franklin and.Kosaki overlooked.one crucial aspect when they tested the contextual theory and the purely individualistic reaction theory using GSS data. It was hypothesized by Franklin and Kosaki that “when the Court rules, initial reactions may be entirely individualistic; but within a short period.of time the individual is exposed to the reactions of others in the social environment. These environmental inputs may change . . . individual’s reactions. . . " As described earlier, the GSS data.were gathered over a period of three months, beginning immediately after the Supreme Court ruling and ending three months after the rulingu ZHowever, the data used by Franklin and Kosaki make no distinction between information gathered at the beginning of the period and that gathered toward the end of the 3—month period. Consequently, 23 such data cannot possibly test a hypothesis concerning individualistic reactions (at the beginning of the time period) and contextually affected reactions (at the end of the time period). The research of this study used two separate but comparable data sets, one gathered immediately after the Supreme Court ruling, and the other three months later. This research design permitted us to compare the adequacy of “contextual theory" and “purely individualistic reaction theory“. CHAPTER III RESEARCH METHODS AND HYPOTHESES 1. RESEARCH DESIGN This study proposed.to focus on.the individual’s attitude change about abortion issues for a three months period after the Supreme Court ruling on Webster vs. Reproductive Health Services in.July 1989. The research was guided.by three major questions. First, is there any effect of a Supreme Court ruling on people's attitudes toward abortion issues? Second, if there is any effect of the Supreme(Court ruling on people’s attitudes toward abortion issues, how does that attitude change occur? Third, is there any race difference in attitudes toward abortion issues? The design of this study could be called quasi- experimental research. One national sample was measured immediately after (Earlier Category) the stimulus (the Supreme Court ruling), an equivalent sample (Later Category) was measured two and a half months after the stimulus. Two different types of research design, experiment and sample survey, are often used to study changes of attitudes. 24 25 The experimental research design is characteristically strong in that a single control group controls for all or most of the threats to internal validity (Campbell and Stanley, 1963). However, experimental research is typically limited to a narrow range of stimuli, to a simple situation and a short time frame (Hovland, 1959). Thus, without being supplemented by descriptive research, experiments cannot explain particular historical occurrence such as the Supreme Court rulings. On the other hand, survey research does not separate out the effects of various variables (Hovland, 1959). The design of this study shares a problem with general survey research and quasi-experimental research design, that is, the lack of a control group. In the context of the present discussion, this point comes down to sorting out the impact of the Supreme Court ruling from the influence of the other factors, such as, difference between people in religious affiliation. This issue of the lack of a genuine control group can be circumvented to some extent by distinguishing between those who had heard of the Supreme Court ruling and those who had not, from both the Earlier and the Later categories. If those who had not heard of the Supreme Court ruling, from both the Earlier and the Later Groups, display about the same attitude toward abortion issues, and there are differences among those who had heard, we could attribute the change to hearing of the Supreme Court ruling. The validity of this procedure is strengthened if statistical control is 26 used to measure and “hold constant" (statistically) the effect of other variables, such as demographic or social-structural variables, which could be expect to be related to (correlated with) the effect (dependent) variables. 2. DATA SOURCE, CONDITIONS, MEASURES AND CATEGORIES A. Data Source and Conditions Data obtained from the Inter-university Consortium for Political and Social Research at Ann Arbor, Michigan was used for this study. CBS News/New York Times (producer of the data) conducted two abortion polls on nationwide samples using direct telephone interviews in 1989. The first nationwide survey on abortion issues was conducted between the 25th and the 28th of July which was three weeks after the Supreme Court decision on Webster v. Reproductive Health Services. The second survey was conducted between the 17th and 20th of September which was two and a half months after the Supreme Court decision. The number of cases of the first survey was 978 and that of the second survey is 1,347. The samples of the CBS/NeW' York Times surveys were selected through. a variation of “stratified random digit dialing“ (see Appendix C for detailed sampling and weighting procedures). There were 77 variables in the first survey and 100 27 variables in the second survey. Twenty three more variables in the second survey included questions like President Bush's performance in specific areas, Bush’s relationship with congress, taxes, and United States aid to Eastern Europe and the Soviet Union. Despite the difference in number of variables between the first and the second survey, crucial information items on abortion issues and respondents' background were exactly the same. The questions on abortion issues which were used in both the first and the second survey included whether respondents approved of the July Supreme Court decision in Webster vs. Reproductive Health Services, whether abortion should be legal and under what circumstances (e.g., possibility of birth defect, unmarried motherhood, health risk to the woman, age of the female, and whether the pregnancy would force a teenager to quit high school or a professional women to interrupt her career), and.whether they favored or opposed possible restrictions on abortion such as requiring parental consent and prohibiting public employees or hospitals from. jperforming abortions. In addition, respondents' opinions were sought concerning government regulation of abortion, the reasonableness/extremism of pro— and anti-abortionists, and whether abortion was murder. Also, the same background questions on respondents such as marital status, parental status, age, income, sex, race, education, religious preference, religiosity, and political party affiliation, were used in both the first and the second 28 surveys (see Appendices A and B for complete content of the data). In sum, data were rich in opinions on abortion issues, demographic and contextual information. Also, the timing of the gathering' of two data sets strengthens their appropriateness for the comparative test of contextual theory and purely individualistic reaction theory. B. Measures Relevant To Abortion Attitudes and Categories As the abortion controversy has developed in the United States, among the factors that appears to have a significant bearing'on most people’s attitudes toward these issues are two factors that we focused on in the analysis. We named the two factors (1) 'Restrictiveness' and (2) “Justifications”. The Supreme Court ruling on Webster vs. Reproductive Health Services had a direct effect on whether the individual states could restrict the right to an abortion. In the data of this study, there were four items which measure the attitudes toward restrictions on abortion by states, i.e., consent of parents for women under 18, a test judging a fetus could live or not outside of womb, prohibiting abortion by public hospitals and employees, and making it difficult for private clinics to perform abortions. These really come down to ways in which the pro-life people try to limit the frequency of abortion either through the wording of legislature or by their actions such as blocking the entrance 29 of abortion clinic. 'Fhis "Restrictiveness“ measure‘was useful in testing the Hypotheses proposed in this research. Because restrict factors referred to three different aspects of restrictiveness, thus, operationalized the general idea of restrictiveness. 131thiS context, “Restrictiveness“ refers to a person's beliefs about those restrictions which should.limit the occurrence or frequency of abortion. Through the years, abortion issues in the United States have specifically elaborated in terms of whether abortion should be legal and under what circumstances. The circumstances‘ that were referred to in the questions of survey used in this research included health risk to the woman, possibility of birth defect, unmarried motherhood, cannot afford more children, teenage girl dropping out of school, and interrupting woman’s occupational career. The above six “circumstances“ have been often used to justify abortion and people of course differ with respect to the acceptability of these reasons. The first two listed above had to do with health risks and the others can be called “discretionary“ factors. Previous researches indicated that these two items, health risk to the woman and possibility of birth defect, are viewed as justifying abortion by high percentages of the general U.S. population, whereas the 1 Another circumstance often mentioned as controversial circumstance is rape, but rape was not included in the above list because different wording was used in the Earlier and the Later surveys. 30 population has split with respect to the other justifications (Wilcox, 1990; Franklin and Kosaki, 1989). Thus, it was reasonable to do an analysis on the items which show variability of response. In this context, "Justifications" refers to a person’s beliefs about those circumstances which should limit the legitimacy of abortion. The respondents in both the “Earlier" and "Later“ categories were divided into two categories; those who had “Heard“ of the Supreme Court ruling and those who had “Not Heard". This categorization was necessary because there was no reason.to believe that the Supreme Court ruling should.have any effect on those who had “Not Heard“ about it. Furthermore, as pointed out above (p. 21), those who had tug; Heard'I groups about the Supreme Court ruling could be viewed as a sort of proxy for a control group if it turns out that there were attitude differences among those who had "Heard“. In our Hypotheses, the “Heard category“ refers to those who had heard of the Supreme Court ruling on Webster vs. Reproductive Health Services, and the ”Not Heard category" refers to those who had not. 31 3 RESEARCH HYPOTHESES As discussed earlier, in general, a Supreme Court ruling on a controversial issues tends to increase the support among the “public“ for the Supreme Court’s position (Marshall. 1987; Johnson and Canon, 1984; Uslaner and Weber, 1980; Caset, 1974; Adamany, 1973; Murphy and Tanenhaus, 1969; Kessel, 1966) . Furthermore, even if the net effect is zero, while the structure of group support and opposition changes, the Supreme Court decision still had an impact (Franklin and Kosaki, 1989) . Also, as discussed earlier, when the Supreme Court rules, initial reactions may be entirely individualistic, but within a short period of time the individual is exposed to the reactions of others in the social environment. Then, these environmental inputs may change or reinforce both the interpretation of the public and the individual’s reactions to it (as emphasized by the contextual theory) (Franklin and Kosaki, 1989; MacKuen and Brown, 1987; Sprague, 1982). One of the most discussed subjects in abortion issues is racial differences in abortion attitudes. And one of the most interesting aspects is that during the last couple of decades, black2 Women were twice as likely to have abortions as whites, 2 In recent years, the term "African—American“ has often been used rather than I'black". As the United States population has become more diverse over the past 50 years, the choice of appropriate terms becomes more difficult. 32 yet, blacks have been less supportive of legal abortion than whites (Henshaw and Silverman, 1988; Secret, 1987; Hall and Ferree, 1986). A couple of studies (Wilcox, 1990; Combs and Welch, 1982), using 17 years of GSS data (1972—1988), demonstrated that racial differences on abortion attitudes are declining. In addition, the introduction of new religious variables in the later surveys suggested that the race differences in abortion attitudes were in part a function of the greater religiosity and doctrinal orthodoxy of the black population” .Although it would be possible to develop and test hypotheses for this study without including the race variable, racial issues are so important in the functioning of this society, that it would be unrealistic to avoid them. Based on the theories and findings of previous studies discussed above, we formulated the following propositions: Proposition 1: A Supreme Court ruling usually affects individual’s attitudes toward abortion issues. Proposition 2: Increased interaction with others leads individual's attitudes closer to the group norms. Proposition 3: In the United States, there are racial differences in abortion attitudes. The Supreme Court ruling on Roe vs. Wade in 1973 appeared to legalize the abortion, but Webster vs. Reproductive Health Services in 1989 put new restrictions on the right to an 33 abortion. A poor, divorced texan named Norma McCorvey, pseudonymously called “Jane Roe” , wanted to end her pregnancy. Even though she had her baby and gave it up for adoption, her law suit was settled belatedly in her favor. This ruling altered the lives of millions of women, and ignited the “civil war' over abortion issues. At that time, Justice Blackmun declared that the constitutional right to privacy was “broad enough to encompass a woman’s decision whether or not to terminate her pregnancy. " Roe vs. Wade divided pregnancy into three trimesters. During the first three months, women were given a right to abortion, protected by the constitutional right to privacy. During the second three .months, an individual state was allowed some say in the control of abortion, but only in such regulations that protected the health of the mother, not the fetus. In the final trimester, as the fetus became viable outside the womb, the state could proscribe abortion unless the mother's life or health was endangered. After the ruling, it was criticized that the Supreme Court established a somewhat “shadowy“ privacy right as to authorize the killing of a fetus without first determining whether the subject in question might not be a human.being (Gillespie, 1989; Anderson, 1989; Salholz, et al., 1989; Carlin, 1989). In retrospect, it appears that the Court was unwillingly to commit itself with regard to the fundamental question about the mbment when the human life begin. 34 Two Missouri pro-life activists, Lee and Andrew Puzder, crafted a bill which stated that "life begins at conception“ and included a number of regulations likely to be accepted by a court majority in the early 19803. Then, pro—life forces in the Missouri Legislature pushed the bill through. After a local abortion clinic challenged the new law, the Supreme Court agreed to hear webster vs. Reproductive Health Services case. On JUly 3, 1989, the Supreme Court handed down the ruling on Webster vs. Reproductive Health Services which required: 1) Missouri doctors to test whether a fetus thought to be 20 weeks or older has reached a stage of development at which it could live outside the womb before considering an abortion; 2) public hospitals or other taxpayer-supported facilities may not be used for performing abortions not necessary to save life, even if no public funds are spent; 3) public employees, including doctors, nurses and other health care providers, may not perform or assist an abortion not necessary to save a woman's life. Even though the ruling had only a small direct effect, most significantly, the majority of justices let stand a preamble to the Missouri act which declares that ”the life of each. human being begins at conception" and.that “unborn children have ... all the rights, privileges and immunities available to other persons, citizens and residents of this state“. By revising the foundations of Roe vs. Wade and allowing states greater latitude in restricting abortion, the Supreme Court apparently opened the 35 door to an avalanche of state legislation that will attempt to test further the right to control abortion (New York Times, 1989; Gillespie, 1989; Anderson, 1989; Salholz, et al., 1989; Carlin, 1989). As discussed earlier, as a result of the Supreme Court ruling, since 1989 the states have had more authority to restrict abortion. In order to respond to the Supreme Court ruling, one must have heard of it. Thus, it is reasonable to assume that those who had "Not Heard” would not change their opinion on the restriction on abortion by the states within a two months period (barring the occurrenceeof other influential events). Formulating Hypothesis 1 might seemed unworthy. However, if Hypothesis 1 is supported, this strengthens the conclusions drawn from. the testing of later Hypotheses concerned with the effects of Supreme Court rulings. The logical relationships between the mean scores on “Restrictiveness“ for the Earlier/Later and Heard/Not Heard characteristics, as expressed in Hypotheses 1 and.2, are shown diagrammatically in Figure l. Hypothesis 1: Between the Earlier and Later categories, those who had not heard of the Supreme Court ruling on Webster vs. Reproductive Health Services, do not discernibly differ on the ”Restrictiveness“ measure. 36 Earlier Later Not Heard Rem -<————> RLN ‘<——- Hypothesis 1: REM ' RLN = 0 Heard R.EH Run <-- (For comparisons between REH and R“, see Hypotheses 4 and 5) Hypothesis 2: REN < REH and RLN < RLH Mean score on Restrictiveness Not Heard categories. Where: REN RLN = Mean score on Restrictiveness Not Heard categories. REH = Mean score on Restrictiveness Heard categories. RLH = Mean score on Restrictiveness Heard categories. in in in in Figure 1: Diagram of Effects on Restrictiveness of Earlier/Later and Heard/Not Heard Categories in Hypotheses 1 and 2. Earlier, Later, Earlier, Later, 37 Previous studies (Franklin and Kosaki, 1989; Combs and Welch, 1982; Uslaner and Weber, 1980) found that persons with less formal education are less likely to have heard about Supreme Court decisions than those with more education (number of years in school). Since education, along with other variables such as race and gender, is correlated with attitudes toward abortion issues, we need to control attitudes toward abortion by demographic and socio-economic variables in order to eliminate “spurious“ influence. The independent variables included here are education, religion, gender, age, race and religiosity. Hypothesis 2: When controlled for demographic and socio—economic variables, the “Heard" category has a higher mean restrictiveness score than the “Not Heard“ category in both the Earlier and the Later categories. It has been suggested that the Supreme Court may give the populace an example of the way good citizens should behave through its explication of the law and its high mpg; standing (Lerner, 1967) . It has been also widely accepted that when the Supreme Court rules, it confers legitimacy on the position it favors (Marshall, 1987; Casey, 1974; Adamany, 1973; Murphy and Tanenhaus, 1969; Kessel, 1966) . As discussed earlier, the majority of the United States Supreme Court Justices let stand a preamble to the Missouri act that sets out that "life begins 38 at conception“. Based on the positive response hypothesis that the Supreme Court ruling increase the support for the Supreme Court's position, we would expect that people who had Heard of the Supreme Court's position on when a human life begins, would agree more with the statement that abortion is the same thing as murdering a child. Hypothesis 3: In both the Earlier and the Later categories, when controlled for demographic and socio—economic variables, the heard category is more likely to agree with the item which states that abortion is same thing as murdering a child, than does the not—heard category. The contextual theory, as stated by Sprague (1982), posits that homogeneity in the social environment is a key variable in attitude change. Franklin and Kosaki (1987) applied this principle in their research on responses to Supreme Court decisions. The theory also predicts that those who are more solidly embedded in an environment are most likely to be affected.by the group norms. In accord with this line of thought, it is reasonable to expect that highly religious persons would express attitudes that conformed to the "official“ views of their religious organization regarding abortion to a greater degree than would “less religious" persons. Based on the principles of the contextual theory, we would expect that measures of "Restrictiveness“ and “Heard“ "Highly Religious“ Earlier Category Hypothesis 4: High Variance on Restrictiveness Measure Hypothesis 5: High Variance on Justifications Measure (Heterogenous) 39 ———> --—> “Heard“ “Highly Religious" Later Category Hypothesis 4: Low Variance on Restrictiveness Measure Hypothesis 5: Low Variance on Justifications Measure (Homogeneous) Predicted Variance of Measures of Restrictiveness and Justifications Among Highly Religious People Who Had Heard of the Supreme Court Ruling on Webster vs. Reproductive Health Services in the Earlier and the Later Category. Figure 2: 4O "Heard" Later Category Hypothesis 6: High Variance on Restrictiveness Measure Less Religious People: Hypothesis 7: High Variance on Justifications Measure (Heterogeneous) Hypothesis 6: Low Variance on Restrictiveness Measure Highly Religious People: Hypothesis 7: Low Variance on Justifications Measure (Homogeneous) Figure 3: Predicted Variance of Measures of Restrictiveness and Justifications Among Less and Highly Religious People Who Had Heard of the Supreme Court Ruling on Webster vs. Reproductive Health Services in the Later Category. 41 "Justifications" varies more among highly religious people who had “Heard" in the Earlier category than among highly religious people who had “Heard“ in the Later category. This proposition is presented diagrammatically in Figure 2. Also, we would expect that a.measure of “Restrictiveness“ and “Justifications“ varies more among less religious people who had Heard“ than among highly religious people who had “Heard“ in the Later category because less religious people‘would.have less chance to be exposed to the conforming pressure of group norms. This is presented diagrammatically in Figure 3. Hypothesis 4: Among those who had “heard", and.were religious, the Variance on the Restrictiveness measure is greater in the Earlier Category than in the Later Category. Hypothesis 5: Among those who had “heard", and.were religious, the Variance on the Justifications measure is greater in the Earlier Category than in the Later Category. Hypothesis 6: Among those who had “heard", and.were secular, the Variance on the Restrictiveness measure is greater than among those who had “heard“, and were religious, in the Later Category. Hypothesis 7: Among those who had "heard“, and were ' secular, the Variance on the Justifications measure is greater than among those who had “heard", and were religious, in the Later Category. 42 According to Ransford and Miller (1983), black men are more traditional in gender-role attitudes than white men, but there is no significant differences among women. Also, Petchesky (1984) suggested the possibility of gender-race interaction on abortion issues. She argued that black men may be less supportive of legal abortion than white men for political reasons, i.e., genocide, while black women may be more supportive than white women because of the greater frequency of abortion in the black community. Hypothesis 8: Black men are less supportive of abortion than white men, and black women are more supportive than white women. 4. STATISTICAL METHODS AND MEASURES As discussed earlier, in order to asses the effect of the Supreme Court ruling, "Restrictiveness“ and “Justifications" were measured by a three and four-item indices (see pages 28 and. 29), respectively; Each. respondent got a score on ”Restrictiveness“ and "Justifications" consisting of the number of items he/she agreed with. The scores of the “Restrictiveness" and ”Justification" measurements were used as the dependent variables. In both the “Earlier" and "Later“ categories, the respondents were divided into two categories based on whether a sample had heard about the Supreme Court ruling or not. 43 Those who had heard of the Supreme Court ruling were categorized as “Heard", and those who had not heard were categorized as"Not Heard“. These categories, the “Earlier"/ "Later", and the “Heard"/“Not Heard", were used as the independent variables. Such demographic and socio-economic variables as gender, education, religion, race, age and religiosity was used.as the control variables. Income of each respondent was not used as a control variable in this study because income was relatively highly correlated with education (r=.39) compare to the other zero-order correlations between those pairs of variables, all of which were lower than .15. The multiple regression analysis, which estimates the relationship between a continuous dependent variable and multiple continuous (n: discrete independent variables, was used to test Hypotheses 2 and 3. The multiple regression analysis was appropriate for testing Hypotheses 2 and 3, because those Hypotheses were stated to find out the relationship between a dependent variable and multiple independent variables. The T-test was used to examine the variance differences among various groups as proposed in Hypotheses 4 tx3'7. The Means test was also used to test Hypotheses 1 and 8. Descriptive statistics such as percentages, variances and correlations, also were used to describe the characteristics of the sample (Knoke and Bohrnstedt, 1991; Schroeder, Sjoquist and. Stephan, 1986; Achen, 1982; Lewis-Beck, 1980). CHAPTER IV DATA ANALYSIS AND HYPOTHESES TESTING In order to test the eight Hypotheses, the “Heard“ and “Not Heard“ categories, and the “Earlier" and “Laterf categories were used as independent variables. To test the Hypotheses, the following variables were either calculated or selected from the data and used as dependent variables: (1) the restrictiveness on abortion, (2) the justification on abortion, and (3) abortion is same thing as:murdering a child. Following variables were also used as control variable in order to test Hypotheses: (1) education, (2) gender, (3) race, (4) age, (5) religiosity, and (6) religious identification. Multiple regression analysis, the T-test, and the Means test were used to analyze the data. The F and t statistics were used to test the statistical significance of the results of the data. Hypothesis 1: Between the Earlier and Later categories, those who had not heard of the Supreme Court ruling on Webster vs. Reproductive Health Services, do not discernibly differ on the “Restrictiveness“ measure. 44 45 Hypothesis 1 was formulated to examine whether any change on the “Restrictiveness“ measure occurred.among people who had not heard of the Supreme Court decision on Webster vs. Reproductive Health Services within a two months period (i.e. , the time interval between the ”Earlier" and the “Later" measurements). As discussed earlier (see page 34), as a result of the Supreme Court ruling, states since 1989 have more authority to restrict abortion. Thus, we hypothesized that those who had.not heard.would.not change their opinion on the restrictions on abortion by the states within the two months period. “Restrictiveness“ on abortion was measured by a three- item-restrictiveness-index for each respondent in which a score of “0" indicates not restrictive and "3" indicates highly restrictive on abortion. The respondents of both "Earlier'I and “Later“ categories were divided into a set of categories: 1) those who had heard of the Supreme Court decision on Webster vs. Reproductive Health Services, and 2) those who had not heard of this decision. Then, the mega score on "Restrictiveness“ was calculated for each of the following categories: (1) those who had.'Heard' of the Supreme Court decision in the “Earlier“ category, (2) those who had “Heard“ in the "Later“ category, (3) those‘who had “Not Heard“ in the “Earlier” category, and (4) those who had “Not Heard“ in the “Later" category. In order to test Hypothesis 1, the “Restrictiveness" score was used as the dependent variable, 46 and the ”Earlier" and the “Later“ categories were used as the independent variable. Also, the “Heard" and "Not Heard" categories were used as a control variable» The mean test was used to analyze the data and the t statistic was calculated to test the significance of the result. The results of the analysis are presented in Table 1. Hypothesis 1 refers only to the ”Not Heard'I condition in Table 1 (the first line of the Table) As Table 1 shows, when cases where the respondents had not heard of the Supreme Court decision were considered, the difference on the mean “Restrictiveness" score between the "Earlier“ and the "Later“ categories was .032 (the first line of Table 1) and increased by 3% from.the “Earlier“ category to the “Later“ category. The t score of the I'Not Heard" category was .598 (p=.55). When alpha (i.e., probability level) is .05 with 1,192 degrees of freedom, the critical value is 1.960 for a two tailed test. Since the test statistic was .598, we could not reject the null hypothesis. Thus, based on the data, we conclude that the “Restrictiveness“ score in.the "Not Heard“ category was not different between the “Earlier“ and the “Later“ conditions. Thus, the data support Hypothesis 1, “Between the Earlier and Later categories, those who had not heard of the Supreme Court ruling on Webster vs. Reproductive Health Services, do not discernibly differ on the “Restrictiveness“ measure.“ As a consequence, this strengthens the conclusions drawn from the testing of later 47 TABLE 1 Mean Score on Restrictiveness, by Earlier/Later and Heard/Not Heard Conditions Earlier Later Diff- %Diff- t p Sub- Category Category erence erence Group (N / %) (N / %) Not Heard: 1.063 1.095 .032 + 3% .598 .55 (429/56) (765/70) Heard: 1.414 1.580 .166 +12% 2.088 .04 (336/44) (326/30) Total: 765 1091 48 Hypotheses concerned with the effects of Supreme Court decisions. As the second line of Table 1 shows, when the cases which people who had heard are considered, a difference of the "Restrictiveness“ score between the “Earlier“ and “Later" categories was .166. The “Restrictiveness“ score increased by 12% from the "Earlier" category to the “Later“ category. We also note that the mean t score of the "Heard“ category was 2.088 (p=.04), contrary to that of “Not Heard" category, larger than the critical value of 1.960 when alpha is .05 with 662 degrees of freedom. The data show that the difference on the “Restrictiveness" score Ibetween. the “Earlier" and “Later“ categories among people who had heard was statistically significant» This is in accord. with. our expectations. However, this line of analysis cannot show a true effect of hearing of the Supreme Court decision because of possible spurious influences by the other factors such as education, gender, religiosity and race. In order to test the true effect of having heard of the Supreme Court decision, we need more elaborate statistical analysis which.can isolate the effect of having heard of the Supreme Court decision from.the effects of the other independent variables on the dependent variable. This line of analysis were performed for the testing of Hypotheses 2 and 3 as follows. 49 Hypothesis 2: When controlled for demographic and socio—economic variables, the "Heard" category has a higher mean restrictiveness score than the “Not Heard" category in both the Earlier and the Later categories. Hypothesis 2 was formulated to examine whether the Supreme Court decision on Webster vs. Reproductive Health Services had any effect on people's attitude toward restriction on abortion by states. As discussed earlier, a number of factors could contribute to people's attitude toward abortion and their attitude change. For example, white Catholics are more likely to oppose abortion than white protestants (Franklin and Kosaki, 1989), black men are more likely to oppose abortion than their white counterparts (Wilcox, 1990) and attitude stability increases with age (Alwin and Krosnick, 1991). Thus, in order to measure an effect of having heard of the Supreme Court decision more clearly, we need to control the influences of other relevant variables on the dependent variable and as well as on the “Heard“ and the “Not Heard“ category. Based on previous literature (Alwin and Krosnick, 1991; Wilcox, 1990; Franklin and Kosaki, 1989), we identified a number of independent variables which may be presumed to have an influence on people's attitude as follows: education, gender, race, age, religiosity and religion. Education was coded into four categories: (1) less than high-school, (2) high-school graduate, (3) some college (trade or business), and (4) 50 college graduate and beyond. Race was divided into white and non-white categories. Religiosity' was divided into two categories, “religious“ and “secular", based on the responses to a question which asked “How important is religion in your daily life?“. Respondents who answered either "extremely'I or ”very“ were coded as “religious“ and respondents who answered either "somewhat'I or “not at all“ were coded as “secular“. Religion was divided into two categories, “Protestant" and "Catholic“. The other religious identifications such as Jewish, "other“ religion, and.'no religion“, were not included in the analysis because there were very few cases with these religious identifications. A multiple regression analysis used to test Hypothesis 2 . The respondents’ “Restrictiveness“ score was used as the dependent variable, and “Heard“ - "Not Heard“ was used as the independent variables. Age, education, gender, religiosity, race, religion, were used as the control variables. In order to determine the effect of having heard of the Supreme Court decision on people's attitude toward restriction on abortion while controlling for (or taking into account) other independent variables, the “Heard“ and the “Not Heard" categories were entered into the multiple regression equation at the last. Correlation coefficients among dependent and independent variables were calculated in order to find out whether high correlations (e.g., .80 or higher) were observed among the independent variables. Because high correlations 51 between independent variables inflate the variances of the estimates, thus, increase a significance of overall regression while ngg§_of the individual coefficients are significant. R2 (coefficient of determination) change when "Heard“ and "Not Heard“ category variable was entered into an equation that already contains the other independent variables, were calculated in order to determine the relative importance of the “Heard“ and the “Not Heard“ category variable. B (multiple regression coefficient), EV (Beta weight), part and partial correlation coefficients were also calculated in order to determine the amount of increase or decrease in the "Restrictiveness” score due to the heard - not heard condition while controlling for the other independent variables (Knoke and Bohrnstedt, 1991; Norusis, 1990; Schroeder, Sjoquist, and Stephan, 1986; Berry and Feldman, 1985; Achen, 1982; Lewis- Beck, 1980; Henkel, 1976). As Table 2 (page 52) shows, correlation coefficients among the independent variables in the ”Earlier“ category were very lowu The highest correlation coefficient of .323 between education and the “Heard" and the “Not Heard“ category (the third line of the last column of the Table) is considerably lower than .80. A value of .80 is generally considered to indicate a multicollinearity among variables, and as a result, significance of the regression coefficients becomes meaningless. Based on the low correlation coefficients among independent variables, we conclude that any significance of .coflmflaom .oomm .uopcow .mufimoflmflaom .Gofiumospm .om¢ u Houucou .pumom boz\pumom u ucmpsoqoch .mmmcm>fluofluummm n unopcomoo .mmm n z uouoz ooo.a .nnmmm noz\6ummm Ame smo. ooo.H "coflmfiamm lee owe.) oea.u ooo.H "comm Ame mmo.- mHo.- mac. ooo.a "smegma Ame 040.- smH.- mad. mmo. ooo.H ”snflmoflmflamm lee mmm. omo. omo.- mmo.- wmo.- ooo.H "coflomosum Ame Hmo.- SHH.- moo.- mofi. boa. mmfl.- ooo.a "was Ame FAA. meo.- mas. mmo.- HNH.- was. mmo. ooo.s "mmmcm>fl -uofluummm lav Amv Abv Amy Amy Avv Amy Ame AHV manmfluo> huoomumu Hwflaumm "gunner boz. pnm .ouwom. one .cofimflaom .ooom .uopcow .huemoflmflamm .COHumosom .mmd .mmmcm>fiuofluummm mcofio mcofiumHmuuou N mam49 mm 53 TABLE 3 Results of Regression of Restrictiveness Index on Age, Education, Religiosity, Gender, Race, Religion, and “Heard“ and "Not Heard": Earlier Category Independent Variables B BETA Part Partial t p Age: .003 .062 .060 .062 1.586 .11 Education: .075 .077 .072 .074 1.898 .06 Religiosity: -.305 —.130 -.128 —.130 3.366 <.01 Gender: -.027 -.013 . -.013 —.014 .359 .71 Race: .132 .039 .038 .039 1.015 .31 Religion: -.140 —.O63 —.O61 —.063 1.621 .11 neural net Heard: .307 .155 .146 .149 3.843 <.01 R Square = .059 R Square Change = .021 (when Sub—Group variable was entered) F Change = 14.770 (p < .01) Note: N = 659. 54 the test result of regression analysis of Hypothesis 2 would be meaningful enough to draw a conclusion. As Table 3 shows, R2 (coefficient of determination) change when ”Heard“ - I'Not Heard“ variable was entered into the regression equation, was .021 and F value of that change was 14.770 (bottom of the Table). when alpha is .05 with 8 (the number of variables entered in the equation) and 651 (N — total number of independent variables in the equation — 1) degrees of freedom, the critical value is 1.88 for an one- tailed test. Since the test statistic was 14.770 (p<.01), we could reject the null hypothesis. The significant change of R2 indicate that the “Heard" and the "Net Heard“ category variable provides unique information about the dependent variable (i.e., restrictiveness score) that is not available from.the other independent variables (i.e., control variables in the analytical sense) in the equation. The multiple regression coefficients (B) of independent variables (second column of the Table) show us the amount (i.e., size of a value) of increase or decrease (i.e., sign.of a value) in the “restrictiveness" score for‘ a one-unit difference in the independent variable while controlling for the other independent variables in the equation. The "B'I of the “Heard" and the 'Not Heard“ category was .307 (the seventh line of Table) which was larger than the other 'Bs“. Based on this result, we tentatively conclude that people among the “Heard" category had .higher scores on the 55 “Restrictiveness“ measure than people among the “Not Heard“ category in the “Earlier" category. However, it is inappropriate to interpret the Bs as sole indicator of the relative importance of variables because the actual magnitude of the coefficients depends on the units in. which the variables are measured. When independent variables differ in units of measurement, the magnitude of their coefficients cannot reveal the relative importancerof the variables. Thus, we need to calculate Beta weights which are the coefficients of the independent variables when all variables are expressed in standardized (Z-score) form. As the third column of the Table 3 shows, Beta weight (.155) of “Heard“ and Not Heard" category (the seventh line of Table) was still larger than the others. Calculating a part and partial correlation coefficient are other ways of showing the relative importance of an independent variable while controlling for other independent variables which are already entered into a regression equation. A part correlation coefficient is a signed square root of change in R2 when a new independent variable is entered into an equation” Thejpart correlation coefficient is a measurement of a proportion of variance in a dependent variable that an independent variable "X" can explain while linear effects of the other independent variables are removed from the independent variable "X“. The partial correlation coefficient is more meaningful than the part correlation 56 coefficient in a sense, because the .partial correlation coefficient can measure a relationship between a dependent variable and an independent variable "X“ while linear effects of the other independent variables are removed from pgph the dependent variable and the independent variable "X". We note that both the part (.146) and the partial correlation coefficients (.149) (the seventh line of the forth and fifth column of Table 3) of the ”Heard" — “Not Heard“ variable are larger than those of the other independent variables. As the bottom of the sixth column of Table 3 shows (the seventh line of Table), the t test statistic for B, Beta weight, part and partial correlation coefficients of the “Heard“ and the "Not Heard'I categories was 3.843 (the seventh line of Table). Since the test statistic of 3.843 was larger than the critical value of 1.645 when alpha is .05 with 657 degrees of freedom, we conclude that the “Restrictiveness" score of the “Heard“ category was significantly larger than that of the “Not Heard“ category in the “Earlier“ condition. The data ShOW’ that the first part of Hypothesis. 2 ‘was supported, that is 'when.controlled.for demographic and.socio- economic variables, the I'Heard" category has a higher mean restrictiveness score than the Not Heard category in the Earlier category“. As Table 4 shows, correlation coefficients among the independent variables in the “Later" category were, again, very low. The highest correlation coefficient of .297 between .coflmflaom .momm .Hopcmw .huflmoflmfiamm .cofiumospm .mm4 u HOHDGOU .puomm uoz\pummm n unopcomoch .mmoco>fiuofluummm u ucmpcoaom .mwm N Z "ODOZ ooo.a "spawn noz\eummm lav oso.- ooo.s "coflmflamm lee mso.- esfi.- ooo.H “meme lee seo.- mmo.- Hmo. ooo.s "sesame Ame oHo.- mma.- was. was. ooo.H "Anamofimfiamm lee ham. Hes. mmo.- mmo.- mso.- ooo.a "coflumospm Ame Ado. mmo.- mmo.- mso. ems. Hao.- ooo.H "was Ame mmm. mmo.- omo. mao.- msH.- mam. Hmo. ooo.a "mmmcm>fl uuofluommm lav Amy Abe Ame Amv Ave Ame Ame AHV manmfiup> xuommumu Hmong "spumom boz. one sensor. use .cofimfiamm .momm .umpcmw .wuflmofloflaom .coflumospm .mmm .mmmcm>fiuofluummm ocean mcoflumHmuuou v mqmde um 58 TABLE 5 Results of Regression of Restrictiveness Index on Age, Education, Religiosity, Gender, Race, Religion, and "Heard“ and “Not Heard": Later Category Independent Variables B BETA Part Partial t p Age: .003 .060 .059 .063 1.928 .06 Education: .152 .164 .156 .163 5.071 <.01 Religiosity: -.415 -.181 -.178 —.185 5.802 <.01 Gender: .023 .012 .012 .013 .389 .71 Race: .168 .053 .052 .055 1.695 .09 Religion: -.090 —.043 -.042 —.044 1.370 .17 Rental Not Heard: .378 .180 .171 .179 5.586 <.01 R Square = .113 R Square Change = .029 (when Sub-Group variable was entered) F Change = 31.204 (p < .01) Note: N = 949. 59 education and the "Heard“ - "Not Heard" category (the third line of Table) was still low enough not to be concerned about multicollinearity among the independent variables. As Table 5 shows, R2 change when the "Heard" - "Not Heart“ category variable was entered into the regreSSion equation (the bottom part of the Table), was .029 and F value of that change was 31.204 which.was significantly larger than the critical value of 1.88 when alpha is .05 with 8 and 941 degrees of freedom, .As the seventh line of Table 5 shows, the multiple correlation coefficient of the "Heard" - “Not Heard" category was .378, which was exceeded in magnitude only by the religiosity variable. The Beta weight (.180), the part correlation coefficient (.171), and the partial correlation coefficient (.179) of the "Heard“ - “Not Heard“ category (the seventh line of Table) variable also*were significantly larger than those of the other independent variables. The test statistic (t=5.586) confirmed the significant increase in the "Restrictiveness“ score'fromlthe “Not Heard! to the "Heard“ in the "Later'I category. When alpha is .05 with 947 degrees of freedom, the critical value is 1.88 for an one-tailed test. The test statistic also confirmed the importance of the "Heard'I - “Not Heard'I category variable in predicting the "Restrictiveness" measure in the “Later" category. Based on the results of two multiple regression analysis, one for the “Earlier" category and the other for the “Later“ category, we conclude that people who had heard the Supreme 60 Court decision scored significantly higher on the "Restrictiveness“ measure than people who had not heard regardless of the time of measurement. Thus, Hypothesis 2, “when controlled for demographic and socio-economic variables, the "Heard“ has higher mean “restrictiveness" score than the “Not Heard“ category in both the Earlier and the Later categories“ was supported by the data. Hypothesis 3: In both the Earlier and the Later categories, when controlled for demographic and socio-economic variables, the heard category is more likely to agree with the itemwwhich states that abortion is same thing as murdering a child, than does the not—heard category. Hypothesis 3 was formulated to examine the influence of the Supreme Court on public opinion. The same control variables (i.e., age, education, religiosity, gender, race, and religion), which were used to test Hypothesis 2 were, again,. used in the testing of Hypothesis 3. The “Heard“ - “Not Heard“ category was used as the independent variable. The responses to the question which states that “abortion is the same thing as murdering a child, or abortion.is not murder because the fetus really is not a child yet?“, were used as the dependent variable. A. multiple regression analysis technique was again 61 utilized tx> test Hypothesis 3. After time “Heard“ - “Not Heard" category 'variable ‘was entered, correlation coefficients among the variables, R2 change, multiple correlation coefficients, Beta. weights, part and. partial correlation coefficients were calculated in order to examine any significant difference between the "Heard" and.“Not Heard" categories on opinion of "abortion is murder" item while controlling for the other independent variables. As Table 6 shows, any correlation coefficient among independent variables was not large enough to indicate multicollinearity (If the independent variables. As the bottom part of Table 7 shows, R2 change was very small (.007) , and F test statistic of R? change was 6.400 (p=.01). When alpha is .05 with 8 and 714 degrees of freedom, the critical value is 1.88 for an one-tailed test. Since the test statistic was 6.400, we conclude that the contribution of the “Heard" -— "Not Heard" category variable to the regression equation was small, but statistically significant. The small values of the multiple correlation coefficient (—1.093), Beta weight (-.092), part (—.087) and partial (-.094) correlation coefficients of the "Heard" - “Not Head“ category variable (the seventh line of‘Table 7) suggested.that having heard of the Supreme Court decision had only a small effect on people’s opinion on the “abortion is murder“ item. Since the test statistic of 2.559 (p=.01) was larger than the critical value of 1.645 when alpha is .05 with 720 degrees of .coflmflaom .oomm .uopcmw .MDAmoflmfiHmm .cofiumoSpm .mm« H Houucou .pummm uoz\pummm u homecomoch .uoousfi no cofluuonm u ucmpcwawn .mmb u z umuoz oooé "Unmmm uOZ\UHmmm Amy ooo. ooo.H "coflmoamm loo omo.- ooo.- ooo.H “meme loo ooo.- oHo.- moo. ooo.H "smegma Imo ooo.- oma.- oHH. moo. ooo.o "mnemoflmflamm loo mom. omo. ooo.- omo.- omo.- ooo.H "coflumospm loo Hmo.- oHH.- ooo.- moH. ooH. mmH.- ooo.H "mma Ame HoH.- moo. oHo. moH. omm. moo.- moo. ooo.H "omens: mfi acouuona loo Amy Abv Ame Ame Rev Ame Amy AHV manmflum> muoaoumu mmfiaumm ".pnmmm uoz. use .pumwm. use .coflmflamm .womm .umpcoo .>uflmoflmflaom .cofluoospm .mo¢ .guopusz we s0ouuond. mcofim mcoHumHouuou m mamfifi mm 63 TABLE 7 Results of Regression of “Abortion is murder“ on Age, Education, Religiosity, Gender, Race, Religion, and “Heard" and "Not Heard“: Earlier Category Independent Variables B BETA Part Partial t p Age: .001 .042 .041 .045 1.207 .23 Education: —.058 -.117 —.109 —.117 3.173 .01 Religiosity: .395 .332 .326 .334 9.469 .01 Gender: .066 .064 .063 .069 1.856 .64 Race: -.065 -.038 -.O37 -.040 1.094 .27 Religion: .071 .063 .061 .066 1.784 .07 Heard/ Net Heard: -.093 -.092 -.087 -.094 2.529 .01 R Square = .155 R Square Change = .007 (when Sub-Group variable was entered) F Change = 6.40024 (p = .01) Note: N = 722. 64 freedom, the difference between the opinions of the "Heard“ and the “Not Heard" category on the "abortion is murder" item was statistically significant. In.addition, negative signs of those values of the “Heard" and the "Not Heard" category variable, suggested that the people who had “Heard" of the Supreme Court decision agreed little less than the people who had “Not Heard“, on the “abortion is murder“ item. Based on the results of the regression analysis, we conclude that in the “Earlier“ category, when controlled for demographic and socio-economic variables, the “Heard“ agrees less than does the “Not Heard“ category on the item which states that abortion is same thing as murdering a child. The correlation.coefficients among variables of‘Table 8, again, confirmed. weak correlations among' the independent variables, thus, proved.the reliability of the test regression model. As the bottom part of Table 9 shows, R? change when the “Heard” — “Not Heard'I category was entered into the regression equation, was very small (.0063). The F test statistic (7.759 and p=.01) of R2 change was significant since the critical value is 1.88 when alpha is .05 with 8 and 1,015 degrees of freedom. As a result, we conclude that the contribution of the “Heard" - “Not Heard“ category to the regression equation was small but statistically significant. Again, small values of multiple correlation coefficient (-.093), Beta weight (-.084), part (-.080) and partial (—.O87) correlation coefficients (the seventh line of Table 9) .cofimflamm .oomm .umpcww .xuflwoflmfiaom .cofluoospm .mm< u Houucou .ouomm ooz\pummm u ucmpcmaoch .uopusE we cofiuuond u ucoocoamm .mmo.o u z "muoz ‘|'|"'|""[8"I|||l"l||l“--|"'llll'l'|"l--'|'Ilil'l'Il""lll'|l"'l'l"'lI. ooo.H ”some noz\pummm loo ooo.- ooo.H "aoflmaomm loo ooo.- ooH.- ooo.o "mumm loo ooo.- mmo.- Hmo. ooo.H "umpamo loo moo.) omH.- ooo. was. ooo.H soflmoflmoamm loo pom. Hoo. moo.- mmo.- ooo.- ooo.H "aconmospm lmo ooo. mmo.- ooo.- moo. poo. Hoo.- ooo.H "was Imo moo.- moo. mmo. ooH. mom. mom.- mHo. ooo.H "omens: mo acouuonm Loo Ame C; 3v Am. :1 Amv Amy A: wanofiuo> xnoomumu Hmong "gpumom uozg poo .pumom. pom .sofimflaom .womm .umpcow .huomOomHHom .QOADmospm .mm< ..nmpusz me c0wuuoem. mcofio meoumHmHuoo w mqmfiB mm 66 TABLE 9 Results of Regression of “Abortion is murder" on Age, Education, Religiosity, Gender, Race, Religion, and "Heard" and "Not Heard": Later Category Independent Variables B BETA Part Partial t p Age: -.003 -.022 —.022 —.023 .749 .45 Education: -.O80 —.l65 -.156 -.166 5.391 <.00 Religiosity: .361 .303 .297 .306 10.246 <.01 Gender: .059 .059 .058 .063 2.013 .04 Race: ~.003 -.002 -.001 —.002 .071 .94 Religion: .083 .076 .075 .081 2.598 <.01 Heard! th Heard: -.O93 -.084 -.080 -.087 2.786 <.01 R Square = .383 R Square Change = .006 (when Sub-Group variable was entered) F Change = 7.759 (p < .01) Note: N = 1,023. 67 indicated that there were small difference between the "Heard" and the ”Not Heard" category on the opinion of “abortion is murder“ item. The test statistic of t (2.786 and p=.01) for those coefficients was statistically significant since the test statistic was barely larger than the critical value of 1.645 when alpha is .05 with 1,022 degrees of freedom. The negative sign of coefficients indicated that people who had "Heard“ of the Supreme Court decision agreed little less than people who had “Not Heard“ on “abortion.is murder“ item in the ”Later" category. In sum, based on the results of the data analysis, Hypothesis 3, ”in both the Earlier and the Later categories, when controlled for demographic and socio-economic variables, the heard category is more likely to agree with the itemtwhich state that abortion is the same thing as murdering a child, than does the not-heard category“ was not confirmed by the data. Hypothesis 4: Ambng those who had “Heard", and were religious, the Variance on the Restrictiveness measure is greater in the Earlier Category than in the Later Category. Hypothesis 5: Among those who had 'Heard', and.were religious, the Variance on the Justifications measure is greater in the Earlier Category than in the Later Category. 68 Hypothesis 6: Among those who had “Heard“, and were secular, the Variance on the Restrictiveness measure is greater than among those who had “Heard", and were religious, in the Later Category. Hypothesis 7: Among those who had “Heard", and.were secular, the Variance on the Justifications measure is greater than among those who had I'Heard", and were religious, in the Later Category. Hypotheses 4 to 7 were formulated to examine the essence of the contextual effects which states that homogeneity in the social environment and strong membership are key variables in attitude change (see Figures 1 and.2 of pages 38 and.39). The religiosity of the respondents was used to divided the sample into the two groups: (1) people who are more solidly embedded in an homogenous environment, and (2) people who are not. Respondents who answered either "extremely“ or “very" to a question “How important is religion in.your daily life?“ were put into the "religious“ group, and respondents who answered either I'somewhat'I or “not at all“ were put into the “secular“ group. Then, in view of the importance of religion-related factors in abortion issues, the respondents were divided into the religious and the secular Protestants and Catholics. The I'Restrictiveness" and the "Justification“ measures were used as dependent variables. The religiosity, religious identification, the "Heard"- “Not Heard“ category, and the 69 “Earlier” and the ”Later“ categories were used as independent variables. The T—test was used to examine the variance differences among various groups as proposed in Hypotheses 4 to 7. The mean scores of the “Restrictiveness" and the “Justification“ measures, the standard deviations, and the variances within each group were calculated. The F value (the ratio of the larger sample variance to the smaller) was calculated.in(order to test whether the population variances in the two groups were statistically different in magnitude. If the observed significance level for the F test was small (less than .05), the null hypothesis that the group variances were equal was rejected. When the population variances in the two groups were not equal, the separate-variance t-test was used to determine the statistical significance of the difference between two group means. The degrees of freedom for the separate t-test is a function of the sample size in the two groups. If the observed significance level for the F test was large (i.e., the population variances in the two groups are equal), the pooled-variance t-test was used. The pooled- variance t-test is based.on.the assumption that the population variances in the two groups are equal and is obtained using a pooled estimate of that common variance. The degrees of freedom for the pooled t-test are obtained as the sum of the sample sizes in both groups minus 2. In the T—test analysis, the observed significance level (p) for the F and ts are used 70 to determine whether the differences of the variances and the means are significant (usually, less than .05 is considered as significant) (Norusis, 1990; Henkel, 1976). In order to test Hypothesis 4, “Among those who had I'Heard", and. were religious, the variance on the Restrictiveness measure is greater in the Earlier category than in the Later category“, the respondents were grouped into four categories: (1) religious Protestants who had “Heard" of the Supreme Court decision in the I'Earlier" category, (2) religious Protestants who had "Heard“ of the Supreme Court decision in the “Later“ category, (3) religious Catholics who had “Heard" of the Supreme Court decision in the “Earlier“ category, (4) religious Catholics who had ”Heard" of the Supreme Court decision in the “Later“ category. Then, the mean “Restrictiveness“ scores, the standard deviations, and the variances within each group were calculated. The F and the t test statistics and the observed significant levels between groups (1) and (2), and (3) and (4), were calculated. As Table 10 shows, the difference of within group variances between the ”Earlier“ and.the "Later“ categories was .127 or 14% increase from the “Earlier“ category to the "Later” category (middle of the fifth column). The test statistic (F=l.14) did not confirm the significant difference of the within group variances between the “Earlier“ and the l'Later" categories, since the calculated probability level (.49) did not attain the .05 probability criterion. 71 TABLE 10 Restrictiveness Score by Category Among Religious Protestant Who had “Heard" Mean Number Restrict- of tiveness Standard Category Cases Score Deviation Variance F p Earlier: 115 1.174 .948 .899 Difference: .124 .065 .127 1.14 .49 %Difference: +11% + 7% +14% Later : 114 1.298 1.013 1.026 72 As the third column of the Table 10 shows, the difference of mean “Restrictiveness“ scores between two categories was .124 or the mean "Restrictiveness“ score in the “Later" category was 11% greater than the “Earlier" category. As mentioned earlier, the observed significance level for the F test was significantly larger than .05, thus, a pooled variance t-test statistic was used to test the statistical significance of the difference of the means between the "Earlier“ and the I'Later" category. Since the indicated probability level for the pooled t—test was .34 which was greater than the criterion level of .05, we could not reject the null hypothesis. This analysis shows that the time lapse (i.e., two months) between the “Earlier" and the "Later“ measurements had no significant effect on. attitude toward. restriction. on abortion by states for the people who were religious Protestants and had "Heard“ of the Supreme Court decision on Webster vs. Reproductive Health Services. As Table 11 shows, the within group variances of the "Restrictiveness" score decreased 20 % from the ”Earlier" category to the “Later“ category or the difference between two categories was .217 (middle of the fifth column of the Table), but this difference‘was not statistically significant (F=1.26, p=.47). As the third column of the Table 11 shows, the mean "Restrictiveness“ scores of the I'Earlier" and the ”Later“ 73 TABLE 11 Restrictiveness Score by Category Among Religious Catholic Who had “Heard“ Mean Number Restrict— of tiveness Standard Category Cases Score Deviation Variance F p Earlier: 49 .980 1.031 1.063 Difference: .005 .111 .217 1.26 .47 %Difference: - 1% -11% —20% Later : 40 .975 .920 .846 74 categories were almost identical, .980 and .975, respectibly. Thus, there is no indication. in these data that “Restrictiveness" changed over this time period. The data showed that the within group variances and the means of the "Restrictiveness'I score ‘were not different between the “Earlier“ and the “Later“ categories for the people who are religious Catholics and had “Heard" of the Supreme Court decision on Webster vs. Reproductive Health Services. When the results of the analysis were considered, the data did not support Hypothesis 4, “Among those who had “Heard" and were religious, the variance on the Restrictiveness measure is greater in the Earlier category than in the Later category." In order to test Hypothesis 5, "Among those who had heard, and were religious, the variance on the Justification measure is greater in the Earlier category than in the Later category", the same respondents groups (i.e., 1 to 4) which were used for the testing of Hypothesis 4, were used again. Then, the mean "Justification" scores, the standard deviations, and the variances within each group were calculated. The F and the t test statistics and the observed significant levels between groups (1) and (2), and (3) and (4) were calculated. As Table 12 shows, a difference of‘within.group variances on the “Justification“ score between the “Earlier" and the “Later“ categories was .094 or 3% decrease from.the "Earlier" 75 TABLE 12 Justification Score by Category Among Religious Protestant Who had ”Heard" Mean Number Justifi— of cation Standard Category Cases Score Deviation Variance F p Earlier: 125 1.672 1.904 3.625 Difference: .344 .025 .094 1.03 .89 %Difference: +21% - 1% - 3% Later : 125 2.016 1.879 3.531 -.-.---—-————-—-————-—-—-—-—---.—_———-———e—-.—_———-----————- 76 category to the "Later” category (middle of the fifth column). The test statistic (F=1.03) was not large enough to confirm the significant difference of the within group variances between the “Earlier" and the “Later" categories p=.89). The null hypothesis could not be rejected. As the third column of the Table 12 shows, the difference of the mean "Justification“ scores between the two categories was .344 or the mean “Restrictiveness“ score in the “Laterf category was 21% increase from the l'Earlier" category. As mentioned earlier, the observed significant level for the F test was significantly larger than .05, thus, a. pooled variance t-test statistic was used to test the statistical significant of the difference of the :means between the “Earlier“ and the "Later“ category. This test (t=1.44) indicates a probability of .15 so we conclude that the data fail to support a rejection of the null hypothesis. The hypothesis of a change in Restrictiveness score is not supported. This analysis shows that during the two months period, there was not much change on attitude toward abortion for the people who are religious Protestants and had “Heard“ of the Supreme Court decision on Webster vs. Reproductive Health Services. As middle of the fifth column of Table 13 shows, the difference of within group variances of the “Justification" score between the “Earlier“ and the "Later" categories was 77 TABLE 13 Justification Score by Category Among Religious Catholic Who had ”Heard" Mean Number Justifi— of cation Standard Category Cases Score Deviation Variance F p Earlier: 47 1.489 1.864 3.474 Difference: .562 .408 1.354 1.64 .11 %Difference: -38% —22% —39% Later : 41 .927 1.456 2.120 78 relatively large (1.354). However, the probability level for the F test (p=.11) did not obtain the .05 criterion of statistical significant. We conclude that although the trend for people in the “Earlier“ category to be more heterogenous than the people of the "Later" category on the justification of the abortion issue, was as hypothesized, the difference‘was not statistically significant. As the second column of the Table 13 shows, the difference of the mean “Justification" score between the “Earlier“ and the "Later“ category was .562 which was a 38% decrease fromlthe “Earlier“ to the "Later" categoryz However, since the indicated probability level of the pooled t—test was .02, larger than the criterion of .05, we could.not reject the null hypothesis. The data showed that although there was a tendency for the people in the I'Earlier" category to be more supportive of abortion than the people of the "Later“ category, the difference was not statistically significant. The data suggested.that the religious Protestants who had ''Heard" of the Supreme Court decision on Webster vs. Reproductive Health Services in both the I'Earlier" and the "Later“ categories were not much different in their supportiveness on.abortion.as well as their homogeneity in the opinion. Although statistically not significant, the religious Catholics who had “Heard“ in the “Earlier“ category supported more on abortion and showed more heterogenous opinion on abortion than their counterpart in the “Later" 79 category. In sum, Hypothesis 5, ““Among those who had "Heard“ and.were religious, the variance on the Justification measure is greater in the Earlier category than in the Later category.“ was not fully supported by the data. In order to test Hypothesis 6, "Among those who had heard, and were secular, the variance on the Restrictiveness measure is greater than among those who had heard, and were religious, in the Later category“, the respondents were grouped into four categories: (1) secular Protestants who had “Heard“ of the Supreme Court decision.in the “Later" category, (2) religious Protestants who had “Heard” of the SupremeTCourt decision in the "Later“ category, (3) secular Catholics who had “Heard" of the Supreme Court decision in the “Later“ category, (4) religious Catholics who had “Heard“ of the Supreme Court decision in the “Later” category. Then, the mean “Restrictiveness" scores, the standard deviations, and the variances within each group were calculated. The F and the t test statistics and the observed significant levels between groups (1) and (2), and (3) and (4) were calculated. .As Table 14 shows, the within group variances of the "Restrictiveness“ score increased 13% from the secular Protestants who had "Heard" of the Supreme Court decision in the "Later“ category to the religious Protestants who had “Heard“ in the “Later“ category or the difference between the two groups was .122 (middle of the fifth column). Since the indicated probability level for the F test (.57) was larger 80 TABLE 14 Restrictiveness Score by Religiosity Among Protestant Who had heard in the Later Category Mean Number Restric— of tiveness Standard Category Cases Score Deviation Variance F p Secular: 71 1.845 .951 .904 Difference: .547 .062 .122 1.13' .57 %Difference: *30% + 7% +13% Religious: 114 1.298 1.013 1.026 81 than the critical point of .05, we could not reject the null hypothesis. As the third column of the Table 14 shows, the difference of the mean ”Restrictiveness“ scores between the two groups was .547 cu ea 30% decrease from the secular group to the religious group. A significant of this difference 'was confirmed by the indicated probability level for the pooled t- test, p<.01, whichtwas smaller than the critical point of .05. The results of the analysis suggested that the secular protestant who had “Heard“ of the Supreme Court decision in the “Later“ category favored more the restriction on abortion by the states than their religious counterparts, however, homogeneity of opinion was about the same between the two groups. As Table 15 shows, the difference of within group variances on the “Restrictiveness” scores between the secular Catholics who had "Heard'I of the Supreme Court decision in the “Later“ category and the religious Catholics who had “Heard“ in the ”Later“ category was .034 or a 4% decrease from the secular group to the religious group (the middle of the fifth column). The indicated.probability level for the F test (.90) does not attain the .05 criterion level, so we cannot claim a difference in homogeneity on this measure, and our substantial hypothesis is not supported. As the third column of the Table 15 shows, the difference of the mean “Restrictiveness" scores between the two groups 82 TABLE 15 Restrictiveness Score by Religiosity Among Catholic Who had heard in the Later Category Mean Number Restric- of tiveness Standard Category Cases Score Deviation Variance F p Secular: 34 1.706 .938 .880, Difference: .731 .018 .034 1.04 .90 %Difference: -43% - 2% - 4% Religious: 40 .975 .920 .846 83 was .731 or a 43% decrease from the secular group to the religious group. Since the indicated probability level for the pooled t-test was less than .01, we reject the null hypothesis. The conclusion is that the secular Catholics who had heard of the Supreme Court decision in the “Later“ category favored more of the restrictions on abortion by the states than did their religious counterparts. In sum, the difference of within group variances on the “Restrictiveness“ scores between the secular and the religious groups of the Protestants and the Catholics who had heard in the “Later" category was not statistically significant. Meanwhile, the secular Protestants and Catholics who had “Heard” in the “Later" category favored more of the restrictions on abortion by the states than their counterparts. When the results were considered, the data did not support Hypothesis 6, which states that “Among those who had “Heard", and were secular, the variance on the Restrictiveness measure is greater than. among those ihad "Heard“, and were religious, in the Later category." In order to test Hypothesis 7, “Among those who had "Heard", and were secular, the variance on the Justification measure is greater than among those who had "Heard", and.were religious, in the Later category“, the same respondents groups (i.e., 1 to 4) which were used for testing Hypothesis 6, were used again. Then, mean “Justification“ scores, standard deviations, and variances within each group were calculated. 84 TABLE 1 6 Justification Score by Religiosity Among Protestant Who had heard in the Later Category Mean Number Justifi- of cation Standard Category Cases Score Deviation Variance F p Secular: 77 3.468 1.209 1.462 Difference: 1.452 .67 2.069 2.41 <.01 %Difference: -42% +55% +142% Religious: 125 2.016 1.879 3.531 -——————---———————-————-—--———-—-—-———--——-—--——-——--—--— 85 The F and the t test statistics and the observed significance levels between groups (1) and (2), and (3) and (4) were calculated. As Table 16 shows, the within group variances of the “Justification“ scores, contrary to our prediction, increased 142% from the secular Protestants who had “Heard“ of the Supreme Court decision in the "Later" category to the religious Protestants who had “Heard" in the "Later" category or the difference between the two groups was 2.069 (middle of the fifth column). Since the indicated probability level for the F test (p<.01) was lower than the criterion level (.05), we reject the null hypothesis. As the third column of the Table 16 shows, the difference of the mean "Justification“ scores between the two groups was 1.452 or a 42% decrease from the secular-group to the religious group. The statistical significance of this difference was confirmed by the indicated probability level for the separate t-test, p<.01, which was smaller than the criterion of .05. The results of the analysis indicate that the secular protestants who had "Heard“ of the Supreme Court decision in the “Later“ category had higher justification score on abortion than their religious counterparts. The data also showed that the religious Protestants who had “Heard“ in the I'Later" category, were significantly more heterogenous in opinions on abortion than their secular counterparts. 86 TABLE 17 Justification Score by Religiosity Among Catholic Who had heard in the Later Category Mean Number Justifi- of cation Standard Category Cases Score Deviation Variance F p Secular: 35 3.027 1.543 2.381 Difference: 2.1 .087 .261 1.12 .71 %Difference: -69% - 6% -11% Religious: 41 .927 1.456 2.120 87 As the middle of the fifth column of Table 17 shows, the difference of within group variances of the “Justification" score between the secular Catholics who had "Heard" in the “Later“ category and the religious Catholics was .261 or a 11% decrease from the secular group to the religious group. The indicated probability level for the F test (.71) was not small enough to rejected the null hypothesis. As the third column of the Table 17 shows, the difference of the mean “Justification“ score between the secular and the religious Protestant groups was 2.1 which was a 69% decrease from the secular group to the religious group. Since the indicated probability level of the pooled t-test was less than .01, we could reject the null hypothesis. The data showed that the secular Catholics who had “Heard“ the Supreme Court decision in the "Later“ group wereumore permissive»on abortion than their religious counterparts. The data indicate that both the secular and the religious Catholics who had “Heard" of the Supreme Court decision in the ‘Later' categories were not much different in their homogeneity in the opinion, However, both the secular Protestants and Catholics were mere permissive on abortion than.their religious counterparts. zusa.resu1t, Hypothesis 7, "Among those‘who had “Heard“ and were secular, the variance on the Justification measure is greater than among those who had heard, and were religious, in the Later category." was not supported by the data. 88 Hypothesis 8: Black men are less supportive of abortion than white men, and black women are more supportive than white women. Hypothesis 8 was formulated to examine whether there were any discernable racial differences in abortion attitudes, as given by these data. As discussed earlier (see pages 40—41), a number of studies have found certain racial differences in abortion attitudes. However, the results of recent studies suggest that racial differences in abortion attitudes were somewhat reduced when religious variables such as religiosity, a :measure of doctrine, and. a .measure of denominational fundamentalismn were included. in the (analysis (Hall and Ferree, 1986; Wilcox, 1990). Although the CBS/New York Times abortion survey was not particularly rich in religious-variable information, we could measure the “religiosity" of the respondents by using the item which asked “how important is religion in your daily life?" The answers of "extremely“ were coded as "religious", and the others were coded as "secular". In order to test Hypothesis 8, the “Justification" measure was used as the dependent variable, and race was used as the independent variable. Also, gender and religiosity were used as the control variables. The mean “justification" score was calculated, and the ; statistic was used to determine the significance of the difference. 89 TABLE 18 Mean Score of Justification, Gender and Race White Black Justifi- Justifi— cation cation Diff- %Diff— t p Sub— Score Score erence erence Group (N / %) (N / %) Male: 2.185 1.641 .544 -25% 1.822 .04 (433/44) (39/35) Female: 1.933 2.206 .273 +14% 1.186 .23 (553/56) (73/65) Total 986 112 Religious Male: .875 .634 .241 -28% .514 .60 (72/31) (ll/28) Religious Female: 1.027 1.655 .628 +61% 1.895 .05 (158/69) (29/72) Total: 230 40 Secular Male: 2.442 2.035 .407 -17% 1.191 .23 (360/47) (28/39) Secular Female: 2.253 2.568 .315 +14% 1.088 .27 (402/53) (44/61) Total: 762 72 90 As Table 18 shows, when the male respondents were considered, the difference of the mean “Justification“ score between the white males and the black males was .544 (first line of the Table). The mean "Justification“ score was decreased by 25% from the white males to the black males. The t-score of the male respondents was 1.822. ‘When alpha is .05 with 470 degrees of freedom, the critical value is 1.645 for a one—tailed test. Since the test statistics was 1.822, we could reject the null hypothesis. Thus, we conclude that the black males were significantly less supportive of abortion than the white males. As the second line of Table 18 shows, when the female respondents were considered, the difference of the mean “Justification“ score between the whites and the blacks was .273. The mean “Justification“ score increased by 14% from the white females to the black females. The t—score of the female respondents 'was 1.186, slightly‘ smaller than the critical value of 1.645 when alpha is .05 with 624 degrees of freedom. The data show that although the mean “Justification" score for the black females was higher than the mean for white females, the difference was not statistically significant. As the third line of the Table 18 shows, difference between the religious white males and the religious black males was .241 or 28% decrease from.the religious white males to the religious black males. Since the t statistic (.514) was smaller than the critical value of 1.658 when alpha is .05 91 with 81 degrees of freedom, the difference was not significant. The data showed that although, the relative size of the two comparable sample means was as expected, with the religious black males supporting abortion slightly less than the religious white:males, the difference‘was not great enough to attain statistical significance (i.e. , the difference could have been attributable to random variation). We note that the difference of the mean I‘Justification" score between the religious white females and the religious black females was .628 or 61% increase from the religious white females to the religious black females. The t statistic of 1.895 was significant, since the critical value was 1.645 when alpha is .05 with 175 degrees of freedom. Based on the data, we conclude that the religious black females were significantly more supportive of abortion than the religious white females. As the fifth line of Table 18 shows, when the secular male respondents were considered, the difference of the mean ”Justification“ score between the whites and the blacks was .407. The mean “Justification“ score was decreased by 17% from the secular white males to the secular black males. We also note that the t score of the secular male respondents was 1.191, slightly smaller than the critical value of 1.645 when alpha is .05 with 366 degrees of freedom. The data showed that the secular black males were less supportive of abortion than the secular white males, however, the difference was not statistically significant. As the last line of the Table 18 92 shows, difference between the secular white females and the secular black females was .315 or 14% increase from the secular white females to the secular black females. Since the t statistic (1.088) was slightly smaller than the critical value of 1.645 when alpha is .05 with 444 degrees of freedom, the difference was not significant. We conclude that although the secular black females tended to support abortion slightly more than the secular white females, this tendency might be attributable to random variation. As Table 18 shows, the religious males and females (the second part of the Table) were significantly less supportive of abortion than the secular males and.females (the third.part of the Table). Also, the result of the analysis showed that in general, the black males were less supportive of abortion than the ‘white :males, and the black females were :more supportive of abortion than the white females regardless of the religiosity of the respondents. In general, the data support a modification of Hypothesis 8 which would state that 'Blackimen are less supportive of abortion than white men, and black women may be more supportive than white women.“ Since this modification derives from.the particular characteristics of these data, the modified hypothesis should be viewed somewhat skeptically until further independent data are provided. CHAPTER V INTERPRETATION OF RESULTS 1. SUMMARY OF DATA ANALYSIS To test the Hypotheses, the “Restrictiveness“ and "Justification" scores, and “abortion is same thing as murdering a child“ itemiwere used as the dependent variables. The 'Earlier“ and “Later" categories, the “Heard“ and “Not Heard“ categories were used as the independent variables. Also, race, religiosity, education, gender, religion, and age were used as the control variables to test the Hypotheses. Hypothesis 1, referring to the opinions of people in the “Earlier' and “Later" categories who had pp; heard of the Supreme Court decision on Webster vs. Reproductive Health Services, was confirmed by the data. In order to test Hypothesis 1, the “Restrictiveness“ on abortion was measured by a three-itemrrestrictiveness-index for each respondent in which a score of “0" indicates not restrictive and “3" indicates highly restrictive on abortion. The respondents who had "Not Heard” of the Supreme Court decision were selected from both the "Earlier“ and "Later“ categories. The test of 93 94 Hypothesis 1 suggested that the hearing of the Supreme Court decision could be attributed to the changing of people’s attitude toward abortion. Hypothesis 2, referring to the effect of the Supreme Court decision on people's attitude toward abortion, was also confirmed by the data. To test hypothesis 2, in addition to the "Restrictiveness" score, the “Justification“ score was measured by a four—item-justification-index for each respondent in which a score of '0' indicates not supportive and '4“ indicates very supportive of abortion. The respondents were divided into the “religious“ and I'secular" groups based on their self stated religiosity. There was a significant increase in support of restriction on abortion by states among people who had “Heard“ of the Supreme Court decision compared with people who had “Not Heard“ from both the “Earlier“ and "Later'I categories. This suggests that people's attitude toward abortion were, indeed, affected by the Supreme Court decision on Webster vs. Reproductive Health Services. Hypothesis 3, referring to the somewhat theological or philosophical issue of the rights of a fetus as a human, was not confirmed by the data. The same independent and control variables which were used to test Hypothesis 2 were again, used to test Hypothesis 3. The item “abortion is the same as murder"was used as a dependent variable to test Hypothesis 3. The "Heard“ and “Not Heard“ categories did not ShOW’ a. 95 significant difference in opinion on "abortion is the same as murder“ item from both the “Earlier“ and “Later“ categories. Hypotheses 4, 5, 6 and 7, referring to homogeneity of abortion attitudes in social context were not fully supported by the data. In order to test Hypotheses 4, 5, 6 and 7, the respondents who had “Heard“ of the Supreme Court decisioniwere selected from both the "Earlier“ and "Later“ categories. The “Restrictiveness“ and “Justification“ measures were, again, used as the dependent variables. The “Heard“ and “Not Heard“ categories were used as the independent variable, and the “Earlier“ and “Later“ categories and religiosity were used as the control variables. Most of the differences in the variances of the “Restrictiveness“ and “Justification“ scores among the “Heard” and “Not Heard“ categories, and the “Earlier“ and “Later“ categories were not large enough to confirm statistical significance. The data supported Hypothesis 8, referring to racial difference in abortion attitudes. In order to test Hypothesis 8, the respondents were divided into black males and females, and white males and females. The “Justification“ scores of each respondent were used as the dependent variable to test Hypothesis 8. .Also, race was used.as the independent variable and gender and religiosity were used as the control variables. The black males scored significantly lower than the white males in the "Justification'I measure and black females scored higher than the white females, but not to a 96 statistically significant degree. This confirmed that black men were less supportive of abortion than white men, and black women may be more supportive than white women. 2. EFFECTS OF THE SUPREME COURT DECISION ON ABORTION ATTITUDES As discussed earlier (see page 34), on July 3, 1989, the Supreme Court handed down the ruling on Webster vs. Reproductive Health Services which required: 1) Missouri doctors to test whether a fetus thought to be 20 weeks or older has reached a stage of development at which it could live outside the womb before considering an abortion; 2) public hospitals or other taxpayer-supported facilities may not be used for performing abortions not necessary to save life, even if no public funds are spent; 3) public employees, including doctors, nurses and other health care providers, may not perform or assist an abortion not necessary to save a woman’s life. In other words, as a result of the Supreme Court ruling, states since July 1989 have more authority to restrict abortion. In addition, the majority of justices let stand.a preamble to the Missouri act which declares that “the life of each human being begins at conception" and that “unborn children have all the rights, privileges and immunities available to other persons, citizens and residents of this state". 97 Also, as discussed earlier (see pages 15-16), the “legitimation" and “positive response“ hypotheses argue that when the Supreme Court rules, it confers legitimacy on the position it favors. The Supreme Court's prestige allows it to bestow legitimacy on others through its rulings. This legitimacy leads to increased public support for the Supreme Court’s position (Franklin and Kosaki, 1989; Marshall, 1987; Johnson andwCanon, 1984; Uslaner and Weber, 1980; Blake, 1977; Casey, 1974; Adamany, 1973; Murphy and. Tanenhaus, 1969; Kessel, 1966). Thus, we assumed that those who had “heard" of the Supreme Court decision on Webster vs. Reproductive Health Services would agree more on the restrictions on abortion by the states, and on the rights of the fetus as a human, while those who had “Not Heard“ would not change their opinions on those issues within a two months period (i.e., the time interval between the “Earlier" and “Later“ measurements). As table 1 (see page 46) shows, when cases where the respondents had “Not Heard“ of the Supreme Court decision, were considered, the difference of the “Restrictiveness“ score between the "Earlier“ and 'Later' categories was very small (.032) and was not statistically significant (first line of Table 1). On the contrary, when the cases of people who had “Heard" are considered, the difference was relatively large (.166) and was statistically significant. These results suggest that having heard of the Supreme Court decision may have been a factor in changing attitudes toward abortion, 98 because people who had not heard did not show a net change in their attitudes toward abortion in the two months period. However, there still remains the question of differences in composition between the “Heard“ and "Not Heard” categories. As Franklin and Kosaki (1989) pointed out, usually those who had.not heard of the Supreme+Court decision, are less educated than those who had heard. Since education, as well as some other variables such as religiosity and gender, is correlated with the attitudes toward abortion issues, the difference of attitudes toward abortion between the “Heard“ and “Not Heard“ category may be due to some other factors rather than having heard of the Supreme Court decision. In order to show the effects of having heard of the Supreme Court decision more clearly (in other words, controlling for other independent variables), we included several independent variables in the multiple regression model which was designed to test Hypotheses 2 and 3. As Table 3 shows (see page 52), those who had “Heard" agreed significantly more on the restrictions on abortion by states than those who had "Not Heard“ in the “Earlier" category (the last line of Table 3). Also, as Table 5 shows (see page 57), the “Heard“ category agreed significantly more on the restriction on abortion by states than the “Not Heard“ category in the “Later“ category (the last line of Table 5). These results confirmed the "legitimation“ and “positive response“ hypotheses which argue that when the Supreme Court 99 rules, it confers legitimacy on the position it favors, thus, leads to increased public support for the position taken by the Supreme Court. Hypothesis 3, referring to people’s opinions on the issue of when human life begins, was pg; supported by the data (In fact, the non-significant trend was contrary to the hypothesis). As Table 7 and 9 shows (see pages 62 and 65), those who had "Heard“ did pg; agree more on the item which stated that abortion is the same thing as murdering a child, compared to those who had "Not Heard“ in both the “Earlier“ and "Later“ categories. The results of the tests for Hypotheses 2 and.3 suggested some very interesting aspects of the effects of the Supreme Court rulings in general, as well as on abortion issues. First of all, as stated earlier, a Supreme Court ruling tends to legitimige the position the Supreme Court favors, thus, leads to increased public support for the position taken by the Supreme Court. The data supported this notion, as those who had "Heard" the Supreme Court decision agreed more on restriction on abortion by the states than those who had 'Not Heard'. However, the question of when human life begins, is app, a legal issue, but is a theological, cultural or philosophical issue. For example, although the majority of the Supreme Court let stand the preamble to the Missouri act which declares that "the life of each human being begins at conception" and that “unborn children have ...all the rights, 100 privileges and immunities available to other persons, citizens and residents of this state”, Justice Stevens (who dissented from the majority View), saw the preamble as essentially a theological argument that was invalid under the First Amendment and was "an unequivocal endorsement of a religious tenet of some but by no means all Christian faiths" (Shulman, 1989). An example of the cultural aspect of when human life begins can be found in the traditional age counting method in some Eastern countries, especially in Korea. When an infant is born, that infant is considered as one day old.in.a Western culture, however, that same infant is considered as one year old in the traditional Korean culture (actually nine months old - however, Koreans are a generally generous people!). This indicates that the Koreans traditionally considered life as having begun at conception or at least considered a fetus as a human being. In the traditional Korean belief system, the fetus could be affected not only by a mother's nutritional system but also by such things as the mental and ethical beliefs of the2motheru These theological and.cultural aspects of the issue of when human life begins, weakens the legal aspect of that same issue. Consequently, it may be that this "philosophical'I aspect weakens or removes the power of the SupremevCourt's legitimatization on this kind of issue. As a result, people were not affected by the Supreme Court's letting stand the preamble to the Missouri act. The test of Hypotheses 2 and 3 clearly showed that the legal prestige of 101 the Supreme Court (i.e., restriction on abortion by the states) and its limits (i.e., when human life begins), and the effects of the Supreme Court rulings on peOple’s attitudes. 3. CONTEXTUAL EFFECTS ON PEOPLE’S ATTITUDES TOWARD ABORTION As Sprague (1982) has stated, social scientists are primarily interested in the contextual effects because the contextual effects serve as evidence for interpersonal influence. Empirical studies of the contextual effects (Brown, 1987; Weatherford, 1982; Putnam, 1966) suggested that individual behavior and attitude formation are partly a function of the social environment. As discussed earlier (see page 20), Sprague (1982) pointed out that homogeneity in the social environment and solid embedment in an environment (i.e., people who have more of their interactions within a group) are two key factors to be affected by the group norms (i.e., contextual effects). Previous studies (Franklin and Kosaki, 1989; Wilcox, 1987) also suggested that religious identification was one of the most powerful predictors of the attitudes of people. Based on the previous studies and contextual theory, we selected religious Protestants and Catholics as the context (or micro-environment) to test the contextual effects. We viewed it as reasonable to assume that those who are more religious would have more contact. with 102 their micro—environments (i.e., Protestant or Catholic church members) than those who are less religious. This line of reasoning also assumes that “abortion" and "religion" are related realms of discourse. Then, we formulated the prediction that those ‘who are religious in the "Later" category would show more homogeneous attitudes toward abortion than those who are religious in the “Earlier category“, and those who are religious would shOW'more homogeneous attitudes toward abortion than those who are less religious. Hypotheses 4 and 5 were formulated to test 3111211 the contextual effects takes place. Previous studies in the contextual effects mestly concentrated on the attitudes or opinions differences between “before" and “after“ the social impact (or 'stimuli'). In order to identify the contextual effects, previous studies have ‘used. data. which. gathered several months after the social impact (Franklin and Kosaki, 1989) or several years after (Wilcox, 1987) as the contextually affected groups. Usually those data were gathered over a period of several months. Therefore, the researchers must have pursued their studies under the two assumptions: (1) it takes some time (probably at least several months) for people to be affected by their micro-environment, (2) there is no difference between the opinions gathered at the beginning of the period of the data gathering (usually several months) and the opinions gathered toward the end of the period. 103 The nature of the data used for this study has two distinctive advantages over the data used for previous studies, that is: (1) modern telephone interview techniques make it possible to gather data from a relatively large national sample in a very short time period (i.e., it took only four days to complete the whole interview), (2) the data included two datum, one was measured immediately after (three weeks after) the social impact (i.e., the SupremeeCourt ruling on Webster vs. Reproductive Health Services), and the other was measured some time later (three months after). This characteristics of the data provided an ideal setting for the test of when the contextual effects takes place. As Table 10 and 11 show (see pages 70 and 72), the within group variances and the.means of the ”Restrictiveness“ scores were not different between the ”Earlier“ and the “Later" categories for those who are religious Protestants and Catholics, and had heard of the Supreme Court decision on Webster vs. Reproductive Health Services. Also as Table 12 and 13 (see pages 74 and 76) show, the religious Protestants and Catholics who had heard of the Supreme Court decision in both the “Earlier“ and “Later“ categories were not much different in.their supportiveness on.abortion.aS‘well as their homogeneity in the opinion. Based on these results, we conclude that the time lapse (i.e., two months) between the “earlier“ measurement and the "later“ measurement had no significant effect on the people's attitudes toward abortion. 104 This finding suggested a somewhat important aspect of the contextual effects and the attitudes toward abortion of those who are religious. This finding suggested that it might not take a very long time period for people to be affected by their micro-environment, probably less than three weeks (the time period our “Earlier“ measurement was gathered after the social impact). If this assumption is correct, we could interpret the results of the data that people were already affected by their micro-environment before the “Earlier" :measurement, thus, there was no differenceeof opinions between the “Earlier“ and I'Later" categories. Another aspect of the results of the data is that those who are highly involved in the church, both Protestants and.Catholics, might be closer to their group norms before the social impact (i.e., the Supreme Court ruling on Webster vs. Reproductive Health Services) , and they :might be :more resistant to outsider's opinions on abortion. Hypotheses 6 and 7 were formulated to test the contextual effects which argues that those who are more solidly embedded in an environment (in this study, religious Protestants and Catholics), are most likely to be affected by the group norms. Previous studies (Franklin and Kosaki, 1989; Johnson and Canon, 1984; Combs and Welch, 1982; Uslaner and Weber, 1980) showed that those who are more religious are less supportive of abortion than those who are less religious, and those who are more religious are closer to group norms in the attitudes 105 toward abortion than those who are less religious. As Table 14 and 15 (see pages 79 and 81) show, the differences of‘within.group'variances on the "Restrictiveness" scores between the secular and religious groups of both the Protestants and.Catholics were not statistically significant. As also Table 16 and 17 (see pages 83 and 85) show, the secular and religious people of both the Protestant and Catholic persuasion were not much different in the homogeneity of their opinions on abortion. However, those who are more religious Protestants and. Catholics are :more opposed. to abortion than their less religious counterparts. The results of the data confirmed that those who are more religious are less supportive of abortion than those who are less religious. However, these data failed to support the aspect of the contextual effects which argues that those who are more solidly embedded in an environment are most likely to be affected by the group norms. These findings suggest that the religious groups might not be as homogeneous in their opinions on abortion as previously believed, .Also, the findings led us to reconsider the adequacy of the size of the sample used in this study. For example, as shown in Table 17 (see page 85), after the elimination of those who had “Not Heard” of the Supreme Court decision, only 41 religious and 35 secular Catholics were left to be analyzed. Thus, only a few extreme scores could skew the results, resulting in, high variances for both groups. In order to test the contextual effects, it 106 may be that one needs data with larger samples which cannot be disturbed by a few “outlying“ cases. 4. RACE DIFFERENCES IN ABORTION ATTITUDES The difference in,abortion rates — black women were twice as likely to have had an abortion as white women - has held constant during the last decade (Henshawrand Silverman, 1988). Yet, previous studies showed.that, in general, blacks are less supportive of legal abortion than whites (Secret, 1987; Hall and Ferree, 1986; Combs and Welch, 1982). According to the findings of some studies (Petchesky, 1984; Ransford and Miller, 1983) thereeis an interaction.between race andmgender, that is, black men were less supportive of abortion than white men, and black women were about the same or slightly mere supportive of abortion than white women. Petchesky (1984) argued.that black:men.may be less supportive of legal abortion than white men because some black.men associate abortion.with genocide for black males, while black women may be more supportive of abortion than white women because of the greater frequency of abortion in the black community. A couple of studies (Hall and Ferree, 1986; Combs and Welch, 1982) have reported that controls for religious variables such as religiosity, a measure of doctrine (interpretation of the Bible), and. a measure of denominational fundamentalism, 107 significantly' reduced. racial differences ix: abortion attitudes. Based on the results of the previous studies, we formulated a prediction that black men are less supportive of abortion than white men, and black women are more supportive than white women. The data used for this study did not provide much detailed information about religious variables; however we were able to get a rough measure of personal religiosity, and used this to test race difference in abortion attitudes when controlled by religiosity. As the first part of Table 18 (see page 88) shows, black men were significantly less supportive of abortion than white men.and black women were relatively more supportive than white women (not statistically significant). Also, as the second and third.part of Table 18 shows, those who are religious were significantly less supportive of abortion than those who are not religious regardless of race. Furthermore, overall, women were more supportive of abortion.than.men regardless of race. Among those who are not religious, difference in supportiveness of abortion was almost non—existent. Among those who are religious men were less supportive of abortion than women, and black men were less supportive of abortion than white men. And black women were more supportive of abortion than white women. These findings suggested that gender and race differences in abortion attitudes still exist and religiosity itself does not significantly reduce the race difference in abortion attitudes. However, the function of 108 religious variables in race difference in abortion attitudes are not very conclusive because of lack of other religious variables such as a measure of doctrine and denominational fundamentalism in the data used in this study. CHAPTER VI CONCLUSION 1. SUMMARY AND DISCUSSION This study was intended to provide answers to following questions: (1) is there any effect of the SupremeTCourt ruling on people's attitudes toward abortion issues, (2) if there is any effect of the Supreme Court ruling on people's attitudes toward abortion issues, how does that attitude change occur, (3) is there any race difference in attitudes toward abortion issues. The data gathered by CBS News/The New York Times from a national sample were described and analyzed, and tables were presented to show the effect of the Supreme Court ruling, the contextual effects on attitude change and the race difference in abortion attitudes. ‘R: fulfill the above purposes the research work was designed on two levels -- descriptive and analytical. On the descriptive level, overall characteristics of the respondents (e.g., religiosity; religious identification; education; age; gender; whether or not heard of the Supreme Court decision; the restrictiveness on 109 110 abortion; the permissiveness on abortion) were examined. On the analytical level, the eight Hypotheses were tested to analyze the effect of the Supreme2Court ruling, the contextual effects on attitude change and the race differencerin abortion attitudes. The followings are a summary of the most significant findings. There was a significant increase in support of restrictions on abortion by the states among those who had heard of the Supreme Court ruling on Webster vs. Reproductive Health Services compare with those who had not heard. Those who are religious ‘were significantly'.l§§§_ supportive» of abortion than those who are not religious. Black men were significantly less supportive of abortion than.white men, and black women were mpre supportive of abortion than white women. Hypothesis 1 was based on the assumption that without any stimuli or social impact (in this study, the Supreme Court ruling), people ‘would. not change their“ attitudes toward abortion. This Hypothesis was confirmed by the data of this study. The result strengthened the conclusions drawn from the testing of the other Hypotheses concerned with the effects of the Supreme Court ruling. Hypothesis 2 and 3 were based on the "legitimation" and the “positive response" hypotheses regarding the effects of the Supreme Court rulings. Hypothesis 2 was confirmed by the data of this study, while Hypothesis 3 was not supported by the data. Based on the result, we concluded that the Supreme 111 Court ruling increased public support for the legal position taken by the Supreme Court Hypotheses 4, 5, 6, and 7 were based on the contextual effects which argues that those who are more solidly embedded in an environment, are most likely to be affected.by the group norms, and the homogeneity in the social environment is the key variable in attitude change. Contrary to expectation, these Hypotheses were not fully supported by the data. The results suggested.that the contextual effects might take place much faster than previously' assumed, The results also suggested that the religious groups might not be as homogeneous in opinions on abortion, In addition, the results suggested that perhaps data from a larger number of people with a strong religious orientation might be needed to test the contextual effects. Hypothesis 8 was based on the findings of previous studies which indicate that there probably are important race differences in. abortion attitudes. This hypothesis was supported.by the data of this study; Based on the results, we conclude that black men are less supportive of abortion than white men, and black women are more supportive of abortion than white women. The results also suggested that controlling for the religiosity variable did pg; reduce the race difference in abortion attitudes. 112 2. SUGGESTIONS FOR FUTURE STUDY The following areas of concern are recommended.for future study. The design of this study was somewhat limited due to available dataq The research would have been.more complete if it had included data gathered prior to the Supreme Court ruling on Webster vs. Reproductive Health Services. Adding a “before“ group in the study would provide valuable information on the issue of effect of the Supreme Court rulings as well as the trends of changes in the attitudes toward abortion issues. For example, the data used in this study showed that support of restrictions on abortion by the states increased slightly as well as justification of abortion during the first three months regardless of whether one had or had not heard of the Supreme Court ruling. However, in order to determine a trend of abortion attitudes, one would.need data gathered before the Supreme Court ruling was issued. The support by these data for Hypotheses 1 and 2, compared to the contrary indication (lack of support) for Hypothesis 3, underlines that the exact way in which measurements are operationalized can lead to initial differences in research results, particularly when dealing with a complex, highly emotional, multi-faceted (probably multi-dimensional), and very controversial issue such as abortion. “Restrictiveness" and Justification“ can be viewed 113 as examples of the kind of research measurement that can perhaps lead toward analyses which reflect the complexity of the various abortion issues, and may lead to greater understanding of these complexities. The tests of the contextual effects were not very fruitful, perhaps because our “earlier" measurement was not early enough to measure the purely individual reactions to the social impact. It would. be meaningful to gather data immediately after a social impact (within a couple of days), and gather follow—up data some time later (say, about three months). This kind of data.would provide valuable information about how fast the contextual effects take place. In addition, information on M one heard about the Supreme Court ruling would be helpful to examine the contextual effects. For example, there could be a difference if one learned about an important social event from.a relatively impartial source, compared to learning about it from a more biased source. Data with more religious variables, such as religiosity ammi more specified religious identifications, is needed to test race difference in abortion attitudes while controlling for the religious variables. Results from the data with.more religious variables would provide more complete findings. Considerably more research.is needed to provide the keys to untangling the complex relationships among (1) socially significant events, such.as Supreme:Court decisions, (2) basic demographic factors, such as education, race, and social 114 status (some of which has been considered herein) and (3) the attitudes of various sectors of the population (which we assume underlie how people participate in social life). BIBLIOGRAPHY BIBLIOGRAPHY Achen, Christopher H. Interpreting and Using Regression. Beverly Hills: Sage Publications, 1982. Adamany, David. "Legitimacy, Realigning Elections, and the Supreme Court.“ Wisconsin Law Review, (1973): 790-846. Cited by Charles H. Franklin, and Liane C. Kosaki, “Republican Schoolmaster: The‘U.S. Supreme Court, Public Opinion, and Abortion.‘I American Political Science Review. 83, 3, (September 1989): 751-771. Allport, G.W. "Attitudes.'I In Handbook of Social Psychology. pp.798-844. Edited by C.M. Murchison. Worcester, MA: Clark Univ. Press, 1935 Cited by G.W. Allport, “The Historical Background of Social Psychology.“ In Handbook of Social Psychology, Vol. I 3rd ed. pp.1-46. Edited by G. Lindzey and E. Aronson. New York, NY: Random House, 1985. Allport, G.W. "The Historical Background of Social Psychology.‘ In Handbook.of Social Psychology. Vol. I 3rd ed. pp.1—46. Edited by G. Lindzey and E. Aronson. New York, NY: Random House, 1985. Alwin, Duane F. and Krosnick, Jon A. ”Aging, Cohorts, and the Stability of Sociopolitical Orientations over the Life Span.“ American Jougnal of Sociology. v.97, 1, (July 1991): pp. 169-95. Anderson, Alun. "Abortion Ruling Divides the United States.“ Nature, July 13 1989, p. 83. Achen, Christopher H. Interpreting and Using Reggession. Edited by John L. Sullivan, and Richard G. Niemi. Beverly Hills, CA: Sage Publishing, Inc., 1982. Babbie, Earl. The Practice of Social Research. 4th. ed. Belmont, CA: Wadsworth Publishing Co., 1986. Berry, William.D. and Stanley Feldman, Multiple Regression in Practice. Beverly Hills, CA: Sage Publishing, Inc., 1985. 115' 116 Blake, Judith. “The Abortion Decisions: Judicial Review and Public Opinion." In Abortion: New Directions for Policy Studies. Edited by Edward Manier, William.Liu, and David Solomon. Notre Dame: University of Notre Dame Press, 1977. Cited by Charles H. Franklin, and Liane C. Kosaki, “Republican Schoolmaster: The U.S. Supreme Court, Public Opinion, and Abortion.“ American Political Science Review. 83, 3, (September 1989): 751-771. Bray, James H. and Scott E. Maxwell, Multivariate Analysis of Variance. Beverly Hills, CA; Sage Publishing, Inc., 1985. Brown, Thad A., Migration and Politics. Chapel Hill: University' of North. Carolina Press, 1987. Cited in Michael MacKuen and Courtney Brown., "Political Context and.Attitude Change.“ American Political Science Review. 81, 2 (June 1987): pp. 471—490. Campbell, Donald T., and Stanley, Julian C. Experimental and Quasi-Experimental Designs for Research. Chicago: Rand McNally College Publishing Co., 1963. Carlin, Jr. David R. 'As American As Freeways: Abortion’s Roots in.Our Culture." Commonweal, July 14, 1989: pp.392- 393. Casey, Gregory. “The Supreme Court and Myth: An Empirical Investigation.” Law and Society Reviews. 8, (1974): 385- 419. Cited by Charles H. Franklin, and Liane C. Kosaki, “Republican Schoolmaster: The U.S. Supreme Court, Public Opinion, and Abortion." American Political Science Review. 83, 3, (September 1989): 751-771. CBS News/The New York Times. CBS NEWS/NEW YORK TIMES ABORTION POLL, JULY 1989 [Computer file]. New York, NY: CBS News [producer], 1989. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 1991. CBS News/The New York Times. CBS NEWS/NEW YORK TIMES ABORTION POLLL,SEPTEMBER—OCTOBER 1989 [Computer file]. New York, NY: CBS News [producer], 1989. Ann Arbor, MI: Inter- university Consortium for Political and Social Research [distributor], 1991. Combs, M., and Welch S. “Blacks, Whites, and Attitudes Toward Abortion.“ Public Opinion Quarterly. 46 (1982): pp. 510- 520. Cited by Clyde Wilcox, "Race Differences in Abortion Attitudes: Some Additional Evidence." Public Opinion Quarterly. (54) 1990: pp. 248—255. 117 Cook , Thomas D . , and Campbel l , Donald T . Quas i - Experimentation: Design & Analysis Issues for Field Settings. Chicago: Rand McNally College Publishing Co., 1979. Dayton, Cornelia H. "Taking the Trade: Abortion and Gender Relations in arm Eighteenth—Century New England Village . “ The William and Mary Quarterly 48 (January 1991) : 19-49. Droba, D.D. “The Nature of Attitude." Journal of Social Psychology no. 4 (1933): 444-463. Cited by G.W. Allport, "The Historical Background of Social Psychology.“ In Handbook of Social Psychology, Vol. I 3rd ed. pp.1-46. Edited by G. Lindzey and E. Aronson. New York, NY: Random House, 1985. Franklin, Charles H., and Kosaki, Liane C. "Republican Schoolmaster: The U.S. Supreme Court, Public Opinion, and Abortion.” American Political Science Review. 83, 3, (September 1989): 751-771. Gillespie, Marcia Ann. "Repro Woman.“ Mg, October 1989, pp. 50-53. Hall, E., and Ferree, M. “Race Differences in Abortion Attitudes.“ Public (pinion Quarterly. 50 (1986): pp. 193- 207 . Cited by Clyde Wilcox, "Race Differences in Abortion Attitudes: Some Additional Evidence.“ Public Opinion Quarterly. (54) 1990: pp. 248—255. Hatman, F.S., 'An experimental study of the effects of ethos in public speaking. " Speech Monograph. 16 (1949) : pp.190- 202. Henkel, Ramon E., Tests of Significance. Edited by Reic M. Uslaner, Beverly Hills, CA: Sage Publications, Inc., 1976 Hensaw, S, and Silverman, J. "The Characteristics and Prior Contraceptive use of U.S. Abortion Patients. " Family Planning Perspectives. 20, 4 (1988): pp. 158-168. Cited by Clyde Wilcox, “Race Differences in Abortion Attitudes: Some Additional Evidence. " Public Opinion Quarterly. (54) 1990: pp. 248-255. Hovland, C.I., and W. Weiss., “The Influence of Source Credibility on Communication Effectiveness.“ Public Qpinion Quarterly., 15 (1952): pp. 635-650. 118 Hyman, Herbert H., “The Psychology of Status.“ Archives of psychology., 269 (1942): Reprinted in Readings in Reference Group Theory and Research. Edited by Herbert H. Hyman and Eleanor Singer. , New York: The Free Press, 1968. Hyman, Herbert H., and Eleanor Singer, Readings in Reference Group Theory and.Research. Edited.by Herbert.H. Hyman.and Eleanor Singer., New York: The Free Press, 1968. Johnson, Charles A., and Canon, Bradley C. “Judicial Policies: Implementation and Impact." Congressional Quarterly. 1984. Cited by Charles H. Franklin, and Liane C. Kosaki, “Republican Schoolmaster: The U.S. Supreme Court, Public Opinion, and.Abortion.‘I American Political Science Review. 83, 3, (September 1989): 751-771. ' Katz, D “The Functional Approach to the Study of Attitudes." Public Opinion Quarterly 24 (1960): 163—204. Cited by Richard E. Petty, and John T. Cacioppo. Attitudes and .P_e_i;suasion: Classic and Contemporaer Apmroaches . Dubuque, Iowa: Wm. C. Brown Co. Publishers, 1981. Kelley, Harold H., “TWO Functions of Reference Groups.“ in Readings in Reference Group Theogy and Research. Edited by Herbert H. Hyman and Eleanor Singer., New York: The Free Press, 1968. Kelman, H. C., and. C.I. Hovland, “Reinstatement of the communicator in delayed measurement of opinion Change.“ Journa; of Abnorma; Social Psychology., 48 (1953): pp. 327—335. Kessel, John. “Public Perceptions of the Supreme Court." Midwest Journal of Political Quarterly. 10, (1966): 167— 91. Cited by Charles H. Franklin, and Liane C. Kosaki, “Republican Schoolmaster: The U.S. Supreme Court, Public Opinion, and Abortion.“ American Political Science Review. 83, 3, (September 1989): 751-771. Knoke, David, and Bohrnstedt, George W., Basic Socia; Statistics. Itasca, IL: F.E. Peacock Publisher, Inc., 1991. Kulp, D.H. “Prestige as Measured by Single—Experience Changes and.Their Permanency.“ Journal of Educational Research., 27 (1934): pp. 663-672. 119 Lerner, Ralph. I“The Supreme2Court.aisepublican.Schoolmaster." In Supreme Court Review. Edited by Philip B. Kurland. Chicago: University of Chicago Press, 1967. Cited by Charles H. Franklin, and Liane C. Kosaki, "Republican Schoolmaster: The U.S. Supreme Court, Public Opinion, and Abortion." American Political Science Review. 83, 3, (September 1989): 751-771. Lewis-Beck, Michael S. Applied Regression: An Introduction. Beverly Hills: Sage Publications, 1980. Liska, J. “Situational and Topical Variations in Credibility Criteria.“ Communication.Monographs, 45, (1978): pp. 85- 92. MacKuen, Michael, and Brown, Courtney. “Political Context and Attitude Change.“ American Political Science Review. 81, 2 (June 1987): pp. 471-490. Marshall, Thomas. “The Supreme Court As an Opinion Leader." American Politics Quarterly. 15, (1987): 147-68. Cited by Charles H. Franklin, and Liane C. Kosaki, “Republican Schoolmaster: The U. S. Supreme Court, Public Opinion, and Abortion.“ American Political Science Review. 83, . 3, (September 1989): 751- 771. Murphy, Walter F., and Tanenhaus. “Public Opinion and the United States Supreme Court: A Preliminary Mapping of Some Prerequisites for Court Legitimation of Regime Changes.“ In Frontiers of Judicial Research. Edited by Joel B. Grossman and Joseph Tanenhaus. New York NY: John Wiley & Sons, 1969. Cited by Charles H. Franklin, and Liane C. Kosaki, "Republican Schoolmaster: The U.S. Supreme Court, Public Opinion, and Abortion.“ American Political Science Review. 83, 3, (September 1989): 751- 771. New York Times, 4 July 1989, p. A1. Norusis, Marija J. SPSS: Statistical Data Analysis. Chicago, IL: SPSS Inc, 1990. Petchesky, R. Abortion and.Woman’s Choice. New York: Longman, 1984. Cited by Clyde Wilcox, “Race Differences in Abortion Attitudes: Some Additional Evidence.“ Public Opinion Quarterly. (54) 1990: pp. 248-255. Petty, Richard E. and. Cacioppo, John. T. Attitudes and Persuasion: Classic and Contemporary Approaches. Dubuque, Iowa. Wm. C. Brown Co. Publishers, 1981. 120 Putnam, Robert D., “Political Attitudes and the Local Community.“ American Political Science Review. v,60 (1966) : pp. 640-54. Cited in Michael MacKuen and Courtney Brown. , “Political Context and Attitude Change. “ American Political Science Review. 81, 2 (June 1987): pp. 471-490. Ransford, E., and Miller, J. “Race, Sex, and Feminist Outlooks.“ American Sociological Review. 48 (1983): pp. 46-59. Cited by Clyde Wilcox, “Race Differences in Abortion Attitudes: Some Additional Evidence.“ Public Opinioanuarterly. (54) 1990: pp. 248-255. Rosenblatt, Roger. “How'To End.The Abortion.War.“ The New York Times Magazine. January 19, 1992, pp. 26, 41-42, 50, 56. Salholz, E.; MaDaniel, A.; King, P.; Joseph, N.; Cerio, G.; and Carroll, G. “The Battle Over Abortion. “ Newsweek, May 1 1989, pp. 28-38. Schroeder, Larry D., Sjoquist, David L., and Stephan, Paula E. UnderstandinggRegression Analysis. Beverly Hills: Sage Publications, 1986. Secret, P. “The Impact of Region on Racial Differences in Attitudes Toward Legal Abortion.“ Journal of Black Studies. 17 (1987): pp. 347-369. Cited by Clyde Wilcox, “Race Differences in Abortion Attitudes: Some Additional Evidence.“ Public Opinion Quarterly. (54) 1990: pp. 248— 255. Shulman, Seth., “Abortion Ruling Divides the United States.“ Nature. V340, n13 (July 1989): p. 83. Smith, M. ., Bruner, J. S., and White, R. W. Opinions and Personality. New York: Wiley, 1956. Cited by Richard E. Petty, and John T. Cacioppo. Attitudes and Persuasion: Classic and Contemporary Approaches. Dubuque, Iowa: Wm. C. Brown Co. Publishers, 1981. Spector, Paul E. Research Designs. Edited by John L. Sullivan. Beverly Hills, CA: Sage Publishing, Inc., 1981 Sprague, John. “Is there a Micro Theory Consistent with Contextual Analysis?“ In Strategies of Political Inguigy. Edited by Elinor Ostrom, Beverly Hills, CA: Sage, 1982. Spector, Paul E. Research Designs. Beverly Hills: Sage Publications, 1981 121 Superior Court of Connecticut, Rex v. John Hallowell et al., Superior Court Records, Book 9, 113, 173, 175. Cited by Cornelia H. Dayton. “Taking the Trade: Abortion and Gender Relations in arm Eighteenth-Century New England Village.“ The William and Mary Quarterly 48 (January 1991): 19-49. Tingsten, H. Political Behavior: Studies in Election Statistics. London: P.S. King, 1937. Cited by John Sprague “Is there a Micro Theory Consistent with Contextual Analysis?“ In Strategies of Political Inquiry. Edited by Elinor Ostrom, Beverly Hills, CA: Sage, 1982. Uslaner, Eric M., and Weber, Ronald B. “Public Support for Pro-Choice Abortion Policies in the Nation and the States: Changes and Stability after the Roe and Doe Decision. “ In The Law and Politics of Abortion. Edited by Carl E. Schneider and Maris A. Vinovskis. Lexington, MA: Lexington, 1980. Cited by Charles H. Franklin, and Liane C. Kosaki, “Republican Schoolmaster: The U.S. Supreme Court, Public Opinion, and Abortion.“ American Political Science Review. 83, 3, (September 1989): 751-771. Warren, H.C., Ed. Dictionary of Psychology. Boston, MA: Houghton Mifflin, 1934. Cited by G.W. Allport, "The Historical Background of Social Psychology.“ In Handbook of Social Psychology, Vol. I 3rd ed. pp.1-46. Edited by G. Lindzey and E. Aronson. New York, NY: Random House, 1985. Weatherford, M. Stephen., “International Networks and Political Behavior.“ American Journal of Political Science. v,26 (1982): pp. 117—43. Cited in Michael MacKuen and Courtney Brown., “Political Context and Attitude Change.“ American Political Science Review. 81, 2 (June 1987): pp. 471-490. Wilcox, Clyde. “Race Differences in Abortion Attitudes: Some Additional Evidence.“ Public Opinion Quarterly. (54) 1990: pp. 248-255. Windham County Superior Court, Rex v. Hallowell, Windham County Superior Court Files, Box 172. Cited by Cornelia H. Dayton. “Taking the Trade: Abortion and Gender Relations in arm Eighteenth-Century New England Village. “ The William and began! Quarterly 48 (January 1991): 19-49. Young, K., Ed. Social Attitudes. New York: Henry Holt, 1931 Cited by G.W. Allport, “The Historical Background of Social Psychology.“ In Handbook of Social Psychology, Vol. I 3rd ed.-pp.l—46. Edited by G. Lindzey and E. Aronson. New York, NY: Random House, 1985. APPENDICES APPENDIX A SELECTED ITEMS USED IN THIS STUDY FROM CBS NEWS/NEW YORK TIMES ABORTION POLLS, JULY and SEPTEMBER, 1989 JULY C# SC Q# APPENDIX A SELECTED ITEMS USED IN THIS STUDYGIEROM CBS NEWS/NEW YORK TIMES ABORTION POLLS, JULY and SEPTEMBER, 1989 C#: Card Number SC: Starting Column Q#: Question Number SEPTEMBER C# SC Q# DESCRIPTION 1 22 2 44 9 2 45 9 1 22 2 55 17 19 2 57 19 2 58 19 Respondent’s Sex 1=Male 2=Female Earlier this month the Supreme Court decided a major case about abortion. Do you approve or disapprove of this decision, or haven't you heard enough about it to have an opinion? v1=Approve 2=Disapprove 3=Haven’t heard enough 9=DK/NA Should it be legal or should it be against the law for a pregnant woman to obtain an abortion if ..... 1=Should be legal 2=Should be against the law 3=Depends (vol.) 9=DK/NA a. the woman’s own health is seriously endangered by the pregnancy? b. she is not married and does not want to marry the father? 3For complete survey questionnaires, contact Inter-university Consortium for Political and Social Research, P.O. Box 1248, Ann Arbor, Michigan, 48106 (July - ICPSR 9488; September - ICPSR 9491). 122 46 47 48 49 55 56 57 61 64 67 15 15 15 15 16 19 20 2 59 19 2 60 19 2 61 19 2 62 19 25 2 69 25 2 71 25 2 72 25 2 74 26 3 72 50 3 64 43 123 c. the family has a very low income and cannot afford any more children? d. the pregnancy would force a teenage girl to drop out of school? e. there is a strong chance of serious defect in the baby? f. the pregnancy would force a professional woman to interrupt her career? Here are some possible restrictions on abortion that are being debated in some states. Would favor or oppose ...... 1=Favor 2=Oppose 3=Depends (vol.) 9=DK/NA a. Requiring the consent of her parents before a girl under 18 years of age could have an abortion? b. Requiring a test to make sure that the fetus is not developed enough to live outside the womb before a woman could have an abortion? c. Prohibiting public employees or public hospitals from performing abortions? ASK EVERYBODY: Which of these statements comes closer to your opinion -- abortion is the same thing as murdering a child, OR abortion is not murder because the fetus really hasn't developed into a child yet? 1=Murder 2=Not murder 3=Depends (vol.) 9=DK/NA How old are you? 99=DK/NA How important is religion in your daily life -- is it extremely important, very important, somewhat important, or not at all important? 1=Extremely 2=Very 3=Somewhat 4=Not at all 9=DK/NA 70 70 75 3 68 46 3 71 49 3 75 51 124 What is your religious preference today? SPECIFY 1=Protestant 2=Catholic 3=Jewish 4=Other 5=None 9=DK/NA What was the last grade in school you completed? 1=Not a H.S. grad 2=H.S. grad 3=Some college (trade or business) 4=college grad and beyond 9=ref Are you white, black, or some other race? 1=White 2=Black 3=Other 4=ref APPENDIX B CBS NEWS/yew YORK TIMES POLL National Survey Methodology 1989 APPENDIX B one NEWS/THE new YORK TIMES POLL National Survey Methodology 1989 Sanpl ing Procedure: The individuals conprising the sanple in the CBS News/yew York Tia nation-wide telephone surveys are selected through a variation of random-digit dialing.[ll Prinary sanpling units (”W-are blocks of 100 telephone numbers, identical through the eighth digit. (Area Code-Exchange-XX_). The PSU's are stratified by geographic region. area code and.sise of place.[2] The first: stage of selection is systenatic, with a uniforn rate Eros a couplets list of all existing area codes and exchanges in the conterninous United States. The last two digits are randonly assigned within each 980. The nunber is called for a screening interview. If it is a working residential nunber. the P80 (i.e.. the block of 100 telephone nunbcrs) is retained in the sanple. Three phone nunbers are called within a PSU for each survey. The actual telephone nusbers for each survey are selected in snall groups (replicates) that pernit varying sanple size while still retaining geographic distribution. The initial screening of a P80 takes advantage of the telephone conpany's practice of assigning working telephone nunbers in blocks of 100's or 1000's before starting a new block. The design is self-weighting, as the likelihood of an .eight digit group or “block“ being selected as a P80 is exactly proportional to the nunbor of working nunbsrs in the block. Respondents are selected within each household using a netbod developed by Kish, and nodifiod by Backstreet and fiursh.[3] The probability of selection depends on the sire of household. The procedure provides an unbiased selection of households of fewer than five adults. This respondent selection controls for sex and the relative age conposition of the household. Nationwide surveys condgctad during the 1916 and 1980 presidential (prinary caspaign include an ovarsanpling of Republicans in order to insure that the nunber of Republicans. in the sanple‘ -would. be large enough for reliable analysis. To do this, additional randon replicates beyond those used to obtain the original sanple were included. In these additional replicates. only Republican respondents were interviewed. The sanple is weighted to reflect: the actual probability of selection, including this oversampling. 125 126 Ueighting Procedure: The weight assigned to each respondent is the inverse of the probability of selection. modified by a series of ratio estimates. The.first.stage estimate is essentially a non-interview adjustment within geographic region. For each region, an adjust-ent is made to approximate the number of adults in.that region. In the next stage, race by sex are the . ratio estimate characteristics, and the final stage is age by education. 9n occasion. because of a small number of sample cases in some cells at the final stage. some educational groups are collapsed within age categories. In some surveys respondents were also given a weight reflecting their probability of ‘voting. The estimate of their likelihood of voting was based on prior voting participation. the individual's stated intention of voting in the fall and a correction for overreporting registration. In 1980 and 1982 the correction for overreporting of registration was based on race. age; and income variables and was devised from ‘the validation studies of Michael Traugott and John Katosh.[4] In-lSdé this correction was based on state registration laws and past voting and registration behavior. This estimate was multiplied by the selection weight to obtain the probable electorate weight of 0. Either the probable electorate weight or the selection weight must be used in any analysis. The weights computed in the manner described are the dimension of the adult population of the United States. For convenience, all weights are reduced by a constant so that the resulting sum of the weights is the sample size. A weight is assigned to each sample record and RUST be used for all tabulations. O l [l] The sampling procedure used for these surveys was designed by CBS News in 1970. The optimisation of the procedure was done by Joseph Haksberg. See Joseph Uaksberg, ”Sampling Methods for Random Digit Dialing”, Journal of the American Statistical Association (vol. 73), March 1978. pp. ao-ss [2] Sise of place strata are determined by using the geographic coordinates of the places where the central offices are located. These coordinates are assigned by the phone company to every exchange. All exchanges 'in the same city have the same coordinates. Large and small city strata.are defined by counting the number of exchanges per pair of coordinates. Suburbs are exchanges with 15 siles of large cities. The last stratum includes the remaining exchanges. ‘The number of working residential phone numbers per exchange varies considerably by stratum. [3] Charles U. lackstrom.and Gerald D. Uursh, Survey Beseargh. Northwestern University Press, 1963, pp. 52-58 _ [é] flichael U. Traugott and John P. Katosh, “Response validity in Surveys of voting Behavior”, Public Qpinion gggrterly (Vol. £3).- 9‘11 1979: pp. 359-371. Also John P. Katosh and Hichael fl. Traugott, “The Consequences of validated and Self-Reported Voting Measures", ‘gggg (Vol. 45),‘Uinter 1981, pp. 519-535. And in discussion with the authors. HICHIGQN STRT UNIV. LIBRQR \\\\\\\\\\\\\\\\\\\\\\\\i\\\\\\\\\\\\\\§‘\§\\\\\\\\\ \\\\\ 312930090 9309