Mapping Opportunities for Strategic Investment in Poverty Reduction: Psychology and Federal Antipoverty Legislation

Abstract

Although high-profile examples of psychology’s contributions to public policy exist—particularly around antipoverty legislation—little systematic review has quantified how the field has informed federal policy across time. Recognizing the importance of exploring psychology’s use in policymaking, we provide an overview of psychology’s presence in federal antipoverty legislation over the last two decades by reviewing the over 6,000 antipoverty bills introduced to Congress since 1993 for mentions of psychology. Further, to explore how psychology’s contributions are related to policymakers’ attributions about the causes of poverty, their public statements and voting behavior is considered. Key gaps in our scientific knowledge for informing poverty-related policy are identified. Opportunities to enhance the relevance of psychology in poverty reduction efforts, including the evidence-based policy movement, are described.

Editor’s note.

This article is part of a special section, “Psychology’s Contributions to Understanding and Alleviating Poverty and Economic Inequality,” published in the Xxxxxx 2019 issue of American Psychologist. Heather E. Bullock and Diane M. Quinn served as editors of the special section.

Introduction

An extensive body of psychological research has documented the negative effects of living in poverty (Brooks-Gunn & Duncan, 1997; Evans, Li, & Whipple, 2013; Yoshikawa, Aber, & Beardslee, 2012). While society has employed an array of strategies to reduce or mitigate poverty, the most visible often come in the form of national policies (Pierson, 1995; Rose & Baumgartner, 2013). Yet, while evidence of psychology’s role in national policy can be illustrated by case study and anecdotal evidence (Blank & Blum, 1997; Yarrow, 2011; Zigler & Styfco, 2010); relatively few quantitative assessments of how psychological knowledge is used in federal policymaking exist. To better quantify how federal legislators have leveraged psychology in antipoverty policy, we present findings from a large-scale review of bills introduced to Congress over the last two decades.

This article describes the evidence-based policy movement as an opportunity for leveraging behavioral science, provides a brief overview of psychology’s role in key federal policy efforts to reduce and mitigate poverty, and examines legislators’ attributions about poverty’s causes. Next an approach to quantifying the magnitude of psychology’s role in antipoverty legislation is outlined. Then potential opportunities to increase the use of psychological knowledge in policymaking based on legislators’ attributions about poverty and voting behavior are considered. How psychology may augment its role in supporting evidence-based antipoverty policy is discussed.

Evidence-based Policymaking and the Role of Psychology

There is a growing trend toward evidence-based policy in which decisions are informed by bodies of rigorously established objective evidence (Commission for Evidence-based Policy, 2017; Haskins & Margolis, 2015; Oliver, Lorenc, & Innvær, 2014; Vandlandingham, 2015). The evidence-based policy movement has, in part, grown out of the experimental research traditions that psychologists first applied to the study of human behavior and there are an increasing number of opportunities to leverage psychological research in federal antipoverty efforts (Pearce & Raman, 2014). Although research offers an important resource for policymakers interested in investing in effective poverty reduction strategies, actually translating scientifically-based strategies into public policy remains difficult (Fox, 2005; Haskins & Margolis, 2015; Pearce & Raman, 2014; VanLandigham, 2015). Nevertheless, the use of research evidence in evidence-based policymaking represents a bipartisan opportunity for developing effective policies and programs informed by psychological science (Gamoran, 2018; Supplee & Metz, 2014).

Although it is important not to equate policymakers’ use of psychology with evidence-based policy, even token uses of a knowledge base are consistent with what is understood about how policymakers leverage empirical evidence. Since many evidence-based policy initiatives emerged with the underlying belief that society is ‘best served’ when scientific evidence is used in a comprehensive manner to inform decision making, this aim is contradicted when policymakers select research findings that serve an already established policy goal (Liebman, 2013; Sanderson, 2002). However, this tactical use is only way in which research is used by policymakers (Nutley, Walter & Davies, 2007). More consistent with the aims of evidence-based policymaking is conceptual use (i.e., broadly impacting the way problems are understood), instrumental use (i.e., directly guiding decision-making), and imposed use (i.e., enforcing the use of research standards; Nutley et al., 2007, Tseng, 2014). In this sense, legislative mentions of psychology could conceivably be used to garner political support, indirectly influence how problems are addressed, justify specific antipoverty strategies, or inform regulations on standards of evidence to be executed by administrative agencies. References to psychology in legislation examined in this article may involve any of these types of research use—as a greater understanding of the general use of psychology in federal antipoverty legislation, irrespective of how it was used, has the potential to strengthen psychologists’ efforts to demonstrate relevance of behavioral science to policymakers.

Psychology’s Role in Antipoverty Policy

While many national policies seek to reduce poverty or mitigate its effects (Currie, 2011; Edin & Kissane, 2010; Yarrow, 2011), key examples exist of how psychology has informed or been leveraged in legislative efforts to combat poverty. For instance, the War on Poverty was launched in 1964 as a multifaceted effort to reduce poverty across the country by introducing a number of government services and functions that comprise much of the social safety net as it operates today (Zigler & Styfco, 2010). Ultimately educational and developmental psychologists played a key role in convincing policymakers to define the end goal of Head Start services in terms of optimizing school readiness and have contributed to its longevity (Head Start Act Amendments, 1998). For many psychologists, addressing poverty is essential to better protect children from abuse and neglect (Drake & Pandley, 1996; Dumont & James-Brown, 2017). A more recent example is the Maternal, Infant, Early Childhood Home Visiting (MIECHV) Program, which primarily serves impoverished families with program models that empirically demonstrate effectiveness in a range of outcomes such as:

“Improvements in child health and development, including the prevention of child injuries and maltreatment and improvements in cognitive, language, social-emotional, and physical developmental indicators.”

Examples such as these help the field understand how federal antipoverty efforts may be informed by psychology, yet the extent of psychology’s use is not well understood despite the field’s contributions to demonstrating the detrimental impacts of poverty and effective strategies that reduce poverty’s impact.

Historically, psychology has produced research that may be relatively well-suited to inform policy. For instance, the field was one of the first disciplines to systematically study the consequences of poverty and economic inequality—particularly within experimental contexts (see Brooks-Gunn & Duncan, 1997; Yoshikawa et al., 2012; Zigler, 1994). Additionally, a broad literature in behavioral and social science, both within and outside of psychology, has demonstrated that growing up in poverty has detectable impacts on both brain development and functioning, physical and mental health outcomes, educational achievement, criminal behavior, workforce participation, parenting and social interactions (Duncan & Murnane, 2011; Evans, Chen, Miller, & Seeman, 2012; Korenman & Sjaastad, 1995; Shonkoff & Phillips, 2000; Sommer et al., 2017). Particularly compelling for policymakers is interdisciplinary research on how poverty gets ‘under the skin’ (e.g., biological embedding) and impacts development across the lifespan (Evans et al., 2012; Frameworks, 2010; Noll & Shalev, 2018). Furthermore, a substantial literature base has given rise to potentially effective strategies for lifting individuals out of poverty, mitigating poverty’s detrimental impacts, and disrupting intergenerational poverty (Bradley & Corwyn, 2002; Howard & Brooks-Gunn, 2009; Kirby & Barzelatto, 2001; Olds et al., 2010; Wolf, Aber, & Morris, 2013). More recently this has included key collaborative work with other disciplines in behavioral economics, engineering and economic evaluation to optimize program impact (Crowley et al., 2018; Collins, 2018; Currie, 2004; Gennetian, Darling, & Aber, 2016). Such applications of psychological research in interventions could be a potentially powerful tool for decision makers.

Legislator Attributions of the Causes of Economic Inequality and Antipoverty Policy

In order to understand the role of psychology in public policy, it is essential to appreciate the perspectives of policymakers themselves, including how a person’s attributions about the causes of poverty and economic inequality are related to their support for poverty-related policy (Abouchedid & Nasser, 2001; Bullock, Williams, & Limbert, 2003; Cozzarelli, Wilkinson, & Tagler, 2001). Further, different types of attributions are more common within different constituencies (e.g., welfare recipients, middle class, key sociocultural groups; Al-Zahrani & Kaplowitz, 1993; Bullock, 1999; Kluegel & Smith, 1986). Research on attributions of the causes of poverty generally identify two pronounced attribution categories—structural and individualistic attributions.

Structural attributions about the causes of poverty focus on the importance of contextual and environmental conditions (Bullock, 2004; Kluegel & Smith, 1986). For instance, individuals that hold structural attributions cite labor market inequality, lack of government support or pervasive prejudice as factors that lead to poverty and make mastery difficult (Bledsoe, Combs, Sigelman, & Welch, 1996). Psychologists, and other social scientists have explored how this attribution relates to prejudice and bias in the hiring process (Duffy, Blustein, Diemer, & Autin, 2016; Koch, D’Mello, & Sackett, 2015; Milkman, Akinola, & Chugh, 2015) or diminishes access to government supports (Suárez-Orozco, Yoshikawa, & Tseng, 2015). Also, the effects of government support programs on structural factors have been examined, providing evidence that mitigating environmental conditions can reduce poverty or risk factors associated with poverty (Auger, Farkas, Burchinal, Duncan, & Vandell, 2014; Lee, Zhai, Brooks-Gunn, Han, & Waldfogel, 2014; Morrissey, Hutchison, & Winsler, 2014).

A different type of attribution focuses on how those living in poverty are themselves culpable for their situation (i.e., ‘individualistic attributions’; Jost, Banaji, & Nosek, 2004). For instance, individualistic attributions include referencing individuals’ lack of ability, lack of intelligence, or failure to maintain a stable household (Bullock, 1999; Hwang, Fitzpatrick, & Helms, 1998; Jost et al., 2004; Kluegel & Smith, 1986). Psychological research has contributed empirical knowledge through research on factors such as task persistence, educational achievement, and family stability (Eskreis-Winkler, Shulman, Beal, & Duckworth, 2014; Fuller-Rowell, Evans, Paul, & Curtis, 2015; Roy & Raver, 2014). Research demonstrating that changes in individualistic factors can reduce the likelihood of living in poverty or mitigate its harm provides further evidence for supporting policies targeting individualistic factors (Hawkins & Erickson, 2015; Zigler & Styfco, 2010).

Importantly, these two attribution types are not the converse of each other as an individual can make both types when explaining poverty (Hunt, 2002; Hunt & Bullock, 2016). While other attribution classes have been considered (e.g., fatalistic), as well as sub-categories, individualistic and structural attributions represent the classes most extensively studied and validated across populations, contexts and settings (Furnham & Gunter, 1984; Hunt, 1996, 2002; Mistry, Brown, Chow, & Collins, 2012). Existing research has explored the relationship between these types of attributions and support for specific antipoverty strategies (Hunt & Bullock, 2016). Previous research has identified that individuals who make strong structural attributions are more likely to report support for antipoverty strategies such as laws to reduce workplace discrimination or affirmative action policies in higher education (Bullock, Williams, & Limbert, 2003). Those who make strong individualistic attributions report favoring strategies such as marriage and relationship supports for low-income individuals or programs aimed at increasing youth motivation to seek and maintain employment (Hunt & Bullock, 2016). This work seeks to understand whether such attributions relate to actual legislator voting behavior.

Goals of the Current Paper

While the field of psychology’s role in the ‘War on Poverty’ during the 1960s has been relatively well documented via a series of case studies (Yarrow, 2011; Zigler & Styfco, 2010), its role in more recent decades is not well quantified. Little is known about the frequency with which the field of psychology directly informs legislative provisions that seek to reduce poverty or the likelihood that such bills are supported by legislators. To examine how psychology may inform national poverty policy, findings from a large-scale review of bills introduced to Congress over the last two decades are presented. In particular, the magnitude of psychology’s use in recent decades is quantified by measuring the frequency of direct references to psychology in proposed or enacted legislation. Further, how such references relate to legislators’ voting behavior is explored.

While researchers have assessed the attributions of a number of different groups, no work to date has directly studied the attributions of US legislators. Calls for rigorous strategies for measuring such attributions underscore the opportunity to assess public statements of elected officials (Hunt & Bullock, 2016; Lepianka, Van Oorschot, & Gelissen, 2009). The rise of digital records and archiving of public communications of elected officials—particularly at the federal level—now makes mining the public statements of US legislators feasible. The measures for assessing poverty attributions lend themselves to a variety of measurement approaches, including coding of public statements based on keywords and phrases. In this context, our work is guided by the following research questions:

1. When and how often has psychology been directly referenced in antipoverty legislative language over the last two decades?

2. Is legislators’ voting behavior different for antipoverty bills that do and do not reference psychology?

3. What structural and individualistic attributions do legislators make in their public statements?

4. To what extent are legislators’ attributions associated with their voting behavior?

Methods

To better understand the role of psychology in poverty-related federal policy, all federal bills and policymakers’ public statements were reviewed and connected with policymaker voting behavior data (Jan 1, 1993–December 31, 2016). Primary documents included all resolutions and bills introduced to Congress over the last two decades (N = 141,026), public statements from legislators made in a broad array of mediums (N = 6,830,560; e.g., floor statements, social media) and the voting record of every member of Congress over the last two decades (N = 38,332 votes). The process for identifying (1) poverty-related legislation, (2) direct references to psychology in legislation and (3) legislators’ attributions about the causes of poverty within these three types of primary documents are described below, as is the approach to modeling the association between these direct references within poverty-related legislation, corresponding voting behavior, and legislators’ attributions about poverty. Constructs described below are summarized in Table 1.

Table 1:

Coded Content in Bills, Voting behavior and Public Statements

Construct	Source	Protocol	N
Poverty-Related Bills	Full-text bills	Selected if CRS legislative subject was ‘Poverty and Welfare Assistance’	6,245 Bills
Bills Referencing Psychology	Full-text bills	Selected if bill contained the keyword: ‘Psychology’	368 Bills
Legislator Voting Behavior	Voting records for sampled bills	Computed a proportion of Yea votes out of total	38,332 votes
Structural Attribution	Legislators’ public statements and social media posts	Classified as structural if contained any keywords: ‘job layoffs’, ‘prejudice,’ ‘social safety net.’ Proportion of total legislator statements was used for analyses.	12,549 mentions
Individualistic Attribution	Legislators’ public statements and social media posts	Classified as individualistic if contained any keywords: ‘unskilled labor,’ ‘stupid’, ‘broken home.’ Proportion of total legislator statements was used for analyses.	2,218 mentions

Poverty-related Legislation.

After a bill is introduced to Congress, the Congressional Research Service (CRS) is required to analyze the bill and provide a determination of the bill’s “policy areas” (N = 32) and “legislative subjects” (N = 1,004). Since most bills cover multiple issues, multiple subjects are assigned to each bill. This classification system began three decades ago; therefore, this categorization offers a consistent approach for identifying poverty-related legislation (CRS, 2017). As the focus of this study is legislation that seeks to reduce poverty, we selected every bill introduced over the last two decades that received a subject tag for Poverty & Welfare Assistance (N = 6,245 bills). In an effort to be inclusive, we included all bills designated in this area whether or not the primary issue area was poverty and welfare assistance. Further, for each poverty-related bill, we identified whether the bill was enacted into law.

References to Psychology in Legislation.

All poverty-related bills were then identified as either directly referencing psychology or not. This involved reviewing all text in each bill to identify whether terms indicative of a direct reference to the field were written into legislative language (N = 368 poverty bills referencing psychology). To maximize comprehensiveness, we employed a state-of-the-art legislative data platform known as Quorum (2019). This platform allows for complete searches of keywords and all derivatives, including an automated search of all stems of the root word. Specifically, this included psychology and all derivations (e.g., psychological, psychiatric, psychologist). Identifying direct-references to psychology provides the most visible evidence of psychology’s role in federal antipoverty policy yet yields a conservative estimate of psychology’s influence. Pieces of legislation likely are influenced by the science and practice of psychology that do not directly reference the field. As such, this methodology underestimates the field’s influence—and we recognize its use and reach are likely greater. Further, considering the interdisciplinary nature of poverty research, it is important to note that this work does not seek to capture the influence of the many different but related areas of study (e.g., child development, environmental health science). Therefore, findings here should be seen as a lower-bound estimate of influence narrowly focused on psychology’s direct reference.

Legislator Voting Behavior.

The voting records of all legislators who had the opportunity to vote for at least one poverty-related bill were reviewed (N = 38,332 votes for poverty-related bills). The proportion of poverty-related bills for which a legislator cast a favorable vote was calculated (i.e., All Antipoverty Votes). Then the legislator’s (1) rate of voting for poverty-related bills that reference psychology (i.e., Psychology Antipoverty Votes) and (2) rate of voting for poverty-related bills that did not reference psychology (i.e., Non-Psychology Antipoverty Votes) were calculated.

Legislator Attributions.

Next, legislators’ public statements were searched and identified for whether or not they included phrases indicative of common structural and individualistic attributions of the causes of poverty (e.g., Bullock et al., 2003; Hunt & Bullock, 2016). All legislators who had the opportunity to vote for at least one antipoverty bill were included in this review. Public statements were drawn from a multitude of sources that were consolidated into searchable, full-text data by Quorum. Public statements included all press releases, floor statements, ‘Dear Colleague letters’, letters to federal agencies, statements in public committee meetings, Twitter posts, Facebook posts, Medium posts, Instagram posts and statements in YouTube videos from the legislator (N = 6,830,560).

Each statement was identified as having a structural attribution (N = 12,549) if it mentioned keywords (or their derivatives) indicating structural factors identified as particularly representative of this attribution type (Bullock 1999; Kluegal & Smith 1986). These included poor labor market (i.e., job layoffs), inequality (i.e., prejudice), lack of government support (i.e., social safety net). A statement was identified as an individualistic attribution (N = 2,218) if it mentioned keywords (or their derivatives) indicating commonly identified individual factors, including statements referencing an individual’s lack of ability (unskilled labor, lacking intelligence (i.e., ‘stupid’;), or unable to maintain a stable household (i.e., broken home). In this way, each statement was dichotomously coded for the presence of both types of attribution based on keywords (Table 1).

To provide a comparable estimate of the use attributions by each legislator, as some legislators make many public statements and some make relatively few (M = 6,954; SD = 4,211), the proportion of total statements that were Structural Attributions or Individualistic Attributions was calculated. In this manner, a legislator with a Structural Attribution rate of 1% makes 1 statement using phrases indicative of structural attributions out of every 100 public statements made.

Analytic Approach.

First, the association between whether a bill does or does not reference psychology and whether or not it is enacted was modeled though a logistic regression analysis. Then legislators’ voting behavior in relation to their attributions about the causes of poverty as indicated in their public statements was modeled within a generalized linear regression. Specifically, All Antipoverty Votes was regressed simultaneously onto Structural Attributions and Individualistic Attributions (Model 1). Then Psychology Antipoverty Votes was regressed simultaneously onto Structural Attributions and Individualistic Attributions (Model 2). Finally, Non-Psychology Antipoverty Votes was regressed simultaneously onto Structural Attributions and Individualistic Attributions (Model 3). Post-hoc analyses were carried out to control for the role of party affiliation on voting behavior and explore its relationship with attribution types.

Results

Findings from this review of legislative language, congressional office communication, and voting behavior are described below. First, findings regarding direct references to psychology in legislation are described. Then we describe the association between legislator attributions of the causes of poverty and voting behavior.

Statutory References to Psychology.

How psychology is directly referenced in legislation highlights how the knowledge from psychological research and practice has been translated into public policy. Such influence is illustrated in a number of interrelated ways. One of the most common references to psychology in legislation is made to justify or support policy. For instance, H.R.5294 Health Equity and Accountability Act of 2014 aimed to support the health of minority populations and summarizes the literature on the harms of poverty:

§ 1002(3)(J) Social class differences account for a large part of health disparities. For example, children living in poverty experience poorer housing conditions, increased exposure to indoor allergens and toxins (such as pesticides, lead, mercury, radon, air pollution, and carcinogens), and more psychological stress. These experiences culminate in worse adult health as compared with children with higher socioeconomic status. Specifically, children living in socioeconomic neighborhoods have higher rates of asthma due to higher rates of psychological stress resulting from higher rates of violence.

References are also made to define an investment in the science and practice of psychology. These often describe new training programs, research funding, or recognition and support of the field. In particular, such investments are often part of efforts to expand access to mental health services. An example of research support can be seen in H.R.803 Workforce Innovation and Opportunity Act; Section 204 (29 U.S.C. 764) that was amended to specify the types of studies the National Institute on Disability, Independent Living, and Rehabilitation Research should be funding:

§ 204(1)(B)(ii) studies and analyses of factors related to industrial, vocational, educational, employment, social, recreational, psychiatric, psychological, economic, and health and wellness variables affecting individuals with disabilities, including traditionally underserved populations as described in section 21, and how those variables affect such individuals’ ability to live independently and their participation in the work force;

Another example of statutory language supporting psychological research on how to reduce or mitigate poverty is found in Chapter 6 of the S. 1932 Deficit Reduction Act of 2005 that gives states the option to allow families of disabled children to purchase Medicaid coverage. Specifically, it authorized demonstration projects on Medicaid coverage of home- and community-based alternatives to residential psychiatric treatment.

§ 6063 (a) In General.--The Secretary is authorized to conduct, during each of fiscal years 2007 through 2011, demonstration projects (each in the section referred to as a ``demonstration project”) in accordance with this section under which up to 10 States (as defined for purposes of title XIX of the Social Security Act) are awarded grants, on a competitive basis, to test the effectiveness in improving or maintaining a child’s functional level and cost-effectiveness of providing coverage of home and community-based alternatives to psychiatric residential treatment for children enrolled in the Medicaid program under title XIX of such Act.

Other references reflect recognition of psychologists’ expertise. For instance, this may entail requiring their representation on committees, boards, and within the peer review processes. This is illustrated in the H.R.34 21st Century Cures Act that describes amended peer review requirements to disburse funding from Public Health Service Act:

§ 6009 Section 504(b) of the Public Health Service Act (42 U.S.C. 290aa–3(b)) is amended by adding at the end the following: “In the case of any such peer review group that is reviewing a grant, cooperative agreement, or contract related to mental illness treatment, not less than half of the members of such peer review group shall be licensed and experienced professionals… and have a medical degree, a doctoral degree in psychology…

Another type of reference to psychology in poverty-related legislation is language that seeks to protect the psychological wellbeing of individuals. For instance, the H.R.2610 - Transportation, Housing and Urban Development, and Related Agencies Appropriations Act, 2014 sought to restrict spending by the United States Interagency Council on Homelessness for workforce development-related training:

§ 406(a) None of the funds made available in this Act may be obligated or expended for any employee training that—(1) does not meet identified needs for knowledge, skills, and abilities bearing directly upon the performance of official duties; (2) contains elements likely to induce high levels of emotional response or psychological stress in some participants;

Bill’s Reference to Psychology and Enactment.

Based on these qualitative analyses, we created a dataset of poverty-related bills that indicates whether they directly reference psychology or not. These data were, in turn, linked to data on whether the bills were successfully enacted. This afforded the opportunity to quantitatively explore whether antipoverty bills directly referencing psychology were more or less likely to be enacted.

Findings from these analyses indicate that 5.91% of all poverty-related bills (N = 6,245) introduced in any year from January 1, 1993 to December 31, 2016 directly reference psychology (N = 368; Figure 1). Of bills that directly reference psychology, 39 were enacted, for a 10.6% enactment rate. In contrast, of the 5,877 poverty-related bills introduced that did not directly reference psychology, 376 were enacted (6.4%). Specifically, poverty-related bills that directly reference psychology are 65.6% more likely to be enacted then poverty-related bills that do not reference psychology (χ² = 11.50, p < .001).

Figure 1:

Poverty-related bills introduced, 1993–2016

Legislator Attributions and Voting Behavior.

Coding legislators’ public statements for beliefs regarding the causes of poverty indicated contexts for key terms and phrase usage. For instance, in an example of a structural attribution, Senator Marco Rubio (R-FL) described his belief in the inadequacy of the current social safety net:

“… Washington doesn’t have all the answers. More than 50 years after President Lyndon Johnson declared war on poverty, it’s clear our social safety net programs are in desperate need of innovation and modernization…”

Senator Tammy Baldwin (D-WI) provided a structural attribution of the contribution of economic and labor market weaknesses to highlight problem families.

“Wall Street was bailed out of this mess but the impact was felt by small businesses and hard working families who are still struggling to recover from reduced economic growth, job layoffs, lost income, home foreclosures and retirement savings that were wiped away.”

Examples of individualistic attributions of poverty’s causes are similarly pronounced in legislators’ public statements. For instance, Senator James Lankford (R-OK) remarked in a Father’s Day interview with U.S. News and World Report about the role of family instability on societal and economic problems:

“To take stress off the government, families must become more stable and self-sufficient. Look at any community across America and you will see the direct correlation. In communities with more broken families, and more absent fathers, you will see higher crime and a weaker economy. The opposite will exist in communities with more stable families.”

By identifying individualistic and structural attributions in each public statement, we were further able to quantify the frequency of each legislator’s use of such attributions in their public statements. Of legislators, 94.4% who had the opportunity to vote for poverty-related legislation made at least one public statement that contained a structural or individualistic attribution. Of all the public statements these legislators made, on average about 3.3% (SD = 3.4%) included a structural attribution and 1.6% (SD = 2.7) included an individualistic attribution. Below, we model the association between this proportion of attributions in public statements (i.e., how often they make an attribution) and legislator voting behavior (Table 2).

Table 2.

Relationship between Attribution Type and Legislative Voting Behavior

Parameter	Estimate	SE	t	p
Model 1: All Antipoverty Votes
Intercept	0.71	0.01	48.52	<.0001
Structural Attributions	0.12	0.03	4.24	<.0001
Individualistic Attributions	−0.10	0.04	−2.83	<.01
Model 2: Psychology Antipoverty Votes
Intercept	0.67	0.02	32.43	<.0001
Structural Attributions	0.19	0.04	4.72	<.0001
Individualistic Attributions	−0.13	0.05	−2.49	<.01
Model 3: Non-Psychology Antipoverty Votes
Intercept	0.67	0.02	42.97	<.0001
Structural Attributions	0.14	0.03	4.52	<.0001
Individualistic Attributions	−0.06	0.04	−1.52	0.13

Attribution data from public statements were then linked to legislator voting behavior to assess association between legislators’ attributions and All Antipoverty Votes. Model 1 revealed that when legislators used more Structural Attributions (_β = .12, p < .001) or fewer Individualistic Attributions (_β = −.10, p < .005), they were more likely to vote for All Antipoverty bills F(2,525)=17.95, p < .001. Specifically, a 1% increase in legislators’ Structural Attributions corresponded with a 12% increase in their All Antipoverty Votes. In contrast, a 1% increase in legislators’ Individualistic Attributions corresponded with a 10% decrease in legislators’ All Antipoverty Votes (Figure 2).

Figure 2:

Association between legislator attributions and All Antipoverty Votes (Model 1)

Model 2 revealed that legislators making more Individualistic Attributions were less likely to vote for Psychology Antipoverty bills (_β = −.13, p < .01), whereas legislators making more Structural Attributions were more likely to vote for Psychology Antipoverty bills (_β = .19, p < .001) F(2,525)=22.30, p<.001 (Figure 3). Specifically, a 1% increase in legislators’ Structural Attributions corresponded with a 19% increase in their Psychology Antipoverty Votes. In contrast, a 1% increase in legislators’ Individual Attributions corresponded with a 13% decrease in their Psychology Antipoverty Votes.

Figure 3:

Association between legislator attributions and Psychology Votes (Model 2) and Non-Psychology Antipoverty Votes (Model 3)

Model 3 revealed that Non-Psychology Antipoverty bills were most likely to be voted for by legislators making more Structural Attributions (_β = .14, p < .001), whereas there was no association among legislators making more Individualistic Attributions (_β = −.06, p = .130) F(2,525)=20.44, p<.001) (Figure 3). Specifically, a 1% increase in legislators’ Structural Attributions corresponded with a 14% increase in their Non-Psychology Antipoverty Votes.

Post-hoc analyses were employed to explore the association between attribution, party affiliation and voting behavior (Table 3). When controlling for party affiliation, legislators structural and individualistic attributions remain significantly related to voting behavior for All Antipoverty Votes. When considering bills that reference psychology, structural attributions remain a significant predictor of Psychology Antipoverty Votes, but neither individualistic attributions nor party affiliation predict voting behavior. Both attribution types and party affiliation predict Non-Psychology Antipoverty Votes.

Table 3.

Relationship between Attribution Type, Party and Legislative Voting Behavior

Parameter	Estimate	SE	t	p
Model 1a: All Antipoverty Votes
Intercept	0.76	0.02	46.94	<.0001
Structural Attributions	0.20	0.03	6.72	<.0001
Individualistic Attributions	−0.19	0.04	−5.10	<.0001
Party Affiliation	−0.14	0.02	−6.33	<.0001
Model 2a: Psychology Antipoverty Votes
Intercept	0.65	0.02	27.56	<.0001
Structural Attributions	0.16	0.05	3.53	<0.01
Individualistic Attributions	−0.10	0.06	−1.70	0.09
Party Affiliation	0.05	0.03	1.60	0.11
Model 3a: Non-Psychology Antipoverty Votes
Intercept	0.73	0.02	42.58	<.0001
Structural Attributions	0.24	0.03	7.34	<.0001
Individualistic Attributions	−0.17	0.04	−4.12	<.0001
Party Affiliation	−0.17	0.02	−7.04	<.0001

Discussion

This work highlights the role psychology is already playing in federal antipoverty policy. It also highlights potential opportunities to increase the relevance of our field to better support policy based on psychological knowledge. In particular, about 5% of more than 6,000 poverty-related bills introduced in last two decades directly reference psychology. This finding points to psychology’s direct and observable role in legislation as well as the possibility for a much larger role that extends beyond specific references to psychology. Notably, poverty-related bills that referenced psychology were more likely to become law than poverty-related bills that did not reference psychology. Historically, across all policy areas, only one of every nine bills introduced by Congress becomes law. Our analyses showed that poverty-related legislation that referenced psychology was over 60% more likely to become law than poverty-related legislation that did not reference psychology. Specifically, 1 in 8.5 antipoverty bills referencing psychology became law, whereas only 1 in 14 antipoverty bills that did not reference psychology became law.

Additional analyses reveal that policymakers’ attributions about poverty significantly predict voting behavior for poverty-related bills. Specifically, building on existing efforts to understand how attributions about the causes of poverty in turn influence decision-making, public statements by legislators over the last two decades were coded to assess legislative members’ individualistic and structural attributions. We found these attributions to be predictive of voting behavior for all poverty-related legislation over the last two decades. For instance, legislators who make structural attributions at a higher rate were more likely to vote for antipoverty bills, whereas legislators making more individualistic attributions were less likely to vote for antipoverty bills. This builds on prior interview research which showed similar associations between attributions of poverty and self-reported support of antipoverty policy (Hunt & Bullock, 2016). Importantly, this work also highlights that policymaker attributions uniquely predict voting behavior even when controlling for party affiliation.

These analyses also reveal differential voting behavior among legislators who make individualistic or structural attributions about the causes of poverty, depending on whether a bill directly references psychology. As illustrated in Figure 3, legislators who made many structural attributions in public statements, while generally more likely to vote for any poverty-related bill, were most likely to vote for antipoverty bills referencing psychology. In contrast, results also revealed that legislators who made higher rates of individualistic attributions, while generally unlikely to vote for poverty-related bills, were even less likely to vote for antipoverty bills referencing psychology. This seems to point to underutilized opportunities for engaging legislators who tend to make more individualistic attributions. Such interactions could focus on ways in which psychology could address their concerns about the underlying causes of poverty (e.g., ‘broken households’) in future antipoverty legislation. In particular, this might be accomplished by forging more collaborative relationships between psychologists and policymakers, as a growing empirical literature demonstrates that interactions are critical to supporting policymakers’ efforts to craft evidence-based policy (Crowley & Scott, 2017; Crowley, Scott & Fishbein, 2018; Oliver et al., 2014). Not only might such efforts strengthen the recognition of how psychology can contribute to antipoverty efforts, these results also suggest that such efforts may increase the likelihood of a bill’s enactment.

This review highlights the role of psychology in federal antipoverty efforts and the need for interdisciplinary research in this area. Relatively little research has investigated the association between the use of evidence in legislation and its enactment. Even less work has studied direct associations between the attributions legislators make about the reasons for a public problem and actual legislative voting behavior. Moreover, the research base is particularly scant when investigating issues related to how legislators’ attributions about poverty relate to their actual voting behavior. This study builds on important efforts to apply attribution theory in examining support for past antipoverty policies.

In addition to contributing to this empirical base, this study’s inclusion of all relevant bills and public statements (as opposed to sampling) strengthens the generalizability of these findings. This work also highlights methodological approaches now available as a result of greater digitization of legislative records. Importantly, while previous survey research has assessed preferences for specific antipoverty strategies (e.g., nutrition assistance, job training; Bullock, 1999; Bullock et al., 2003), bills generally include a mix of approaches that might be considered progressive as well as regressive. This arises from the need for compromise between and within political parties and constituencies. In recognition that votes are cast in support of or against a bill, it is most generalizable to consider antipoverty bills as a whole since a legislator much choose to vote for or against a whole bill—not only the parts with which they agree. This means that bills represent a natural unit of interest, as opposed to attempting to parse how legislators’ attributions align with a specific policy strategy. In this context, we sought to expand the literature to consider the association between attributions and support for specific policies, as well as the link between attributions and federal antipoverty efforts as a whole.

Limitations and Future Work

While this study makes a contribution to current knowledge of psychology’s role in US antipoverty legislation, it also highlights opportunities for future study. For instance, the results are generalizable to recent U.S. federal policy, not for state or local policy. Further, while this review focused on strategic keywords and phrases drawn from the massive number of primary documents (i.e., bills and public statements), advances in semantic analysis, machine learning, and qualitative methods are likely to allow for greater depth in future research regarding psychology’s role in antipoverty policy (e.g., Zhu & Mitra, 2009). Such analyses of public statements may also detect additional structural and individualistic attributions beyond the approach used in this study. Also, while this study considers the attributions of a legislator based on their public statements, it does not capture the internal process of deciding how to vote on an individual bill—or how privately held attributions inform this process. The politically charged nature of federal policymaking encourages covert behavior that will require alternative methods for assessing privately-held attributions and decision-making processes. Ultimately, we believe the present review likely provides a conservative estimate of psychology’s role in federal antipoverty policy and offers a baseline for interdisciplinary study of psychology’s utilization.

While this work provides a broad overview of psychology’s use in legislation, many future research directions should be explored to improve our understanding of the role psychology plays in informing antipoverty policy—and public policy in general. In particular, legislative reviews, such as the one introduced here, highlight the opportunity and importance of systematically quantifying the role of psychology in policymaking. Doing so not only highlights the existing value for the field but can indicate the decision makers likely to be receptive to outreach from psychologists. As work in this area progresses, we encourage the development of increasingly nuanced metrics for assessing psychology’s role in legislation and public policy.

Over the years, the field of psychology has made many important contributions to federal antipoverty legislation. This work indicates that psychology’s role may be larger and more complex than many realize. Further, this work provides some of the first quantified estimates of psychology’s use in antipoverty legislation. Ultimately, such insights can guide future research and outreach efforts to support evidence-based policymaking.

Biography