Joint Center for Housing Studies Harvard University, RL Wiener, K Gross, S Block

Tags: Coalition, Debtors, Logistic Regression, financial literacy, Law and Human Behavior, Journal of Applied Social Psychology, consumers, respondents, attitudes, Sexual Harassment, Psychology, financial management, Amos Tversky, University of Chicago Press, Sexual Conduct, Wiener, Daniel Kahneman, Richard L., Harvard University Press, Journal of Social Issues, Richard, Cognitive Psychology, Student Knowledge of Personal Finance, Journal of Financial Education, Journal of Drug Education, Predicting Debtor, Journal of Consumer Affairs, Linda E. Hurt, Saint Louis University Law Journal, California Western Law Review, Social Cognition, Financial Literacy Education, minimum payments, Coalition for Consumer Bankruptcy Debtor Education, bankruptcy in the United States, human behavior, consumer debt, experimental group, consumer, empirical data, Financial Literacy The Coalition, financial literacy program, bankruptcy filers, debtor education, psychological theory, psychological issues, self-reported, positive attitude, attitude, Journal of Applied Psychology, logistic regression analysis, savings account, credit card, credit cards, credit reports, Workplace Financial Education
Content: Joint Center for Housing Studies Harvard University High and Low SES Debtors: The Use of Psychological Measures to Determine Differences Richard L. Wiener, Karen Gross, Susan Block-Lieb, and Corinne Baron-Donovan BABC 04-2 February 2004 This paper was produced for Building Assets, Building Credit: A Symposium on Improving Financial Services in Low-Income Communities, held at Harvard University on November 18-19, 2003. Richard L. Wiener is a professor at the University of Nebraska at Lincoln. Karen Gross is a professor at the New York Law School. Susan Block-Lieb is a professor at Fordham University - College of Law. Corinne BaronDonovan is a Ph.D. Candidate at the Graduate Center of the City University of New York. © by Richard L. Wiener, Karen Gross, Susan Block-Lieb, and Corinne Baron-Donovan. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is given to the source. Any opinions expressed are those of the author and not those of the Joint Center for Housing Studies of Harvard University, or of any of the persons or organizations providing support to the Joint Center for Housing Studies. This work was supported by grants from the National Endowment for Financial Education (NEFE) ®, the National Conference of Bankruptcy Judges, the American College of Bankruptcy to the Coalition for Consumer Bankruptcy Debtor Education, and the Ford Foundation. Correspondence concerning this article should be sent to Richard L. Wiener at the following address: Department of Psychology, 238 Burnett Hall, University of Nebraska at Lincoln, Lincoln NE 68588. Electronic mail may be sent to [email protected]
Abstract This paper used a psycho legal methodology, which combined empirical research with legal scholarship to study debtors in bankruptcy comparing them to a sample of non-debtors on knowledge, mood, attitude, and behavioral measures of spending and saving. Logistic regressions showed patterns of behavior that differentiate debtors from non-debtors. Debtors, compared to non-debtors, were more likely to have unpaid bills and less likely to deposit money into saving or checking accounts. Cluster and profile analyses found differences between types of debtors producing high and low SES profiles with training and educational implications. Most importantly, higher SES debtors' expenditures for unneeded but desired products were largely determined by their attitudes toward irresponsible spending, in conjunction with normative considerations of significant others. On the other hand, the spending and saving actions of low SES debtors relied on positive attitudes toward unnecessary purchasing and their perceptions of their own self-control, but not on normative considerations. 1
Social Analytic Jurisprudence and Bankruptcy In a series of papers, we developed a model of psycholegal analysis that mixes experimental psychology's scientific empiricism with law's humanistic approach to form an interdisciplinary methodology that relies on social scientific theory and method to explore issues of law and public policy (Wiener 1993; Wiener 1995; Wiener, Hurt, Russell, Mannen, and Gasper 1997; Wiener and Hurt 1997; Wiener, Watts, Goldkamp and Gasper 1995, Wiener and Hurt 1999; Wiener and Hurt 2000; Wiener, Hackney, Kadela, Rauch, Seib, Warren, L. and Hurt 2002) . We termed the model social analytic jurisprudence because it describes social and psychological reality with an eye toward influencing policy formation and dispute resolution. Our approach provides descriptive and causal knowledge to the legislative and public policy processes in the legislatures, courts, and other law making bodies. The purposes of this chapter are to introduce social analytic jurisprudence as a tool to scholars interested in studying how to build assets and credit in low-income populations and to offer an application of social analytic jurisprudence to the problem of understanding the psychology of consumers who seek bankruptcy relief for their financial difficulties. This application of our hybrid method can inform policy makers in the public and private sectors about existing barriers to wealth accumulation in populations that exist at the margins of our economic marketplace. We begin with a discussion of the major tenets of our approach, which has as its overarching goal the empirical test of the assumptions that law and policy makers readily accept as the underpinnings of existing rules, regulations, and dispute resolution processes. Social analytic jurisprudence makes three important assumptions about the role of psychology in law and public policy. First, psychology as it relates to law is an empirical science (Wiener 1993). There exists a collection of psycholegal scientists who share a common commitment to a set of scientific beliefs and values and who agree upon the particular problems and solutions that are relevant to issues of policy and law. These scientists constitute a scientific research paradigm (Lachman, Lachman, and Butterfield 1979; Kuhn, 1962) and it is the work product of these scientists that ought to influence public policy. Second, the psycholegal scholar can best contribute to public policy and legal debate with the tested results of Psychological Research. The legitimacy with which scholars can apply psychological knowledge to issues of law and policy directly relates to the quality of the psychological facts that are tested and accumulated. Legal psychologists ought to base their conclusions more on data and less on the 2
shared social and political ideologies of the members of the scientific paradigm. When social scientists rest their arguments on a value consensus rather than a reliable database, they risk the danger of forming a political platform agreed to by only some members of the paradigm and which deviates from the accumulation of knowledge in the social sciences. Finally, the proper role of the scientific psychologist in policy debate and conflict resolution is that of consultant rather than advocate (Wrightsman 1991). Advocates use the product of research to support the side of a policy debate that agrees with their own political or ideological value position. They search psychological knowledge for research results that confirm a chosen position. Of course, it would be naпve to suggest that any psycholegal scholars conduct value free research. It is unquestionably the case that one's value positions influence the choice of one's research questions, methods, data interpretations, and reported conclusions. However, consultants adopt a disconfirming point of view and search research results for evidence that can refute all plausible rival explanations including and especially their own. It is in the best interest of research scientists to expose their own disconfirming data because the peer review process will likely uncover this evidence if the researchers fail to do so. In addition, those psycholegal scholars who choose to communicate psychological knowledge in the form of expert testimony to inform policy debates also face the adversarial approach that is at the heart of our judicial system or the penetrating light of public scrutiny that accompanies open debate in legislatures and other policy formation bodies. The adversarial system and the system of the open market place of ideas supplements the already rigorous peer review process for psycholegal scholars who communicate psychological knowledge with the intention of informing public policy deliberations. Step 1: Finding the Law and the Policy Social analytic jurisprudence begins with an analysis of legal doctrine or policy rules and procedures, carefully looking for assumptions about human behavior (Wiener 1993; Wiener et al. 1997; 1999; 2002). To be useful for adjudicative, legislative, or executive decision-making, psycholegal research must address questions at the heart of the normative model inherent in the policy position. The language and concepts used in the investigation should track closely the language and concepts that make up the law. The work of psychological researchers is most useful in policy analysis if the investigators understand the existing rules and procedures and 3
pose research questions in a way that bears directly on the policy issues that give rise to the questions. While a detailed review of bankruptcy law is beyond the scope of this chapter, the following paragraphs outline the major issues that pertain to the empirical efforts of social analytic jurisprudence described in the current manuscript. Consider consumers who find themselves in chronic and unrelenting states of financial debt. Federal bankruptcy law carefully outlines two procedures by which defaulting consumers may obtain a fresh financial start. Under current law, consumers can choose to file under chapter 7 of the Bankruptcy Code, enabling a trustee to sell a debtor's unencumbered non-exempt assets and distribute the proceeds to priority and general unsecured creditors. Moreover, successful conclusion of this process results in a discharge, freeing the debtor from personal liability on many remaining pre-bankruptcy claims. The Code protects a debtor's wages from the reaches of creditors as soon as a chapter 7 bankruptcy case commences. Alternatively, individuals can file under chapter 13 and hold on to all of their property, whether or not it is exempt, and instead repay their creditors through confirmation of a three to five-year long repayment plan. Repayment plans are funded with the debtors' disposable income that is not needed to support themselves and their dependents for the repayment of secured and unsecured creditors' claims. Debtors only receive a discharge in chapter 13 bankruptcy cases after they have completed payments under their debt repayment plans (Block-Lieb, 2004). As a practical matter, the protection afforded by a discharge in bankruptcy is fragile. First, in the event of a post-petition reaffirmation of otherwise dischargeable debt, the creditor can pursue the debtor to enforce the debtor's obligation. Second, the law allows a debtor to borrow and start the cycle again immediately after securing a discharge. Some segments of the finance industry view discharged chapter 7 debtors as ideal borrowers because they are unable to file for bankruptcy protection for another six years. In response to some of the apparent ills in the current bankruptcy system, a movement emerged at the end of the last decade to reform the current provisions with a modification of the existing statute. Despite the failure of proposed bankruptcy legislation to pass in the 107th Congress, the Bankruptcy Abuse Prevention and consumer protection Act of 2003 (the 2003 Act) was reintroduced in the 108th Congress, and passed the House on March 19, 2003 by a veto proof majority. Like its predecessor, the 2003 Act would do away with a debtor's unfettered choice of filing under chapter 7 or chapter 13 (American Bankruptcy Institute, 2003). The 2003 4
Act would replace debtors' choices with a means test intended to ensure that debtors repay their creditors all that they can afford (HR 975 § ______ ). While many in the consumer finance industry endorse the means test, dissenting legislators attempted to balance several proposals that they believe favor creditors with additional disclosure guarantees that would protect potential debtors from excessive credit card obligations and ensure that they understand the nature of purchasing with a credit card rather than cash. The 2003 Act would amend the Truth-in-Lending Act (15 U.S.C. § 1637, 1994) to require that billing statements disclose how long it would take consumers to pay off hypothetical balances by making only minimum monthly payments (Price 2002). In addition, credit card issuers would be required to maintain a toll-free, continual access telephone service so that debtors could find out at any time how long it would take to pay off their balances if they make only minimum monthly payments (H.R. 975, § 1301(a)). Card issuers would also be required to provide clear and conspicuous information about the mechanics and duration of introductory or "teaser" interest rates (H.R. 975, § 1303) (Price 2002). In short, these amendments, all contained in Title XIII of the 2003 Act, try to curb unwise use of credit. In addition pending legislation would mandate processes and procedures to reeducate consumers about wise spending, saving, and credit use. Although other countries provide credit counseling and financial literacy training to consumers while they are in the bankruptcy system, current U.S. consumer bankruptcy law does not do so in any systematic fashion. However, the 2003 Act would require mandatory financial literacy training for all individual debtors seeking protection under either chapter 7 or chapter 13 (HR 975 § ______ ). Step 2 and Step 3: Applying Psychology to the Problems of Chronic Debtors Returning to our model of policy investigation, the second stage of social analytic jurisprudence consists of a careful psychological analysis of the law, policy, and administrative rules. Statutory, administrative, and common law doctrines are comprised of legal tests and rules that direct the decision maker to apply social facts to specific issues of concern. The tests and rules sometimes appear in language that invites a social scientific and perhaps a psychological investigation. After identifying the empirical issues in the language that makes up the tests and rules, the psycholegal scholar reviews the psychological literature to identify theories, research results, and methodologies that are most suitable to answer legal and policy questions. It is at this 5
stage in the analysis that the investigation takes on a psychological flavor. While it may be helpful to apply the powerful research methodologies of the social sciences to test intuitive answers to empirical questions, that process by itself falls short of realizing the interdisciplinary promise that psychology makes to the interaction of law, public policy, and social science. In this sense, the third stage of social analytic jurisprudence follows directly from the first two. A psycholegal analysis of the knowledge base related to any area of law or public policy is likely to point out gaps in our understanding of the psychological and social reality that lie beneath legal and policy issues. There may be a dearth of information on topics that speak directly to the issues in the debate or more likely, there will be relevant studies that researchers conducted without benefit of a thorough legal or policy analysis, so that the results may not be directly relevant to the substantive issues. In the third stage of social analytic jurisprudence, psycholegal scholars conduct research that tests the psychological models that they applied to answer the empirical issues identified in stages 1 and 2. It is at this point that the powerful methodological and statistical tools of the behavioral scientist comes into play for the purpose of gathering data that speak directly to the issues in the legal or policy debate. Social scientists can bring their work to bear upon legal doctrine and policy issues in three different ways (Wiener et al 1999; 2002). Social science can take on an adjustment or assessment role, an implementation role, or an evaluative role. Law presents normative theories of behavior as depicted in the language of statute, court opinions, and administrative rules and regulations. Psychological research can assess the fit between the law's normative model and people's conduct to assess the correspondence of the law with everyday behavior. In other words, research fulfilling an adjustment or assessment function can help determine whether people behave as the law presumes that they do. For example, Wiener et al. (2002) studied the way in which male and female workers naturally think about gender discrimination without the rule of law and found patterns of judgment that deviate in predictable ways from the principles of hostile work environment harassment codified in federal law. Therefore, our assessment of how workers think about gender discrimination led to some adjustments to the existing tests so that they better fit the judgment processes of men and woman at work. In a separate line of work, Wiener and colleagues (Wiener, Pritchard, and Weston 1995; Wiener, 1998; Wiener, Hurt, Thomas, Sadler, Bauer, and Sargent 1998; Wiener 2003; Wiener, Rogers, Winter, Hurt, Hackney, Kadela, Seib, Rauch, Warren and Morasco in press) also applied social analytic 6
jurisprudence to jury sentencing in capital murder cases to understand how juries reach penalty decisions (prison vs. the death penalty) within the constraints of the normative rules outlined in jury instructions. Here, the goal of the psycholegal analysis was to assess the fit between the cognitive system of jurors and the penalty phase instructions as offered in the jury charges. Ultimately, social analytic jurisprudence produced some viable suggestions for altering the instructions to improve the enactment of the existing law. A second function of psycholegal research is to assist with the implementation of current law. It is perhaps in this role that research psychology has had its most direct impact. Examples include the study of police discretion, jury decision making, judicial decision-making, parole decision making, and eyewitness identification. (See Ogloff 2001 for a current review of each of these areas). It is not a coincidence that each of these focuses on the decision-making and/or judgment capability of significant actors in the legal and policymaking process because the implementation of law depends upon actors making decisions under the constraint of substantive and procedural rules. Acting in its implementation role, psychology measures against statutory, administrative, and judicial principles the effectiveness of the people who execute the law. The purpose of this research is not only to improve the quality of legal process but also to apply our knowledge of human behavior to understand how people function as legal actors and policy makers. Finally, in its evaluative function social science measures the impact of the law on the everyday lives of citizens. Psychological research can either directly test the effects of legislation or judicial holdings on the behavior of those citizens targeted by the law or it can examine how different formulations of law influence the social and cognitive behavior of those citizens. For example, some of our prior work in policy analysis applied social analytic jurisprudence to policy problems in gender discrimination (i.e., sexual harassment) focusing on the influence of federal law on the behavior in workplaces (Wiener et al. 1997; Wiener and Hurt 2000; Wiener et al 2002; Wiener, Winter, Rogers, Arnot in press). This work specifically evaluated the effectiveness of several legal standards (i.e., reasonable person, reasonable woman, reasonable victim, and rational woman) in guiding employees' own perceptions of social sexual conduct at work (Wiener et al., 2000; 2002; in press). The current chapter lays out a program of research that examines current issues in bankruptcy policy from the perspective of social analytic jurisprudence. Our general approach is 7
to identify issues with empirical assumptions about human behavior and to test directly those assumptions with psychological theory and most importantly with empirical data. We firmly believe that policy makers need to understand the interaction between human behavior and bankruptcy law if they are to assist low-income families to gain the capital that they need to realize the promise that the American marketplace makes to all consumers. That is, all consumers can enjoy a reasonable level of prosperity if they prepare themselves to spend and save wisely and to make careful use of credit products. Social Analytic Jurisprudence, Consumer Bankruptcy Law, and Financial Literacy Both legal and economic literatures discuss the social and political antecedents, correlates, and outcomes of consumer bankruptcy in the United States, yet no existing work considers these issues from a psychological perspective. Nonetheless, bankruptcy raises psychological issues in the sense that people are responsible for their own finances and they behave in ways that either prevent them from getting into trouble or that force them to rely on bankruptcy for debt relief. Furthermore, if they make the wrong choices about how to use their money people may experience directly the anguishes that filing for bankruptcy can create in all of the realms of their personal lives. The Coalition for Consumer Bankruptcy Debtor Education (the "Coalition") a not-for profit Section 501(c)(3) organization, founded in 1998 to study closely the bankruptcy process and intervene to assist consumers to reenter the marketplace as skillful users of money, set out to address several issues regarding the psychology and law of consumer debt and bankruptcy. Because there is no prior systematic attempt at understanding debt and bankruptcy from a psycholegal perspective, the Coalition developed some early descriptive questions to lay the groundwork for a more advanced study of the problem. These include: 1) What distinguishes bankrupt consumers from others who do not get into trouble with debt? 2) Are there categories of bankruptcy filers who share separate patterns of psychosocial characteristics? 3) If there are categories of bankruptcy filers, what types of knowledge, attitudes, and behaviors are characteristic of the categories? 4) Are the current bankruptcy policies commensurate with the characteristics of people who need relief? 5) How do debtor consumers make decisions about spending, saving, and using credit? In what ways do they deviate from a rational choice model? 6) Is the system of credit rating for economic risk commensurate with the way in which consumers make their economic choices? Are there 8
individual differences in consumer risk that determine risk better than simple credit scores? 7) What is financial literacy and how can we teach it to bankrupt consumers so that they can become skillful users of money? Financial Literacy The Coalition started with the last question and launched a financial literacy program to train bankruptcy debtors. The Coalition's financial literacy training program is, of course, not the first attempt at studying financial literacy; in fact, many colleges and universities offer financial management courses, and some researchers have collected and published performance data from these classes. There are studies that document the positive relationship between student course performance and student ability (Johnson, Joyce, and Sen 2002) or test the relationship between course performance and effort (Didia and Hasnat, 1998; Nofsinger and Petry 1999). One enterprising group (Bixler and Squires 1998) brought university students to visit New York's financial district to learn about financial management. Others (Tennyson and Nguyen 2001) found that high school students did score higher on tests of financial literacy in states that required specific financial literacy course work as compared to states with only general mandates or no mandates at all. Other published financial literacy program evaluations have shown success within specific domains of financial education including employer based programs (Bernheim and Garrett 1996) and retirement seminars (Bayer, Bernheim, and Scholz 1996). While there are numerous examples of financial literacy programs,1 there are relatively few published evaluations of these programs (Braunstein and Welch 2002). Unpublished studies show mixed results, some successful and others unsuccessful, in areas as diverse as homebuyer training (Hirad and Zorn 2001), workplace financial education (Kim, Kratzer, and Leech 2001), saving effectiveness (America Saves 2002), high school training (Boyce and Danes 1998; Jump$tart Coalition 2002), and general financial training (Bradley, Hirad, Perry, and Zorn 2001). The Coalition found the resulting data wanting not only because the results were confusing and inconsistent, but also because there were no studies that used experimental or quasi-experimental designs to test directly the effects of financial education on financial literacy 1 One review study commissioned by Fannie Mae, Vitt, Anderson, Kent, Lyter, Siegenthaler, and Ward (2000) found providers as diverse as private employers, the military, state cooperative Extension Services, Community Colleges, faith-based groups, and community organizations offered 90 financial literacy education programs. 9
acquisition in consumer debtors. A sub-committee of the Coalition's Board of Directors (i.e., law professors, bankers, debt counselors, money therapists, and adult education specialists) carefully examined existing financial literacy materials and the literature on the psychology of money (c.f., Furnham and Argyle 1998). They participated in lengthy discussions to determine what a financially literate consumer needs to know about credit, spending, and saving money and produced a 38 page long Financial Management Guide comprised of seven units written at a 7th to 9th grade reading level. The first three units (i.e., Money Thoughts, Planning and Goal Setting, and Creating Financial Plans) direct consumers to examine their own attitudes and knowledge about money and finances in order to teach the importance of developing and recording one's own financial goals, carefully separating consumer needs from consumer wants. The first three units culminate in a two-page tracking sheet that helps consumers create their own spending plans. The second half of the Guide (units 4 -6 Sales, Ads, Scams, and Other Traps; Credit Usage; and The Future) focuses on the types of errors that consumers make that lead to unwise spending and poor use of credit. The curriculum ends with unit 7, Resources, which stresses that financial literacy education is an ongoing process, and provides a variety of additional resources (websites, agencies, and books) about financial management. The Coalition-trained instructors (Baron-Donovan, Wiener, Gross, Block-Lieb 2003) administered the curriculum plan as part of a demonstration project, "Making Sense of Cents," in Manhattan, Brooklyn and Long Island between April 2002 and May 2003. Under the Pilot Project, more than 600 individual debtors residing in New York received a free three-hour financial management class taught by one of the more than 125 debtor educators trained by the Coalition in one of its sponsored teacher training programs. The overall goals of the pilot project were to convey basic vocabulary and knowledge about money, credit, and managing finances; alter attitudes and social norms about spending and saving money; and influence debtors' actual financial behavior (spending, saving, budgeting, and using credit) in a positive direction. The study, reported in Wiener, Baron-Donovan, Gross, and Block-Lieb (under review), used a quasi-experimental design consisting of one experimental group and two comparison groups (Cook and Campbell 1979). All respondents in each group completed a pretest questionnaire and an identical posttest survey. Individuals in the experimental group completed a three hour-long financial management class (trained debtors). Participants in one comparison group were debtors who did not receive any debtor education (untrained debtors), while those in 10
a second comparison group were representatives from the general population who had not filed for bankruptcy (non-debtors). Participants completed the questionnaires between April 2002 and May 2003 with an approximate 3-month delay between the pretest and posttest administration. Wiener et. al (under review) found the curriculum effective in a number of areas. First, with regard to the knowledge about wise and unwise spending, saving, and credit neither the untrained debtors nor the non-debtors gained from the pretest to the posttest, but the debtors completing the Coalition's training course did show significant increases in their knowledge scores, catching up to the other two groups. At the onset of the training session, the untrained debtors started out significantly lower than the other two groups but caught up to them at the end of the training sessions. Wiener et al (under review) also found positive results concerning attitudes toward irresponsible spending. After controlling pretest differences, the trained debtors showed more negative attitudes toward unnecessary spending relative to the other two groups and less intention to buy than the non-debtors (the untrained debtors were not significantly different from the non-debtors). Finally, with regard to self-reported behavioral outcomes due to the financial literacy course, pretest to posttest changes showed that compared to the untrained debtors and the non-debtors, the trained debtors increased in the percent of respondents who failed to report unpaid bills and in the percent that put together household budgets. Further, after controlling for pretest differences, the trained debtors showed fewer credit card purchases than did the non-debtors and they borrowed less from predatory lenders (e.g., payday loans, rent to buy, and pawnshops) than did the non-debtors. The untrained debtors did not show differences from the non-debtors on either of these outcomes. Finally, trained debtors reported holding fewer credit cards and those who obtained credit cards showed a lower monthly balance than either the untrained debtor or non-debtor groups. Characteristics of Debtors The purpose of the current data report is to begin to address the first three descriptive questions raised above: 1) What distinguishes bankrupt consumers from others who do not get into trouble with debt? 2) Are there categories of bankruptcy filers who share separate patterns of psychosocial characteristics? 3) If there are distinct categories of bankruptcy filers, what types of knowledge, attitudes, and behaviors are characteristic of the categories? To address these issues we rely on the pretest data collected in the "Making Sense of Cents" pilot study 11
(Wiener et al, under review). Participants in the debtor group were bankruptcy filers in the Eastern and Southern Districts of New York and those in the non-debtor group were undergraduate and graduate students at a large, northeastern, urban university as well as samples of non-professional staff members at two northeastern law schools. Before presenting the results of the pretest data, we reconsider the bankruptcy process in light of our psychological model of consumer debt. In the current manuscript, we examine a sample of individual consumers who have filed for bankruptcy in order to describe the knowledge, moods, attitudes, and economic behavior that sets them apart from each other and others who stay out of the bankruptcy system. Adopting a tripartite approach that is common in applied social psychology, we assumed that knowledge and attitudes are antecedents of economic decision-making and that spending and saving behaviors are the outcomes of those decisions. We measured the financial knowledge of participants with the multiple choice test that Wiener et al. (under review) used to test the effectiveness of the Coalition's financial literacy training program and we used Wiener et al.'s self-report measure of spending and saving to examine consumer behaviors among debtors and non-debtors. Our goal was to study differences in these constructs between debtors and non-debtors and to determine if different types of debtors displayed distinguishing patterns of psychosocial measures. With regard to attitudes about irresponsible spending (buying goods that consumers want, but do not need and cannot afford) the evaluation team borrowed Ajzen's theory of planned behavior (Ajzen 1988). According to the model, behavior is a function of the intention to behave moderated by the actor's perceptions of control. With perfectly perceived control, behavior reduces to the intention to act but without perfect control, the actor's level of perceived control moderates the intention to act. Intentions, in turn, result from the interaction of one's own attitudes toward the action and the attitudes of others (social norms) toward the action. The model measures the actor's attitudes and social norms with Likert type, self-report rating scales. Numerous applied studies have used the model to examine a variety of different types of behaviors. A few of the most recent applications of the model include explanations of healthy eating behavior (Conner, Povey, Sparks, James, Shepherd 2003), ecologically sensitive behaviors such as driving and recycling (Bamberg and Schmidt 2003; Chu and Chui 2003; Kaiser and Gutscher 2003), substance abuse (Kuther and Higgins-D'Alessandro 2003), completing high school (Davis, Ajzen, and Saunders 2003), and shopping on the web (Gentry 12
and Calantone 2002). We extend the model to understand consumers' conscious plans to purchase unneeded but desired products and theorized that attitudes toward buying such products and subjective norms, (i.e., the attitudes of significant others regarding these purchases) would each interact with perceived control to produce intentions to buy (or not to buy). We wondered if this attitude process would be different between debtors and non-debtors and among different types of debtors. Research Methodology Design Overview The current study is a secondary analysis of data collected in Wiener et al. (under review), which used a quasi-experimental design to test the effectiveness of the Coalition's financial literacy program. The design included one experimental group and two comparison groups. All respondents in each group completed a pretest questionnaire and an identical posttest survey. Individuals in the experimental group completed a three hour-long financial management class (trained debtors). Participants in one comparison group were debtors who did not receive debtor education (untrained debtors), while those in a second comparison group were representatives from the general population who had not filed for bankruptcy (non-debtors). Beginning in April of 2002, the Coalition identified from a government public record database all individuals who had recently filed for chapter 7 or 13 bankruptcies in the Eastern District of New York (Manhattan, Staten Island, Queens, Brooklyn and Long Island). After removing duplicate addresses, the research team sent information describing the debtor education project, information about available financial literacy classes, and a pretest questionnaire to individuals living at each of the addresses in the database. The mass mailing included an addressed, stamped return envelope so that the respondents could mail the completed questionnaire directly to the researchers. In addition, several students working for the Coalition passed out and posted flyers at local courthouses during Section 341 meetings in which trustees and creditors were questioned about their assets, liabilities, and other issues. These student assistants announced the project and asked interested debtors to contact the Coalition to participate in the study. Finally, the Coalition recruited a set of intermediaries (i.e., The Legal Aid Society, local union offices, Chapter 13 Trustees - court appointed case administrators, and 13
other agencies who assist individuals filing bankruptcy petitions) to inform individual debtors about the financial literacy classes and the study. Participants in the non-debtor group include undergraduate and graduate students at a large, northeastern, urban university as well as samples of non-professional staff members at two northeastern law schools. Participants in all groups received $10 upon completion of both questionnaires (separated by 3 months). All participants signed an informed consent sheet at the pretest and were treated in accordance to the American Psychological Association's guidelines for ethical treatment of human research participants. The final number of participants who completed a pretest was 417 trained debtors, 304 untrained debtors, and 146 non-debtors (N = 867). Wiener et al. (under review) analyzed and reported posttest differences between the three groups, as a function of training controlling for initial pretest differences. These data are not discussed further in this report. The current report considers differences between debtors and non-debtors, as well as types of debtors using only he pretest data. Statistical summaries of the demographic factors for the final sample of respondents showed that the project included a diverse sample of participants. Table 1 displays the demographic breakdown for non-debtors and debtors who answered the demographic questions (few left the questions blank). In both groups there were many more female than male respondents and the debtors were older than the non-debtors (in part, because many of the nondebtors were drawn from a college sample). White eastern Europeans did not constitute a majority in either sample; however, the debtor group contained a larger percentage of Black respondents than did the non-debtor group. The majority of individuals in both groups reported at least a high school education and in many cases a college education, although, the non-debtors were slightly better educated than the debtors. Married respondents were in the minority in both groups and the modal household incomes in both groups at the time of the survey and 12 months prior to the study were near $30,000. In both groups, the majority of respondents were employed at the time of the survey but a substantial number of respondents were unemployed. English was the primary language for most respondents. Finally, the sample of non-debtors included a substantially higher number of full or part time students than did the sample of debtors and they were less likely to report that they were responsible for their own finances. 14
Materials The pretest consisted of three basic sections (counterbalanced across participants within groups to control for order effects): financial knowledge, attitudes towards buying and spending, and self-reports of spending, credit, and saving.2 Knowledge Survey This section of the questionnaire consisted of 12 multiple-choice questions, each with one correct answer. The items asked about credit, minimum payments, wise buying/spending habits, credit reports, credit scoring, payday loans, rent-to-own programs, and rights and responsibilities following bankruptcy discharge. For example, one question read: If you make only the minimum monthly payment on your credit card bill each month, the amount of interest you will pay will be (correct answer is a): a. more than if you had made larger (greater than minimum) payments each month. b. less than if you had made larger (greater than minimum) payments each month. c. the same no matter what the size of your monthly payment. d. zero, because credit cards don't charge interest. The score for the knowledge section was the number and/or percent of items answered correctly. Attitude Survey The section on attitudes toward unnecessary spending (i.e., buying wanted but unneeded goods) consists of four hypothetical scenarios in which individuals decided whether they would make a purchase. The four scenarios consist of 1) buying a new car once a loan on the current car was paid, 2) going on a cruise with a friend when money is tight, 3) furnishing an entire apartment with new furniture for a monthly rental payment, and 4) purchasing a new washer/dryer when the current machine was still functional. Respondents answered each of the questions on nine-point, Likert-type response scales with verbal anchors. Two questions asked about respondents' valence/feelings toward making the purchase (1= displeased to 9 = pleased) and the importance of those feelings (1 = unimportant to 9 = important). The next two questions asked individuals to rate the valence/feelings that their friends and/or family would feel about the 2 The questionnaire was developed and pre-tested in a pilot study of non-professional staff and Undergraduate students to develop questions that were comprehensible by the average person. 15
purchase (1= displeased to 9 = pleased) and how important that attitude is to the individuals (1 = unimportant to 9 = important). The final two questions asked about the amount of control the individuals perceived they had over the purchase (1 = no control to 9 = complete control) and finally their likelihood to make the purchase (1 = unlikely to 9 = very likely). As described above, we scored the attitudes according to the model of reasoned action (Ajzen, 1991), first recoding the respondents' own valence and perceived valence of others so that the most displeased equaled -4, neither pleased or displeased equaled 0, and most pleased equaled 4. Next, we calculated individual scenario attitude scores by multiplying the valence by importance of that valence for each vignette. The mean attitude score was the average of these products calculated across the four scenarios. The most positive attitude in favor of purchasing could equal 36 and the most negative opposed to purchasing, -36. A subjective norm score resulted from multiplying the perceived valence assigned to others by the importance of that valence. Once again, a mean normative score resulted from an average calculated across the four scenarios with the most positive normative view of buying equal to 36 and the most negative, 36). Averaging the control scores across the four vignettes produced a mean score for control (1 = least control and 9 most control) and averaging individual likelihood to buy scores across the four scenarios (1 = least likely to buy and 9 most likely to buy) yielded a mean intention score. Behavioral Self-Report The self-report questions in this section measured respondents' recent spending, saving, and credit use. Most of these questions were in a yes/no format (e.g., Within the past month, did you review your bills to make sure that there were no mistakes in them?). Some required supplying information (e.g., How much did you withdraw from your savings account last month?) and others simply required a checkmark to describe how the behavior occurred (e.g., How did you get your cash over the last month? Please check all methods that apply). The topics covered included use of checking accounts, use of saving accounts, use of credit cards, generic purchases, paying bills, borrowing money, budgeting, and shopping in negative mood states. Procedure All respondents completed the same questionnaire twice. Debtors participating in the financial literacy-training program completed the survey before they participated in the class, 16
while all others completed the pretest at staggered times throughout the one year project. Approximately three months after completing the first questionnaire, participants received a second questionnaire however this report is only concerned with data collected in the pretest survey. The survey package included a cover letter re-describing the research process and reminding participants that upon completion of the second questionnaire they would receive $10 compensation. Results Overview The analyses are presented in 3 sections. First, we compared the debtors to non-debtors on the measures of demographics, knowledge, shopping mood, attitudes, self-reported spending, and self-reported saving behaviors. Second, we conducted a cluster analysis on debtor demographic factors to divide the debtors into homogeneous groups that showed minimal variation within categories, but maximum variation between categories. Third, we compared the resulting debtor groups on the knowledge, shopping mood, attitudes, and self-reported spending and saving behaviors. Comparing Debtors to Non-debtors This analysis was a forward stepwise logistic regression analysis using non-debtor vs. debtor membership as the criterion variable and 10 predictor variables coded as follows to represent potential demographic differences between the groups. 1. Age (0 = younger than 39 and 1 equals 39 or older) 2. Gender (females = 0 and males = 1) 3. Ethnic status (0 = white and 1 = non-white) 4. Education (0 = high school or less and 1 = at least some college) 5. Responsible for financial affairs (0 = someone other than the self and 1 = self) 6. Student status (0= student and 1 = not a student) 7. Employment (0 = employed and 1 = unemployed) 8. Marital status (0 = non-married and 1 = married) 17
9. Current annual household income (0 = 30,000 dollars or more and 1 = less than 30 thousand dollars) 10. Annual household income 12 months earlier (0 = less than 30 thousand dollars and 1 = 30,000 or more) Table 2 displays the results of the logistic regression analyses predicting debtor status coded 1 (non-debtor status was coded 0) for all the analyses that compare non-debtors to debtors. The second column shows the odds ratios that result from a forward stepwise analysis (Wald criterion) along with the Nagelkerke R2 (proportion of explained variability for the model) and a chi square test of the model fit. The third column shows the same equation except that the program forced each demographic factor into the regression in a single step. While the results are similar using both approaches, the forced entry model displays group differences after statistically controlling for all demographic differences. Model 2 shows that debtors (compared to non-debtors) were about 20 times more likely not to be students, over 6 times more likely to be older than 39, 5 times more likely to have income below $30,000, 3 times more likely to manage their own finances without help from others, and almost 3 times more likely to be married. No other demographic differences between the groups were significant. Forced regression using all 10 predictors produced a significant prediction equation that accounted for 59 percent of the variability between groups and, in fact, classified 90 percent of the respondents correctly into debtor and non-debtors groups. (Note that the chance level of accurate classification for non-debtors and debtors equals 78%). These results show that the remaining cognitive, attitudinal, and behavioral factors have 10% of the variability available to explain. Table 3 is concerned with the relationship between debtor status, knowledge of financial management, and participant mood. Column 2 displays the results of the stepwise logistic regression in which shopping to end a bad mood (coded 0 = did not shop to end a bad mood and 1 = did shop to end a bad mood) and knowledge (0 to 100 percent of the questions answered correct) served as predictor variables without taking into account the demographic differences between the samples. While the results show that debtors scored lower on the knowledge scale (the actual weight was equal to - .35) and that they were almost twice as likely to shop to end a bad mood, as were the non-debtors, these significant relationships disappeared in the forced entry model, which controls for the demographic differences between the groups. Without the 18
demographics, the model was significant but only explained 3 percent of the variability, while the full model with demographic factors forced entered first, explained 60% of the variability. Table 4 shows two similar logistic regression equations pertaining to attitude differences in debtors and non-debtors. Debtors were more likely to show positive subjective norms toward buying unneeded but desired goods, but actually showed lower intentions to buy those products. (Remember that higher values on a -36 to 36 scale indicates more positive attitudes and subjective norms to buy, and that higher numbers on the perceived control and intentions to buy scales indicate greater control and greater intentions to buy.) Once again, the two significant relationships between the attitude measures and the debtor group membership dropped out in the forced entry model, which controlled the demographic differences between the non-debtors and debtors. Table 5 shows a different story for the self-reported consumer behavior factors. With the stepwise logistic regression in which only behavioral measures were entered into the regression equation, column 2 shows an R2 value of .25 indicating that 3 variables did indeed account for 25 percent of the available variability in distinguishing between non-debtors and debtors. The logistic regression coded as "0" having paid all bills in the prior 3 months, depositing money into a savings account, not creating a budget, depositing money in a checking account, reviewing bills at payment time, and withdrawing money from a savings account. Coded with the value "1" were having unpaid bills, no deposits into a saving account, not creating a budget, not depositing money into a checking account, not reviewing bills, and not withdrawing money from a savings account. Number of credit cards owned and percentage of loans from predatory lenders were treated as continuous variables (higher numbers indicate more credit cards and more borrowing from predatory lenders.) Column 2 of Table 5 shows that without statistically controlling for the demographic differences between the samples, debtors relative to non-debtors were more likely to have unpaid bills, more likely not to deposit money into a savings account, and surprisingly, more likely to have created a budget. Even after controlling for the demographic differences between groups, the debtors were more than 4 times more likely to have unpaid bills, 3 times more likely to have not deposited any money into a savings account, and twice as likely to have not deposited money into a checking account. After controlling for demographic differences, self-reported budgeting dropped out to be replaced by the unlikelihood of depositing money into a checking account. 19
Comparing Column 3 in Table 5 to Column 3 in Table 2, shows that the behavioral measures forced into the regression equation after the demographic measures explains 7% of the variance. The bottom of Table 5 shows that the behavioral measures as a block added significantly beyond the demographic factors forced into the equation. Column 2 in Table 6 displays the results of the stepwise logistic regression in which credit card measures, including balance paid (1=nothing, 2=minimum, 3=more than minimum, 4=full amount), number of credit card purchases, current credit card balance (in natural log transformed units), and increase in credit card balance (-1 decrease, 0 no change, 1 increase) over the month served as predictor variables, without taking into account the demographic differences between the samples. 3 Results showed that debtors were likely to pay less of their minimum balance (Beta weight = -1.32) but were likely to make fewer credit card purchases than nondebtors were. The 39 percent of the variability in debtor classification explained by these two predictors was significant. Further, after forced entry of the demographic factors, the same credit card measures (also force entered) show the same relationships with debtor classification, explaining about 13 percent of additional variability after the demographic factors. It is clear that a significant activity that discriminates between non-debtors and debtors is the use of credit card shopping. Distinguishing Between Types of Debtors The next analyses set out to look for some natural distinctions between different types of debtors who filed for bankruptcy relief. To evaluate the higher order structure, we performed hierarchical cluster analysis with a binary squared Euclidean distance algorithm using SPSS and followed it with a profile analysis. Participants in these analyses were debtors only and ratings on the 10 demographic characteristics described in the first set of analyses that compared debtors to non-debtors served as differentiators (see Table 7 for a list of demographic factors in the cluster analysis). The profile analysis tested for significant differences among the percentages of demographic characteristics reported by the respondents in the separate membership categories. 3 These analyses were separated from the self-report spending and saving behaviors because it includes only those debtors with 1 or more credit cards. A significant number of debtors (n = 264) reported having no credit cards after filing for bankruptcy. Apparently, these debtors discarded the cards and did not replace them. Adding these analyses to the spending and saving regressions would have lowered unnecessarily the sample sizes of those analyses. The same separation logic was used for the credit card measures differentiating high and low SES debtors (see Table 9). 20
We examined the resulting 2, 3, 4, and 5 groups cluster solutions. We sought a cluster solution that divided the sample into roughly even sized groups that would allow further profile analyses. The 2-cluster solution produced one group with 583 debtors and a second with only 14 respondents. The size of the three, four, and five cluster groups were respectively, (n1 = 562, n2 = 21, and n3 = 14); (n1 = 287, n2 = 275, n3 = 21, and n4 = 14); and (n1 = 280, n2 = 275, n3 = 7, n4 = 21, n5 = 14). The 4 cluster solution divides the sample of debtors into two similarly sized samples, with 35 additional respondents comprising additional clusters. The five cluster solution (like higher order solutions) splinter off respondents from the larger groups, forming more and more groups of smaller numbers of filers. The 4-cluster solution classified 93 percent of the participants for whom demographic information was available into two clusters with only 7 percent (42) without group membership. Table 7 shows contingency tables displaying the distribution of the demographic factors for both membership clusters and accompanying chi-square statistics that test the difference between the two membership groups. Group 2 (N = 275) are filers who have less current income, had lower income 12 months prior to completing the survey, are less likely to be currently employed, have had no college education, are less likely to be married, and are more likely to manage their finances by themselves. On the other factors (sex, ethnicity, student status, and age) there were no statistically significant differences between these groups. In addition, Table 8 shows a forced entry logistic regression analysis distinguishing between the two cluster groups predicting group 2 to be almost 230 times more likely to have low current income (less than $30,000) and over 100 times more likely to have lower income 12 months prior. Here, group 2 members were almost 6 times more likely to be unemployed and to manage their own finances. They were 4 times more likely to be students and 3 times more likely to be males. Each of these variables contributed significantly to the prediction equation after controlling for all other demographic factors force entered into the logistic regression analysis ultimately explaining 92 percent of the variability in the 2 membership groups and accurately predicting membership for 96% of the group 1 (N = 287) and 97% for group 2 (N = 274) (overall accuracy = 96%). Based upon the data displayed in Tables 7 and 8, we labeled the first category (N = 287) as a high social economic status (SES) group of bankruptcy filers and the second category (N=275) as a low SES group of bankruptcy filers. 21
Comparison of the low and high SES Groups on the Behavioral Measures Table 9 reports on the results of three additional logistic regression analyses. In the first analysis (column 2), we force entered knowledge score and shopping in a bad mood into an equation predicting low SES debtor membership. (Note: the demographic factors were not entered into these equations because we were not interested in controlling these factors; rather we were interested in interpreting the differences between the two debtor groups.) The results, significantly explaining 9 percent of the debtor membership category, show that the lower SES debtors demonstrated less financial knowledge than did the higher SES bankruptcy filers. Column 3, reporting a forced logistic regression analysis of spending and saving behavior, shows that low SES debtors were more likely not to deposit any money into their checking accounts in the last month. These debtors are likely to be using other forms of financing to cover their expenses. The regression analyses reported in column 4 shows that low SES debtors pay off less of their monthly balances (Beta weight = -.31) but still have lower overall card balances (Beta weight = -.11).4 The model with credit card predictors explained 7% of the variability in group membership with a significant prediction equation. Attitude Measures Finally, we conducted a General Linear Model (GLM) analysis of the intention to buy unneeded but wanted products following the Ajzen (1988) theory of planned behavior to try and discriminate between low and high SES debtors. The first GLM treated group status (high vs. low SES) as a between subjects factor, dichotomized perceived control of purchasing into high and low levels using a median split, and added two continuous measure covariates (attitudes toward buying and subjective norms toward buying). The model included all the two-way and three-way interactions between perceived control and SES status, and between these two variables and attitudes and subjective norms toward buying. The results showed significant main effects for perceived control, F(1, 549) = 20.35, p < .001, attitudes toward buying, F(1, 549) = 104.39, p < .001, and subjective norms, F(1,549) = 7.39, p < .007. While there no main effect for debtor status, F(1,549) < 1.00 there was a significant interaction between debtor status and positive normative attitudes, F(1,549) = 3.68, p < .056. 4 See footnote #2 above. 22
A follow up analysis examined more closely the relationship between debtor status and the three predictors derived from the theory of planned behavior. It treated debtor status as the only dichotomous factor and the other 3 predictor variables (attitudes toward buying, subjective norms toward buying, and perceived control) as continuous covariates. Results produced main effects for attitudes, F(1,553) = 107.88, p < .001, subjective norms, F(1,553) = 5.96, p < .015, perceived control, F(1,553) = 13.49, p < .001, and debtor status (low vs. high SES status), F(1,553) = 4.49, p < .035. (Low SES debtors, M = 3.08, were more likely to purchase unneeded products than were high SES debtors, M = 3.13.) More importantly, debtor status interacted with positive norms toward buying, F(1,553) = 4.23, p < .04, and perceived control, F(1,553) = 5.02, p < .026. To determine the nature of the interactions between debtor status and the covariates in the model, we regressed intention to buy on the three predictors separately for the Low SES participants and for the High SES participants. Table 10 displays the Beta weights for these final results. The regression weights listed in Table 10 (column 2) show that high SES debtors rely on their attitudes toward buying and the normative views of their family and friends in deciding whether to buy unneeded but desired products. The more their own attitudes and the beliefs of significant others support unnecessary purchasing the more likely the high SES debtors are to buy. For this group unnecessary purchasing is unrelated to perceived control of their intentions to buy. Similarly, the lower status debtors also relied on their attitudes toward purchasing (the more supportive their own attitudes the more likely they are to buy) but showed no significant relationship between norms and intention to buy unneeded products. Instead, for low SES participants, the greater the perceived control the less the debtors intended to buy the unneeded products. Discussion Summary of the Results If we disregard the demographic differences between our sample of consumer debtors and non-debtors, logistic regression analyses suggest that debtors are more likely to shop in a bad mood and score lower on our measure of financial literacy. Debtors rely more on positive (and negative) norms toward buying but they show less intention to buy unneeded products. Debtors have more unpaid bills, are less likely to deposit money into saving accounts, but are more likely 23
to create personal budgets. Finally, they report paying less towards their credit card balances when they used credit cards. These results suggest that there are chronic and systematic differences that set bankruptcy filers apart from their non-debtor counterparts. The law takes no recognition of individual differences in debtor populations and instead assumes, perhaps wrongly, that debtors are different only in degree and not qualitatively from non-debtors. The law assumes all actors in the marketplace regulate themselves in accordance with a rational actor model and that policy need not make special efforts to secure more positive outcomes with debtor populations. Our data suggest that this assumption may not hold up to empirical scrutiny. There are systematic differences between debtors and non-debtors. However, our sample of debtors was not equivalent in basic demographics to the non-debtors. The debtors were less likely to be students, they were older, had lower current income, they were more likely to be married and more likely to manage their own finances. After controlling for these demographic differences between debtors and non-debtors, we were left with only behavioral differences between the two groups. The debtors were more likely to have unpaid bills and less likely to deposit money into saving or checking accounts. In short, debtors find themselves in financial trouble because of their own conduct even after they have filed for chapter 7 or chapter 13 relief. It seems that the behavioral patterns that create financial problems in the first place, persist even after debtors try to find relief for their financial woes. Perhaps a more aggressive approach is needed to help debtors avoid future debt. One approach that the Coalition is pursuing is holding financial literacy training classes for those motivated to learn better ways to spend and save (Wiener et al. under review; BaronDonovan et. al 2003). The goal of the Coalition financial literacy training course is not simply to provide knowledge about financial management, but also to help consumers understand more about the dynamics of unwise spending, saving, and use of credit that contributes to their financial difficulties. Legislation that creates financial literacy requirements would do well to consider the purpose and direction of financial education in light of the recent findings reported in this and other Coalition reports. Our application of the theory of planned reasoning (Ajzen, 1991) shows that it has merit for more than understanding deliberate judgments about healthy eating behavior (Conner, Povey, Sparks, James, and Shepherd 2003), driving and recycling (Bamberg and Schmidt 2003; Chu and Chui 2003; Kaiser 2003), substance abuse (Kuther and Higgins-D'Alessandro 2003), completing 24
high school (Davis, Ajzen, Saunders, and Williams 2003), and shopping on the web (Gentry and Calantone, 2002). It also may direct us either to develop different approaches to take with lower SES and higher SES persons who are currently in financial trouble or about to lose control of their finances. Our data show that higher SES debtors' spending for unneeded but desired products may be most easily altered with an approach that aims directly at decreasing their positive attitudes toward unneeded spending, as well as decreasing the positive subjective norms of significant others. This approach should work because the intention to purchase unneeded goods emerges from the debtors' attitudes and subjective norms. For higher SES debtors, we need to strengthen their own negative attitudes toward unneeded spending and at the same time direct them to consider and weigh heavily the views of significant others who would also constrain their unnecessary spending. On the other hand, attempts to rehabilitate the spending and saving actions of lower SES debtors should focus on decreasing their positive attitudes toward unnecessary purchasing and increasing their own sense of control of their intentions to buy. Our data show that intention to buy among lower SES debtors is a function of positive attitudes toward spending and perceived control, but that it does not draw on the value of subjective norms that debtors' infer from the actions of significant others. Lower SES debtors who rely on their own sense of control are more able to constrain their own unnecessary spending and seem to do so largely without the assistance of their perceptions of the views of significant others. Social Analytic Jurisprudence and Policy Issues We began this paper with a discussion of social analytic jurisprudence and we return to it to reach some tentative conclusions. First, we have shown the value of using this model to measure the fit between the conduct of lower SES (and higher SES) debtors and the current assumptions in the law. We identified some specific factors that 1) distinguished bankrupt consumers from others who do not get into trouble with debt and 2) we found categories of bankruptcy filers who share separate patterns of psychosocial characteristics. Indeed lower and higher social economic status debtors demonstrated different patterns of knowledge, attitudes, and behaviors that relate and predict their market place behaviors. It is our position that future reform efforts need to begin with an understanding of the type of people who are in need of bankruptcy relief and the naturally occurring differences among debtors who file for bankruptcy. 25
A simple rational actor model that assumes all actors are primarily motivated by self-interest and market forces will continue to miss significant differences between debtors and non-debtors and perhaps most importantly between lower and higher SES debtors. Our analyses advance our understanding of these important issues, but they are only a beginning of an empirical foundation for understanding ch. 7 and ch. 13 filers and the effect of legal rules on their decisions to borrow, default, and file for bankruptcy. In our current work, we are applying the adjustment or assessment function of psycholegal research to examine the implicit assumptions of proposals in the 2003 Act. The proposed bankruptcy bill before Congress assumes that more adequate disclosure rules will afford consumers protection against unwise use of credit cards. The assumption behind these credit card disclosure rules is that direct and understandable disclosure will shape debtors' spending and assist them to make rational decisions about credit card use. In this way, credit card disclosures permit debtors to avoid incurring credit card balances beyond their ability to repay and enable them to better assess and compare credit card offers. Disclosure also seeks to deter predatory lending practices by providing consumers with needed information. As such, disclosure will lead; the argument goes, to diminished debt. However, recent theory and research in the psychology of decision-making have witnessed an assault on the received view that people make choices based upon the expected utility of potential outcomes (Hastie and Dawes 2001; Lowenstein, Weber, Hsee, Welch 2001). The standard "rational actor" or "consequentialist" view has been that people assess the severity and likelihood of possible outcomes, integrate this with an expectation-based calculus, and then choose the best alternative. In other words, consumers first evaluate outcomes of their choices and assign value to the consequences of purchasing or not purchasing goods or services with a credit card (Hastie and Dawes 2001). Consumers then assign a subjective probability or likelihood to each outcome and evaluate their choices by summing the probability of the outcomes times the value (or cost) of the consequences associated with each choice. Beginning with the pioneering work of Daniel Kahneman and Amos Tversky (1979; Tversky and Kahneman 1974), many started to question the completeness of the expected utility model for understanding how people make decisions. Some of the early questioning arose from behavioral decision research and the study of cognitive heuristics under conditions of uncertainty (Kahneman, Slovic, and Tversky 1982). However, much of the more recent work focuses on the 26
roles of emotion in decision-making (Hastie and Dawes 2001). While behavioral decision theory psychologists do not question the usefulness of the expected utility approach, some have challenged the consequential model with a "risk as feelings perspective" (Lowenstein, Weber, Hsee, Welch 2001). According to the newer model, people evaluate outcomes and weigh them with both subjective probabilities and anticipated emotion, but they do so in different ways depending upon the feeling states that they experience while making their decisions (Kahneman and Ritov 1994; Kahneman, Ritov, and Schkade 1999; Kahneman, Schkade, and Sunstein 1998; Loewenstein, et al. 2001). In one program of current research, we contend that those who anticipate experiencing unpleasant emotions after making a credit card purchase will decide not to buy, and those who anticipate experiencing pleasant emotions will decide the other way. While decision makers may use clearly stated credit card disclosures to calculate anticipated emotions when they are not invoked from ongoing experience, disclosures are very unlikely to influence purchasing decisions when alternative sources of anticipated emotion are accessible. Thus, the proposed disclosure requirements in the 2003 Act may not temper purchasing decisions as expected because the disclosure does not influence consumers' anticipated emotions at the time of purchase. We are currently using the social analytic model to test these ideas and to determine whether the new disclosure law provisions fit well with the way in which consumers actually make credit card decisions. The data that we collect should be helpful in suggesting ways to revise the proposal to make the disclosure rules more effective. At the present, we can only speculate on the importance of emotion, motivation, and cognitive complexity as mediators in efforts to prevent consumers from faltering in the market place and to retool those who have already demonstrated distress and, sometimes failure. It is clear that we need additional research to learn more about experienced emotion, anticipated affect, mood related shopping, credit card disclosure practices, and alternative approaches to financial literacy. Such work is underway in the Coalition laboratories both on basic topics (i.e., the role of emotion in credit use and proposed disclosure laws) and in more applied areas (i.e., developing financial literacy training that takes into account differences in the social psychology of debtors). In the meantime, despite the limitations of the current study (i.e., the use of a comparison group different from ch 7 and ch 13 filers, and the need for more directed measurement techniques), this initial investigation of consumer debtors makes a pioneering 27
contribution to the literature. In the end, it is very clear that consumer debt is a complicated phenomenon that demands both psychological and empirical analyses, if the public policy debate about consumer debt is to take on a constructive direction leading to meaningful changes in law and in the market place. We offer social analytic jurisprudence as one approach to meet these goals and contribute to scholarship on the problem of building assets and credit among lowincome consumers. 28
References 11 U.S.C. § 1322. 1994. 11 U.S.C. § § 523, 727, 1994 & Supp. V (1999). America Saves. (2002, May 13). Survey About personal finances. http://www.americasaves.org/back_page/savinginamerica_first.cfm. Ajzen, Icek. 1988. Attitudes, Personality, and Behavior. Homewood, IL: Dorsey Press. Ajzen, Icek. 1991. Predicting Dishonest Actions Using The Theory of Planned Behavior. Journal of Research in Personality 25(3): 285-301. Bamberg, Sebastian. and Peter Schmidt. 2003. Incentives, Morality, or Habit? Predicting Students' Car Use for University Routes With the Models of Ajzen, Schwartz And Triandis. Environment and Behavior 35(2): 264-285. Bankruptcy Abuse and Prevention and Consumer Protection Act of 2003, H.R. 975, 108th Cong. 1st Sess. 2003. Baron-Donovan, Corinne, Richard L. Wiener, Karen Gross, and Susan Block-Lieb. 2003. Multiple Measure Evaluation of 'Training the Trainers': Implications for Pending Bankruptcy Legislation. Paper read at the Psychology and Law International, Interdisciplinary Conference, July, Edinburgh, Scotland. Bayer, Patrick., Douglas B. Bernheim, and John K. Scholz. 1996. The Effects of Financial Education in the Workplace: Evidence from a Survey of Employers. Working paper. No. 5655, p. 27. National Bureau of Economic Research. Bernheim, Douglas B. and Daniel Garrett. 1996. The Determinants and Consequences of Financial Education in the Workplace: Evidence from a Survey of Households. Working paper. No. 5667, p. 35. National Bureau of Economic Research. Bixler, Vencil, J. and Jan R. Squires. 1998. Enhancing the Field Experience in Financial Education. Financial Practice and Education 8(2): 83-87. 29
Block-Lieb, Susan. 2004. Mandatory protections as veiled punishments: Debtor education in H.R. 975, The Bankruptcy Abuse and Consumer Protection Act of 2003. Brooklyn Law Review, forthcoming. Boyce, Laurie and Sharon M. Danes. 1998. Evaluation of the NEFE High School Financial Planning Program, 1997-1998. Report of a study sponsored by the National Endowment for Financial Education, www.nefe.org/pages/educational.html. Braunstein, Sandra and Carolyn Welch. 2002. Financial Literacy: An Overview of Practice, Research, and Policy. Federal Reserve Bulletin 88(11): 445-57. Bradley, Donald, Abdighani Hirad, Vanessa G. Perry, and Peter Zorn. 2001. Is Experience The Best Teacher? The Relationship between Financial Knowledge, Financial Behavior, and Financial Outcomes. Paper submitted to Rodney L. White Center for Financial Research, University of Pennsylvania, Workshop on household financial decision making, March 2001. Chu, Pin-Yu and Ja-Fun Chiu. 2003. Factors Influencing Household Waste Recycling Behavior: Test of an Integrated Model. Journal of Applied Social Psychology 33(3): 604-626. Conner, Mark, Rachel Povey, Paul Sparks, Rhiannon James, and Richard Shepherd. 2003. Moderating Role of Attitudinal Ambivalence within The Theory of Planned Behaviour. British Journal of Social Psychology 42(1): 75-94. Cook, Thomas. D. and Donald T. Campbell. 1979. Quasi-experimentation: Design and analysis for field settings. Boston: Houghton Mifflin Company. Davis, Larry. E., Icek Ajzen, and Jeanne Saunders. 2002. The Decision of African American Students to Complete High School: An Application if the Theory of Planned Behavior. Journal of Educational Psychology 94(4): 810-819. Didia, Dal and Baban Hasnat. 1998. The Determinants of Performance in the University Introductory Finance Course. Financial Practice and Education 8(1): 102-07. 30
Furnham, Argyle and Michael Argyle.1998. The Psychology of Money. Florence, KY: Taylor and Francis/Routledge. Gentry, Lance and Roger Calantone. 2002. A Comparison of Three Models to Explain Shop-Bot use on the Web. Psychology and Marketing 19(11): 945-956. Hastie, Reid and Robyn M. Dawes. 2001. Rational Choice in an Uncertain World: The Psychology of Judgment and Decision Making. Thousand Oaks, CA: Sage. Hirad, Abdighani. and Peter M. Zorn. 2000. A Little Bit of Knowledge is a Good Thing: Empirical evidence of the Effectiveness of Pre-Purchase Home Ownership Counseling. Working paper. Freddie Mac sponsored study. H.R. 333, § 1301(a) (2002). H.R. 333, § 1303 (2002). Johnson, Dean L., Patrick Joyce, and Swapen Sen. 2002. An Analysis of Student Effort and Performance in the Finance Principles Course. Journal of Applied Finance 12(2): 67-72. Jump$tart. 2002. www.jumpstart.org/download.cfm Kahneman, Daniel and Ilana Ritov. 1994. Determinants of Stated Willingness to Pay for Public Goods: A Study in the Headline Method. Journal of Risk and Uncertainty 9: 5-38. Kahneman, Daniel, and Amos Tversky. 1979. Prospect Theory: An Analysis of Decision Under Risk. Econometrica 47: 263-291. Kahneman, Daniel, Ilana Ritov, and David Schkade. 1999. Economic Preference or Attitude Expressions: An Analysis of Dollar Responses to Public Issues. Journal of Risk and Uncertainty 19: 203-237. Kahneman, Daniel, David Schkade, and Cass R Sunstein. 1998. Shared Outrage and Erratic Awards: The Psychology of Punitive Damages. Journal of Risk and Uncertainty16: 4986. 31
Kahneman, Daniel, Paul Slovic and Amos Tversky. 1982. Judgment Under Uncertainty: Heuristics and Biases. Cambridge: Harvard University Press. Kaiser, Florian G. and Heinz Gutscher. 2003. The Proposition of A General Version of the Theory of Planned Behavior: Predicting Ecological Behavior. Journal of Applied Social Psychology 33(3): 586-603. Kim, Jinhee, Constance Y. Kratzer, and Irene E. Leech. 2001. Impacts of Workplace Financial Education on Retirement Plans. Paper read at the Annual Conference of the Association for Financial Consulting and Planning Education. Kuther, Tara L. and Ann Higgins-D'Alessandro. 2003. Attitudinal and Normative Predictors of Alcohol Use by Older Adolescents and Young Adults. Journal of Drug Education 33(1): 71-90. Kuhn, Thomas, S. 1962. The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Lachman, Roy, Janet Lachman, and Earl C. Butterfield, S. 1979. Cognitive Psychology and Information Processing. Orlando, Fl: Academic Press, Inc. Nofsinger, John and Glenn Petry.1999. Student Study Behavior and Performance in Principles of Finance. Journal of Financial Education 25: 33-41. Ogloff, James R. P. (Ed.). 2000. Taking Psychology and Law into the Twenty First Century. New York: Kluwer/Plenum Publishers. Price, Zachary. 2002. The Bankruptcy Abuse Prevention and Consumer Protection Act. Harvard Journal on Legislation 39: 237-260. Tennyson, Sharron and Chau Nguyen. 2001. State Curriculum Mandates and Student Knowledge of Personal Finance. Journal of Consumer Affairs 35(2): 241-262. 32
Tversky, Amos and Daniel Kahneman. 1974. Judgment Under Uncertainty: Heuristics and Biases. Science 185: 1124-1131. Vitt, Lois, A., Carol Anderson, Jamie Kent, Deanna M. Lyter, Jurg K. Siegenthaler, and Jeremy Ward. 2000. Personal Finance and the Rush to Competence: Financial Literacy Education in the U.S. Working paper. Fannie Mae Foundation, http://www.fanniemaefoundation.org/programs/pdf/rep_finliteracy.pdf. Wiener, Richard, L. 1993. Social Analytic Jurisprudence And Tort Law: Social Cognition Goes To Court. Saint Louis University Law Journal 37: 503-551. Wiener, Richard, L. 1995. Social Analytic Jurisprudence in Sexual Harassment Litigation: The Role of Social Framework and Social Fact. Journal of Social Issues 51: 167-180. Wiener, Richard, L. and Linda E. Hurt. 1997. Social Sexual Conduct at Work: How do Workers Know when it is Harassment and When it is Not? California Western Law Review 34: 53-99. Wiener, Richard, L. and Linda E. Hurt. 1999. An Interdisciplinary Approach to Understanding Social Sexual Conduct at Work. Psychology, Public Policy, and Law 5: 556-595. Wiener, Richard, L. and Linda E. Hurt. 2000. How Do People Evaluate Social-Sexual Conduct: A Psycholegal model. Journal of Applied Psychology 85: 75-85. Wiener, Richard, L, Melanie Rogers, Ryan Winter, Linda Hurt, Amy Hackney, Karen Kadela, Hope Seib, Shannon Rauch, Laura Warren and Ben Morasco. in press. Guided Jury Discretion in Capital Murder Cases: The Role of Declarative and Procedural Knowledge. Psychology, Public Policy, and Law. Wiener, Richard, L. and Christine C. Pritchard, and Mindy Weston.1995. Comprehensibility of Approved Jury Instructions in Capital Murder Cases. Journal of Applied Psychology 80: 455-467. 33
Wiener, Richard. L., Linda E. Hurt, Susan L Thomas, Melody S Sadler, Craig A. Bauer, and Tracy M. Sargent. 1998. The Role of Declarative and Procedural Knowledge in Capital Murder Sentencing. Journal of Applied Social Psychology 28: 124-144. Wiener, Richard L., Corinne Baron-Donovan, Karen Gross, and Susan Block-Lieb. under review. Debtor Education, Financial Literacy, and Pending Bankruptcy Legislation Behavioral Sciences and the Law. Wiener, Richard L., Amy Hackney, Karen Kadela, Shannon Rauch, Hope Seib, Laura Warren, L. and Linda E. Hurt. 2002. The Fit And Implementation of Sexual Harassment Law to Workplace Evaluations. Journal of Applied Psychology 87: 747-764. Wiener, Richard L., Linda E. Hurt, Brenda Russell, Kelley Mannen, and Chuck Gasper. 1997. Perceptions of Sexual Harassment: The Effects of Gender, Legal Standard, and Ambivalent Sexism. Law and Human Behavior 24: 71-93. Wiener, Richard L., Barbara Watts, Kristen H. Goldkamp, and Charles Gasper. 1995. Social Analytic Investigation of Hostile Work Environments: A Test of The Reasonable Woman Standard. Law and Human Behavior 19: 263-281. Wiener, Richard L., Ryan Winter, Melanie Rogers, and Lucy Arnot. in press. The Effects of Prior Workplace Behavior on Current Sexual Harassment Judgments. Law and Human Behavior. Wrightsman, Larry. S. 1991. Psychology and the legal system.Pacific Grove, CA: Brooks/Cole Publishing Company. 34
Table 1: Demographic Characteristics of Non-Debtors and Debtors
Measure Characteristic
Non-Debtors
Debtors
Full Sample
Number Percent Number Percent Number Percent
Gender Male Female
146
698
844
31
21
249
36
280
33
115
79
449
64
564
67
Age (Mean, S.D.)
146
26.01
687
43.49
833
40.42
8.50
12.32
13.49
Ethnicity
146
687
829
African American
11
12
286
42
303
37
White
45
31
181
26
226
27
Hispanic
26
18
122
18
148
18
Asian American
34
23
22
3
56
7
Other
24
16
72
10
96
11
Education < High School High School At Least Some College
145
687
832
1
1
37
5
38
5
54
37
326
48
380
45
90
62
324
47
414
50
Marital Status
146
696
842
Married
22
15
191
27
213
25
Single
118
81
290
42
408
49
Formerly Married
6
4
56
31
221
26
35
Table 1 (continued): Demographic Characteristics of Debtors and Non-Debtors
Measure Characteristic
Non-Debtors
Debtors
Full Sample
Number Percent Number Percent Number Percent
Current
146
680
826
Household Income
10,000 or below
18
12
90
13
108
13
10,001 to 20,000
14
10
117
17
131
16
20,001 to 30,000
14
10
147
22
161
20
30,001 to 50,000
46
32
196
29
241
29
50,001 to 80,000
33
23
104
15
137
17
80,001 or above
21
14
27
4
48
6
Last Year
146
683
829
Household Income
10,000 or below
21
14
98
14
119
14
10,001 to 20,000
18
12
110
16
128
15
20,001 to 30,000
22
15
141
21
163
20
30,001 to 50,000
40
27
200
29
240
29
50,001 to 80,000
24
16
100
14
124
15
80,001 or above
21
14
34
5
55
7
Employment UnEmployed Employed
146
687
833
67
46
265
39
332
40
79
54
422
61
501
60
Language
146
692
838
English Primary
103
71
606
88
709
85
English Secondary
43
29
86
12
129
15
Responsibility for Finances Self Other
146
691
837
67
46
524
76
591
71
79
54
167
24
246
29
Student Status Student Not a student
146
631
777
121
83
86
14
207
27
25
17
545
86
570
73
36
Table 2: Logistic Regression: Debtors and Non-Debtor differences on Demographic Variables (Predicting Debtor Status) (N=870)
Variable
Model 1 Forward Stepwise
Model 2 Forced Entry
Not a student
19.43***
20.41***
Old
7.49***
6.70***
Low Current Income
3.12**
5.25***
Finances - Self Managed
2.97***
3.19***
Married
3.25**
2.92**
High Income Last Year
-----
2.17
Male
-----
1.76
Not White
-----
1.40
Currently Unemployed
-----
1.32
At Least Some College
-----
.95
Nagelkerke R2
.59
.60
Model X2 (d.f.)
343.78 (5)***
331.38 (10)***
Notes: The numeric cell entries associated with each variable or odds ratios from the logistic regression equations. The Nagelkerke R2 represents the proportion of explained variability in a logistic regression equation (range 0.0 to 1.00). * p < .05, ** p < .01, *** p < .001.
37
Table 3: Logistic Regression: Debtors and Non-Debtor differences on Knowledge and Mood (Predicting Debtor Status) (N=870)
Variable
Before Demographics Forward Stepwise
After Demographics Forced Entrya
Shopping to End a Bad
1.89***
.82
Mood
Score on Knowledge Test
.87*
.98
Nagelkerke R2
.03
.60
Model X2 (d.f.)
16.94 (2)***
347.39 (12)***
---Block X2 (d.f.)
.45 (2)
Notes: The numeric cell entries associated with each variable or odds ratios from the logistic regression equations. The Nagelkerke R2 represents the proportion of explained variability in a logistic regression equation (range 0.0 to 1.00). * p < .05, ** p < .01, *** p < .001. a All 10 demographic predictors were forced entered on Block #1.
38
Table 4: Logistic Regression: Debtors and Non-Debtor differences on Attitude Measures (Predicting Debtor Status) (N = 870)
Variable
Before Demographics Forward Stepwise
After Demographics Forced Entrya
Positive Norms to Buy
1.02*
.99
Likelihood to Buy
.88*
.92
----
Positive Attitudes to Buy
1.02
----
Perceptions of Control
.89
Nagelkerke R2
.01
.61
Model X2 (d.f.)
10.03(2)**
354.70 (14)***
Block X2 (d.f.)
----
.45 (4)
Notes: The numeric cell entries associated with each variable or odds ratios from the logistic regression equations. The Nagelkerke R2 represents the proportion of explained variability in a logistic regression equation (range 0.0 to 1.00). * p < .05, ** p < .01, *** p < .001. a All 10 demographic predictors were forced entered on Block #1.
39
Table 5: Logistic Regression: Debtors and Non-Debtor differences on Self-Reported Consumer Behavior Measures for the Prior Month (Predicting Debtor Status) (N=870)
Variable
Before Demographics Forward Stepwise
After Demographics Forced Entrya
Unpaid Bills
5.29***
4.22***
No Saving Deposit
3.89***
3.15***
Created a Budget
1.98**
1.77
No Checking Deposit
----
2.05*
Did not Review Bills
----
1.06
Number of Credit Cards
----
.95
No Withdrawal from
----
.81
Savings
Borrowing from Predatory
----
.66
Lenders
Nagelkerke R2
.25
.67
Model X2 (d.f.)
127.86(3)**
381.75 (18)***
Block X2 (d.f.)
----
50.36 (8)***
Notes: The numeric cell entries associated with each variable or odds ratios from the logistic regression equations. The Nagelkerke R2 represents the proportion of explained variability in a logistic regression equation (range 0.0 to 1.00). * p < .05, ** p < .01, *** p < .001. a All 10 demographic predictors were forced entered on Block #1.
40
Table 6: Logistic Regression: Debtors and Non-Debtor differences on Self-Reported Use of Credit Cards in the Prior Month (Predicting Debtor Status) (N=522)
Variable
Before Demographics Forward Stepwise
After Demographics Forced Entrya
Credit Card Balance Paid
.27***
.37***
Number of Credit Card
.87*
Purchases
.77**
Credit Card Balance
----
1.14
(Natural Log Units)
Increase in Credit Card Balance Nagelkerke R2 Model X2 (d.f.) Block X2 (d.f.)
---- .39 133.60 (1)*** ----
1.21 .73 276.52 (14)*** 50.52 (4)***
Notes: The numeric cell entries associated with each variable or odds ratios from the logistic regression equations. The Nagelkerke R2 represents the proportion of explained variability in a logistic regression equation (range 0.0 to 1.00). * p < .05, ** p < .01, *** p < .001. a All 10 demographic predictors were forced entered on Block #1.
41
Table 7: Profile Analysis: High vs Low SES Debtors on Demographic Variables
Variable
High SES Debtors (N=287)
Low SES Debtors (N=275)
%Yes (n) %No (n) %Yes (n) %No (n) Chi Square (d.f. = 1)
30 Thousand+ Current
94 (269) 6 (18)
5 (15)
95(260) 437.77 ***
30 Thousand+ Last Year 91 (262) 9 (25)
8 (22) 92 (253) 389.73***
Currently Unemployed
22 (63) 78 (224) 52 (143) 48 (132) 54.61***
At Least Some College
61 (175) 39 (112) 36 (100) 64 (175) 34.04***
Married
31 (89) 69 (198) 16 (44) 84 (231) 17.52***
Finances - Self Managed 71 (204) 29 (83) 86 (238) 14 (37) 20.00***
Male
36 (103) 64 (184) 32 (89) 68 (186)
.78
White
29 (82) 71 (205) 22 (61) 78 (214)
3.02
Student
12 (35) 88 (252) 11 (31) 89 (244)
.12
Old
54 (156) 46 (131) 60 (164) 40 (111)
1.60
42
Table 8: Logistic Regression Profile: Low and High Status Debtors differences on Demographics (Predicting Low SES)
Variable
Model 1 Forced Entry
Low Current Income
228.44***
Low Income Last Year
106.25***
Currently Unemployed
5.78***
Finances - Self Managed
5.69*
Student
4.20*
Male
3.03*
Married
2.18
Not White
1.38
Old
1.19
At Least Some College Nagelkerke R2 Model X2 (d.f.)
.44 .92 654.89 (10)***
Notes: The numeric cell entries associated with each variable or odds ratios from the logistic regression equations. The Nagelkerke R2 represents the proportion of explained variability in a logistic regression equation (range 0.0 to 1.00). * p < .05, ** p < .01, *** p < .001. 43
Table 9: Forced Entry Logistic Regression (Predicting Low SES Status): High and Low
SES debtor differences in knowledge, mood, spending behavior, and credit card use
Variable
Knowledge/Mood Spending Credit Card Use
1.21 Unpaid Bills
No Saving Deposit
1.34
Created a Budget
1.05
No Checking Deposit
1.66**
Did not Review Bills
.94
Number of Credit Cards
1.01
No Withdrawal from Savings
.68
Borrowing - Predatory Lenders Credit Card Balance Paid
.78 .74***
Credit Card Balance
.89**
Increase in Credit Card Balance
1.17
Number of Credit Card Purchases
1.00
Score on Knowledge Test
.72***
Shopping to End a Bad Mood
1.29
Nagelkerke R2
.09
.03
.07
Model X2 (d.f.)
38.10 (2)***
11.79 (8)ns
12.29(4)*
Notes: The numeric cell entries associated with each variable or odds ratios from the logistic regression equations. The Nagelkerke R2 represents the proportion of explained variability in a logistic regression equation (range 0.0 to 1.00). * p < .05, ** p < .01, *** p < .001.
44
Table 10: Multiple Linear Regression: Predicting Irresponsible Buying in Low and High SES Debtors
Variable
High SES Debtors
Low SES Debtors
Attitudes
.38***
.46***
Subjective Norms
.20***
.01ns
Perceived Control R2
-.07ns .26
-.18*** .28
Full Model
F (3,270) = 31.35***
F (3,271) = 37.02***
Notes: The numeric cell entries associated with each variable are Beta weights from the multiple regression equations. The R2 represents the proportion of explained variance in each regression equation (range 0.0 to 1.00). * p < .05, ** p < .01, *** p < .001.
45

RL Wiener, K Gross, S Block

File: joint-center-for-housing-studies-harvard-university.pdf
Title: Microsoft Word - babc_04-2.doc
Author: RL Wiener, K Gross, S Block
Author: ltrayes
Published: Tue Feb 24 14:27:20 2004
Pages: 47
File size: 0.19 Mb


PARENTERAL, 4 pages, 0.2 Mb

, pages, 0 Mb

The new greatest generation, 11 pages, 0.14 Mb

s and Descriptions, 19 pages, 0.06 Mb
Copyright © 2018 doc.uments.com