An Analysis of Voting and Legislative Behavior

Abstract

Despite the scope and breadth of applied behavior analysis (ABA) over its 60-year history, little attention has been directed toward the formulation and implementation of public policy. This lack of attention is notable because Skinner (1953) posited that government is probably the most obvious agency engaged in the control of human behavior. Although behavioral strategies have been employed to address policy issues, most studies examined small groups in circumscribed settings. Glenn’s (1988) conceptualization of the metacontingency provided a framework for examining public policymaking, with culturo-behavioral science rapidly emerging as a means to further advance our understanding of the complex interactions involved in social and cultural systems (Glenn, 2003; Malott & Glenn, 2019) and the continuing evolution of public policy. This article focuses on voting as an operant behavior and the interlocking behavioral contingencies (IBCs) at play when citizens vote at the polls and lawmakers are voting on potential legislation. Because virtually all legislative bodies have specific protocols regarding everything from legislative drafting to floor debate, the majority of their activity involves rule-governed behavior. In contrast, the votes which a legislator casts, like those of the general public in an election, are contingency-shaped behaviors. One key difference between the vote cast by a private citizen and a legislator are the external consequences that can be imposed following the vote by legislative or governmental leaders. Despite having only a small number of behavior analysts serving in legislatures, recent successes surrounding licensure and mandated insurance coverage for behavior analysis have resulted in a greater awareness by legislators and policymakers of the need and value of such services, suggesting that this is an opportune time for behavior analysts to become more involved and shape public policy.

For over 60 years, applied behavior analysis (ABA) has affected millions of individual consumers, their families, practitioners, and professionals across disciplines and geographical boundaries. The scope and breadth of behavior analysis can be seen in various ways including the 64,626 individuals holding certification from the Behavior Analyst Certification Board (BACB, n.d.) the 41 special interest groups (SIGs) that exist within the Association for Behavior Analysis International (ABAI, 2023) and the 350 domains of socially significant behavior compiled by Heward et al. (2022)

Despite this popularity and growth, one area that has been underrepresented in the literature is public policy and policymaking, which has been defined as the effort by government actors to produce socially desirable results (Dworkin, 1975). This lack of attention to the political process and its impact on human behavior is notable because Skinner (1953) posited that government is probably the most obvious agency engaged in the control of human behavior. He added that an adequate analysis of such a government would include a study of the techniques used by the individual in becoming a member of the agency and in maintaining himself as such.

Goldstein and Pennypacker (1998) noted that a genuine understanding of the behavioral processes in representative government requires an analysis of the behavior of political incumbents. In particular, they emphasized the need to “identify the environmental events that are antecedent and consequent to identifiable classes of ‘political behavior’ and to establish functional relations between such behavior and its environmental determinants” (Goldstein & Pennypacker, 1998, p. 2). In essence, to understand politics and public policy, one needs to understand the politician and policymaker. Moreover, Spiller (2008) stated that “The workings of the policymaking process are determined (to some extent) by the political institutions in each country, such as the presidential/parliamentary nature of the government, the electoral rules in place, the federal structure of the country, and the existence of an independent judiciary.”

To illustrate, the United States and numerous other countries have an open list voting system where people cast votes for individual candidates and the legislators seeking election actively seek support from those voters. This contrasts with closed-list proportional representation systems in countries such as Argentina, Spain, Turkey, and Uruguay, where voters are presented with a list of candidates and they vote for the party, rather than the individual candidates they prefer. In such closed-list systems, the candidates are rank-ordered by the party and these individuals do not appeal directly to voters, but emphasize their support for the policies and positions of the party leadership, thereby hoping to improve their standing within the party and their relative position on the party’s candidate list. Despite the fact that these different electoral systems provide different incentives for candidates (Hix, 2004), almost all legislators are subordinate to party leadership within their assemblies (Carey, 2007) regardless of the electoral system in place in their country.

Although this article focuses on the public policymaking process at the state and federal level within the United States, many of the underlying assumptions and analyses could be applicable to other countries and provinces, if one studies those systems and the relevant contingencies affecting their policymaking process. However, one potentially significant difference between countries is the use of gender quotas for elected offices. Although they do not exist in the United States, 85 other countries and territories, representing more than half the total number, have enacted legislation at the national or subnational level implementing such quotas (Dahlerup et al., 2013).

As Koop et al. (2018) observed, legislators in many political systems serve multiple principals who compete for their loyalty in legislative votes. Knowing how the legislative process works and the idiosyncrasies within the relevant assembly or agency are important first steps in understanding and influencing legislative behavior, regardless of the electoral process or governmental structure.

Thirty-four years following the publication of Science and Human Behavior, Skinner acknowledged a collective failure to identify and manage the key contingencies that are influencing the behavior of public servants (Skinner, 1987). There was a concerted effort during that same time period to increase the participation and contributions of psychologists and behavior analysts in the public policy process. A body of research evolved that demonstrated the efficacy of behavioral strategies to address policy issues such as recycling (Witmer & Geller, 1976; Jacobs et al., 1984), obesity (Wooley et al., 1979), crime prevention (Schnelle et al., 1979; Kirchner et al., 1980), and social and community action (Thyer et al., 1986). Although groundbreaking research had begun to emerge, the studies were generally limited to small groups of individuals in circumscribed and usually controlled settings (Lamal, 1991) and little attention was directed toward the policymaking process itself.

Thus, committees and task forces were established by the Association for the Advancement of Behavior Therapy (Jones et al., 1983), American Psychological Association (Task Force on Psychology & Public Policy, 1986) and Association for Behavior Analysis (Morris, 1988). These efforts sought to educate their members on the policymaking process (e.g., Fawcett et al., 1988; Seekins & Fawcett, 1986) and make specific recommendations to better prepare behavior analysts to actively participate in the formation and adoption of public policy (Task Force on Public Policy, 1988). This was critical because knowledge and experience in the application of behavior principles, methodology, and interventions is insufficient in and of itself to achieve meaningful social change (Ardila-Sánchez et al., 2020b). Behavior analysts also needed to acquire specific knowledge and skills to understand and navigate the political and cultural environment of public policymaking.

A turning point came when Glenn (1988) recognized the distinction between the contingencies of reinforcement at the individual level from those operating at the cultural level and introduced the concept of metacontingency, defined as “a contingent relation between (1) recurring interlocking behavioral contingencies (IBCs) having an aggregate product; and (2) selecting environmental events or conditions” (Glenn et al., 2016). The conceptualization of the metacontingency was a significant contribution that bridged the gap between the individual and cultural levels of analysis and broadened the scope of behavior analysis considerably (Houmanfar & Rodrigues, 2006). Scholarship and research have continued in this area, focusing on the impact of organizational leaders in affecting cultural change (Houmanfar et al., 2015) and the role of cultural milieu as a mediating factor in cultural change (Houmanfar et al., 2020). Expanding the scope of inquiry and working to adopt consistent terminology across studies, Glenn et al. (2016) helped to provide a theoretical framework for research and analyses dealing with a wide variety of public policy areas, including welfare reform (Lamal, 1997), pedestrian traffic laws (Sénéchal-Machado & Todorov, 2008), national health systems (Martins, 2009), climate change (Alavosius et al., 2016), invasive species management (Malott & Glenn, 2019), and systemic racism (Saini & Vance, 2020).

Although these efforts addressed different public policy areas, the policymaking process itself has received little attention and not just by behavior analysts. As Anderson et al. (2022) stated, “political scientists have often displayed more facility and zeal for theorizing about public policy making than for actually studying policy and the policymaking process” (p. 21). A notable exception was Malott (2015) who analyzed the role of the 1947–1948 U.S. Senate Foreign Relations Committee in the establishment of the Cold War as a metacontingency.

Policymaking in the United States occurs within all three branches of government, but it generally occurs most commonly within the legislative and executive branches (Smithey & Robison, 2021) at both the state and federal level. In fact, no major policy can be adopted, no major program initiated, no taxes levied or monies appropriated without legislative involvement and approval (Rosenthal, 2009). But, although the media often refers to the legislature having passed a particular bill, the legislature cannot really act. Rather, it is the people within the legislature, the individual members who are the real actors (Rubin & Feeley, 2003) and who are the focus herein.

Fenno (1973) interviewed more than 600 members of six congressional committees and concluded they had three goals—reelection, influence within the House, and making good public policy. Mayhew (1974) concurred that legislators are focused on getting reelected and this pursuit often dictates how they spend their time, the positions they take, the publicity they seek, and with whom they interact (Arnold, 2004). But running for office and holding office are two different things. Akirav (2015) contended that the skills required for election and reelection are different from those required for enacting legislation and distinguished between two types of legislators—those who are electable and those who are successful in the legislature.

Most people often view Congress and their state legislature as similar bodies, but there are several key differences. First, state legislators are closer to the people, with each member of a state House of Representatives or Assembly representing approximately 61,000 citizens and each member of Congress representing approximately 488,000 citizens (Ballotpedia, 2021; U.S. Census Bureau, 2021). Second, in general state legislatures are not full-time jobs. Only three states have full time legislatures (i.e., California, New York, Pennsylvania), meaning that most state senators and representatives have “day jobs” (FiscalNote, 2017). Third, state legislatures move faster and pass legislation at a higher frequency than Congress (FiscalNote, 2021). For these reasons and the fact that most research has studied congressional behavior (e.g., Lamal & Greenspoon, 1992), this article will primarily focus on legislative behavior at the state level rather than at the federal level.

Although legislatures vary in several ways (e.g., size, composition, length of session, partisan makeup, rules and protocols), they share one commonality—voting. Whether one serves in the New Hampshire Senate, the California State Assembly, or even as a member of Congress, legislators are concerned with two types of votes—the votes from constituents that get them elected and the votes they take once they hold their elective office.

Nearly a century ago, it was noted that “the most tangible and measurable units of political behavior are votes” (Rice, 1925). Voting has been seen as the key operant in our electoral process with researchers relying on a simple three-term contingency and analyzing voting from the perspective of the individual voter (e.g., Visser, 1996). Candidates want the public to vote for them and the average citizen is bombarded in “election season” with campaign ads, mailings, debates, robocalls, and other prompts to “get out the vote.” For some people and demographic groups (e.g., senior citizens), these prompts are unnecessary because voting appears to be a rule-governed behavior as they look forward to “the Tuesday following the first Monday in November” and never miss an election. For other groups (e.g., 18- to 29-year-olds), however, turnout is generally lower and inconsistent, so that voting is a contingency-shaped behavior.

Visser (1996) assumed that voting was an learned or acquired behavior, initially through parental modeling and reinforcement for party preference and then subsequently modeled and reinforced in adulthood by one’s partner. According to this model, consistency and the direction (i.e., party preference) of voting depends heavily on the voting choices of those closest to the individual. Visser (1996) also assumed that voters need not be informed on campaign events, or political issues and that simply knowing when and where the elections would take place was sufficient to get them to vote, suggesting that voting was a rule-governed behavior that was reinforced and maintained by the social reinforcement from one’s parents and partner. The recurrence of voting for candidates from the same party or reflecting similar ideology and the replication of voting behavior across individuals within families and communities can be seen as an emerging culture-behavioral lineage (Glenn, 2003; Glenn et al., 2016). If, however, an individual voted for a different candidate or party, Visser (1996) assumed that the parent or partner would disapprove of the choice, thereby decreasing the likelihood of its reoccurrence.

Figure 1 provides a hypothetical example of the numerous factors that could potentially influence whether or even how a citizen may vote in an election. The arrows illustrate that any of the antecedents on the left side could affect the likelihood of individuals casting their vote, which could result in one or more of the consequences on the right side serving to reinforce that vote. For example, the high voter turnout for the 2020 presidential election and the 2022 midterms suggest that voters are becoming more energized as a result of greater partisanship (Pew Research Center, 2022; Frey, 2023). Thus, one’s political party or ideology may be the relevant stimulus that motivates the person to go to the polls and cast a vote for the candidate who shares that ideology. As an alternative, a candidate’s position or an incumbent legislator’s voting record on a controversial issue (e.g., abortion, gun control) might be the key stimulus motivating a citizen’s vote. For other people, the calendar and “election day” itself may be sufficient to get them to the polls.

Fig. 1.

Potential Factors Influencing a Citizen’s Vote

Lovett (2022) noted that voters have multiple motivations and their vote may be driven in part by issues, in part by group identities, and in part by the expected performance of the candidate. Thus, there may be multiple relevant antecedent factors that can affect whom an individual votes for, or even whether they vote, or not. Some antecedents might favor Candidate A whereas others would suggest that Candidate B would be the better choice. Each of these factors may differ in importance and their cumulative effect may determine who receives the person’s vote. It is interesting that although political scientists have focused on whether citizens should vote at all, because each vote has a small chance of influencing an election, little attention has been directed toward systematically investigating what motivates people to vote (Lovett, 2022).

As seen in Fig. 1, once individuals have made a decision and cast their vote, there are a broad range of consequences, which may serve as reinforcers and the arrows on the right side show the connection between a citizen’s behavior (i.e., voting) and these potential reinforcers. Although reinforcement may be immediate (e.g., fulfill your civic duty, your candidate wins or loses the election) or delayed (e.g., the winner’s legislative votes support your ideology, your specific policy concern is addressed), the relevant, functional consequences may not only influence the likelihood of voting in future elections, but which candidate receives their vote as well.

It is obvious that Fig. 1 provides an illustrative, but by no means exhaustive, list of relevant antecedents and potential reinforcers affecting voting behavior. A person’s vote may be seen as a differential response to these contextual stimuli, controlled by its consequences. Because voters tend to use different criteria in voting in local elections than in national ones (Oliver et al., 2012), the relevant stimuli and controlling contingencies can differ from election to election, or even on the same ballot within a given election. For example, party affiliation may be a critical factor determining who to support for the state senate seat, but may be irrelevant in local races (e.g., school committee) where the candidates often do not declare any party affiliation.

The emphasis thus far has been on the vote of an individual citizen and the contingencies associated with that individual vote and the 2022 elections provided two illustrations of the importance of that individual vote. In Massachusetts, the margin in one legislative race was a single vote (Cawley, 2023) and a New Hampshire state representative race required a separate, special election after the initial race ended in a tie vote (Porter, 2023). In both cases, one vote made a difference, reinforcing the importance of getting out every possible vote, in every election.

Following the election and after assuming office, legislators are confronted with a different type of vote; namely, voting on the bills before them. These votes, like citizen votes, are operant behaviors with various situational antecedents and reinforcers, which may vary in terms of importance and potency and their relative influence on the actual vote taken.

Figure 2 provides a hypothetical depiction of the potential contingencies affecting a legislator’s vote. As with the citizen vote, there are multiple antecedents, reinforcers, and contingencies which can affect a legislator’s votes, some of which are similar and some of which are different from those seen in election voting by the general public. Once again, the arrows show the relationship between the antecedents, the target behavior of voting, and the potential reinforcing consequences.

Fig. 2.

Potential Factors Influencing a Legislator’s Vote

One of the most potent antecedents for legislators is input from constituents. These are the individuals who elected the legislator, who the legislator represents, and who will determine whether the legislator will continue to represent them in the future. Thus, constituent service is seen as a key element of representation (Bussell, 2019) and some have argued that it probably takes up more of a state legislator’s time and energy than any other task (Keefe & Ogul, 1989, p. 21). Although anecdotal evidence indicates that state legislators generally receive fewer than 5–10 contacts (i.e., phone call, email, letter) on an issue, Bergan and Cole (2015) found that state legislators who received at least one phone call from a constituent asking them to support a certain bill were 11%–12% more likely to support the legislation, demonstrating the potential impact of a person contacting their legislators and expressing their views on a particular issue or bill.

Legislative colleagues are another important antecedent factor that can affect a legislator’s vote. Given the wide scope of issues addressed by state legislatures each term, it is nearly impossible for individual legislators to be knowledgeable on every matter that comes before them. As a result, they often rely on colleagues, particularly “go-to” legislators who developed specialized expertise on a topic area for information and guidance about particular legislation (Sarbaugh-Thompson et al., 2004; Bogenschneider & Bogenschneider, 2020). Likewise, because legislators need to secure the support of a majority of the membership for their particular bill to pass, they will reach out to colleagues to solicit support and their vote.

Lobbyists, political action committees (PACs), and other special interest groups also can potentially influence a legislator’s vote either by making financial contributions or providing information which is relevant in deciding to support one bill rather than another (Austen-Smith & Wright, 1992). Political contributions are seen as a vehicle to gain access to legislators (Grenzke, 1989), build and maintain connections (Bertrand et al., 2014), alter the language of a bill (Hall & Wayman, 1990), or influence votes (Roscoe & Jenkins, 2005).

There is a popular belief that you can buy a legislator’s vote through campaign contributions and that money influences legislators to vote against the best interest of their constituents. However, most researchers have concluded that donors have little influence on roll call votes in Congress (Powell, 2013) and although Powell (2012) found that donor influence did vary substantially among the 99 legislative chambers,1 measuring the influence of contributions on roll call votes proved difficult in state legislatures (Powell, 2014). In a subsequent study, however, Matsusaka (2017) found no evidence that campaign contributions changes legislative votes or persuaded legislators to “stray from what the constituency wants” (Schechter, 2017). Furthermore, the data in this study indicated that state legislators’ roll call votes were congruent with the majority views of their constituents 65% of the time and when legislators voted contrary to their constituency, it was not due to contributions. Rather, they were voting in accordance with their own interests, beliefs and ideologies over those of the people they represented. It is interesting that Matsusaka (2017) also noted that state legislators who are more attentive to constituent interests are better able to raise contributions, suggesting that their vote may be an important stimulus for future contributions from donors who are supportive of the vote that was cast.

Due to concerns that people could conceivably “buy a vote” through their political contributions, many states have placed limits on individual contributions and PAC contributions to political candidates, but these restrictions vary widely. For example, Colorado limits individual contributions to candidates for state legislative races to $200, whereas Ohio allows individual contributions of $15,499.69 to candidates for state house and senate races and Alabama, Nebraska, Oregon, Utah, and Virginia have no contribution limitations whatsoever (Coalition for Integrity, 2022). Likewise, 43 states either impose the same limitations on PACs as those for individuals or provide a separate contribution limit, with 7 states allowing PACs to contribute unlimited amounts to state campaigns (National Conference of State Legislatures, 2023). It is interesting that Massachusetts is the only state that imposes stricter limits for lobbyists with individuals able to contribute up to $1,000 per candidate, but registered lobbyists may only contribute up to $200 per candidate (National Conference of State Legislatures, 2023).

Although lobbyists, like other individuals, can contribute to any legislator’s political campaign within the state’s statutory limits, Wright (1990) argued that the act of lobbying has a greater impact on voting behavior than a donation from a PAC or a special interest group. He also examined votes at the committee level and found that the number of times a member was contacted by a lobbying group was much more important in explaining voting behavior than financial contributions, suggesting that information sharing had greater value than money.

Looking again at the potential reinforcers for a legislator’s vote in Fig. 2, some may be immediate and others may have a more long-term impact. The nature of the reinforcer may also vary, with some having a direct effect on the legislator’s constituents (e.g., constituent issue, district project funding) and others having little or no bearing on constituents or the district (e.g., committee assignments, financial contributions). Although either category can be seen as potentially reinforcing for a legislator’s vote, citizens’ votes are more likely to be reinforced by those consequences that directly affect them as opposed to those which do not.

Although there is considerable similarity between the citizen and legislator voting scenarios, there are some significant differences. First, all citizen votes are private so we never know how an individual voted and why they cast the votes that they did. In contrast, virtually all legislators’ votes, whether by electric voting machine or by an oral roll call, are public votes that are recorded and published in the journal of the legislative body. This allows constituents and other interested parties to know how a legislator voted on a particular bill. The only exceptions are voice votes where the presiding officer calls for the “yeas” and “nays” and determines which side has the most votes without the names or tallies of members on each side being recorded. For the most part, therefore, the public nature of legislative voting enables constituents and other interested parties to easily determine how their legislators voted on a particular bill and where they stand on a particular issue.

Second, citizens generally cast votes in elections once or twice every 2 to 4 years. On the other hand, legislators may cast dozens of votes during a particular session and hundreds of votes during a legislative term, and the strategies, dynamics and contingencies can vary significantly with each vote. Nevertheless, constituents and other interested parties can review legislators’ voting records to see where they stand on any given bill or issue and how these legislators “represent” their district and constituents.

Third, another critical difference, seen in Fig. 2, is the relationship and interactions between an individual legislator and legislative leaders or the governor. The legislative leader (i.e., House speaker or Senate president) is responsible for setting the legislative agenda, scheduling bills for floor action, making committee assignments and determining the role and workload of each member. Likewise, the governor has a policy agenda and can, through the power of the pen, sign or veto any bill or budget item in order to advance that agenda. As a result, governors and legislative leaders can have significant influence on legislators and be the key factor determining whether particular legislation dies or becomes law.

Figure 2 includes bidirectional arrows to illustrate potential interactions among the governor, legislative leaders, and the individual legislator, which could affect the legislator’s vote on a given bill. Given their power and authority, the legislative leaders or the governor can, and sometimes do, impose consequences for a legislator’s vote or actions. Within the past year, for example, a Pennsylvania state representative was told to vacate his Capitol office in February 2023 and was not reassigned office space for 2 months for refusing to sign a letter needed for a parliamentary maneuver to call the House back into session (Murphy, 2023). Other notable examples include a state legislator having been banned from speaking on the chamber floor for the remainder of the 2023 session in Montana (Reiman, 2023), another suspended and stripped of seniority in Arkansas (Arkansas Democrat Gazette, 2022) and two expelled in Tennessee (Greenfield, 2023) for statements they made or actions that they took in their role as state legislators. It should be noted that the Arkansas case was the only one where the legislator who was punished was a member of the same party as the legislative leadership.

A final difference between citizen and legislative voting are the operational contingencies. Although voting has been characterized in terms of a three-term contingency at the individual level (Visser, 1996), voting in an election can also be seen as involving IBCs as posited by Glenn and Malott (2004). In particular, they defined IBCs as contingencies in which the behavior or behavioral product of one individual acts as the antecedent for another individual’s behavior and the consequences for both individuals are shared. Citizen voting is an antecedent which results in a product (election of a legislator) that serves as the antecedent for another individual’s behavior (legislator vote), leading to a consequence (district project being funded) benefiting both individuals. That consequence also serves as an antecedent stimulus for the citizen to vote, presumably for the legislator in the next election, thereby maintaining the IBC and increasing its likelihood to occur in the future.

In contrast, legislator voting is more complex, in part, due to how they receive and retain their position and how bills become law. Unlike other occupations where people are hired and job tenure is, in large part, dependent upon a performance evaluation from one’s supervisor and where performance appraisals are required in 75% of individual jobs in U.S. state government (Selden et al., 2001), legislators are beholden solely to voters and subjected to a performance review once every 2–4 years at election time. Regardless of their success in getting bills passed or securing funding for their district, legislators must secure a plurality of the votes cast in the next election to essentially receive a renewal of their contract.

It is not surprising, therefore, that members of Congress appear to be single-mindedly interested in reelection (Fenno, 1973; Mayhew, 1974) and the same appears to be true for state legislators (Perry, 2013). Elections can be viewed as a metacontingency where the predominant behavior patterns of legislators are those that will result in their election or reelection (Lamal & Greenspoon, 1992). It is this event (election or reelection) that sets the occasion for legislative voting and “the consequence of reelection subsumes contingencies that account for the voting behavior of incumbents” (Lamal & Greenspoon, 1992, p. 73). The actual votes taken by legislators can be seen as an aggregate product that result in cultural consequences (e.g., policy changes, project funding). Through their votes, legislators seek to achieve their electoral goals by working to provide benefits and services for the district or for individual constituents within the district (Martin, 2011) and gaining their vote in the next election. But, as discussed below, a legislator’s vote may also have other consequences from leadership that have no direct impact on the citizen voter. Although the aggregate product or cultural consequences of a legislator’s vote may or may not affect individual voter’s behavior, it determines how the behavioral contingencies of the participants (i.e., citizens and legislators) interlock (Tadaiesky & Tourinho, 2012).

Legislatures are complex systems with formal structures (e.g., leadership hierarchy, committee assignments) and specific rules, procedures, and practices guiding their operations, but they also are institutions where each member is equal to every other member and individuals do not formally owe each other obedience and do not necessarily accept any division of labor among themselves (Loewenberg, 2015). Within the United States, state legislative branches range in size from 40–400 members (National Conference of State Legislators, 2021) and the IBCs described earlier for one candidate and the eligible voters in the legislative district are essentially replicated for every other legislator. There also are additional IBCs that reflect the interdependent social contingencies (de Carvalho & Sandaker, 2016) among members. With every bill requiring a plurality to secure passage, some degree of cooperation is necessary and social structures matter (Sandaker et al., 2019) in order to achieve the desired goal.

Because a legislator needs to secure support from 50% of the members in a legislative body, one would expect that a bill sponsor will actively solicit support and votes from their colleagues. Where does one look to get these votes? The obvious suspects are members of their own party, colleagues who share the same ideology (Cox & McCubbins, 2006), legislators whose district is contiguous or close to theirs (Pellegrini & Grant, 1999), desk mates (Masket, 2008), members who have co-sponsored their bills or whose bills they have co-sponsored in the past (Fowler, 2006), legislative caucuses (Kingdon, 1989), and colleagues who they routinely interact or socialize with (Peoples, 2008) both during and outside of legislative sessions. These interactions with people sharing some similarity or mutual interest assume reinforcing value and the “similar other serves as a discriminative stimulus leading to more interaction and IBCs” (Bento et al., 2020b).

As noted earlier, the governor and legislative leaders may have competing agendas, especially in the absence of unified party control of both legislative houses and the governorship (Lupia & McCubbins, 1994). These elected officials (i.e., governor, House speaker, Senate president) may also weigh in prior to a legislative vote, leading to more complexity and potentially greater difficulty is gaining the plurality necessary to secure passage of any bill.

Figure 3 shows a subset of the multiple interconnections that exist between elected officials and the voters they represent. Each of the various elected officials depicted are accountable to the voters who elected them and several (e.g., mayor, state representative, state senator) work collectively on behalf of their shared constituency. In contrast, although individual state representatives and state senators may collaborate with colleagues and co-sponsor specific legislation and work together on mutual areas of interest, the primary focus is the impact on their own district and the constituents who live there, rather than their colleagues’ districts. But, if legislators hope to see their priority bills pass, they must work to gain the support of their colleagues, knowing that they may be asked to reciprocate and vote to support their colleagues bills as well.

Fig. 3.

Metacontingencies in State Government

It should be noted that the aforementioned analysis of the processes and contingencies underlying legislative voting is intended to reflect a particular moment in time and does not address the ever-changing composition of the membership, flows of information, and availability of social and other forms of reinforcement (Bento et al., 2020a). For example, a partisan swing from Democratic to Republican control, or vice versa, or the retirement or election defeat of members on the leadership team or chairing powerful committees could change the dynamics of the legislative body and the relevant IBCs and metacontingencies at play.

History has shown that fewer than 3% of members of Congress are psychologists (Martin, 2003; Newbould, 2007; Sullivan, 2001, 2003; Sullivan & Reedy, 2005) with an even smaller number, perhaps fewer than 1% serving in state legislatures (O’Connor, 2001). Furthermore, of the total number over the years, probably less than a handful in Congress or state legislatures have been behavior analysts.

As noted earlier, despite these small numbers and a lack of training in the application of behavioral principles, there is clear evidence that legislative leaders and governors whether in Congress or state legislatures have utilized reinforcement or punishers as consequences for legislative votes. For example, the former speaker of the U.S. House of Representatives, John Boehner (R-Ohio), once warned the members of his party that leaders are “watching” how the rank and file vote to determine committee assignments (Hooper, 2012). More recently, a Florida state senator claimed that Governor Ron DeSantis vetoed $29 million in budget items for his District because the senator endorsed former President Donald Trump over the governor for president in 2024 (Nusbaum, 2023).

Although legislative leaders are inclined to apply behavioral principles with their members, the contingencies may not always have the desired effect. Although rank-and-file members who vote with leadership often receive choice committee assignments and support for their legislation (Coker & Crain, 1994), anecdotal evidence suggests that punishment for “voting off” does not necessarily decrease the likelihood of doing so for future votes, as seen recently with the Freedom Caucus in Congress under Speaker McCarthy (Joseph, 2023). In fact, if the leader is unpopular in a member’s district, the contrary vote may enhance the member’s reelection chances.

The use of reinforcers or punishers as described above should not be surprising and although the significance of these various consequences will vary from member to member, perhaps the most potent reinforcer for a legislator is the passage of their priority bills. Just as legislative leaders have a policy agenda for each legislative session, all members have their own legislative “wish list” that they would like to see enacted for their constituents and district. Passage of these bills demonstrates effectiveness, often measured by tangible benefits for the legislative district and its constituents.

To help ensure passage of their priority bills, legislative leaders will bring their agenda items forward early in the legislative term and often hold members’ bills until later in the session. Members who vote to support the bills on the leadership agenda are more likely to see their own priority bills move. Those who do not support the leadership bills are less likely to see their priority bills reach the floor.

But, it is important that a percentage of an individual member’s bills be brought forward by leadership and are not “killed” early in the legislative session. Otherwise, leadership may not get the support needed from rank-and-file members to ensure passage of leadership’s priority bills. Although a failure by legislative leadership to advance their agenda could be personally disappointing, there may be an ever greater consequence. Members whose bills are repeatedly “killed” are less likely to cast a vote for the current leaders, thereby jeopardizing their own ambition to continue or ascend further in the hierarchy in the next legislative session. A version of this scenario played out in the U.S. Congress in January 2023 when Representative Kevin McCarthy (R-CA) needed 15 rounds of voting and a week of negotiations and concessions before he was able to win his speaker bid (Goodwin et al., 2023) and similar examples were recently seen in the Ohio House (Trau, 2023) and California Assembly (White, 2023).

Given the strategy of legislative leadership to hold bills until late in the session or to kill them, often in response to an individual member’s voting record, it is not surprising that legislative sessions conclude with only a small percentage of bills passing before the clock runs out during a legislative term. In 2021, over 150,000 bills were introduced in the 50 state legislatures but only 21% were enacted. This contrasts with the U.S. Congress, which enacted 317 bills out of the roughly 13,000, or roughly 2.4%, which were introduced during the same timeframe (FiscalNote, 2021). It should be noted that while Congress has typically passed 4–6 million words of new law in each 2-year session, they have passed fewer than 10% of the bills before them over the past 10 years (Govtrack, 2023).

The process of getting legislation passed involves multiple steps with various procedural requirements along the way, including committee hearings and votes, prior to being brought forward to the full chamber where it originated (i.e., house or senate) for a vote. If it passes in that chamber, it goes to the other for review and possible action. If the bill passes in the second chamber, the two versions will be compared and, if there are differences, a conference committee with members from both the house and senate will work to resolve the differences. Once a consensus has been reached, the bill will be voted on again in both chambers. If both chambers pass the bill, it will be sent to the governor for signature.

It is obvious that the bill could fail at any point in the process or the legislative session ends before all the steps can be completed. If a bill does not pass in a particular session, a legislator will often refile the bill and hope to move it further along during the next legislative session. For example, a bill may die in committee during the first session in which it was introduced. The legislator’s goal in the following session may be to simply receive a favorable vote to move it out of the committee where it was assigned. The goal for the following session would be for the bill to come to the floor for a vote by the entire membership. Although virtually all legislators know and understand the procedure and process just described, few would recognize the shaping process that is occurring, although they will be more likely to refile their bill as they overcome each hurdle and each successive approximation is reinforced.

Despite this lack of familiarity with some of the specific procedures and strategies of our profession, more lawmakers are aware of applied behavior analysis and the impact that behavior analysts have on many of their constituents. This can be attributed to two significant public policy changes—mandated insurance benefits for autism spectrum disorder and licensure of behavior analysts.

Insurance-reform efforts, led by groups such as Autism Speaks, successfully advocated for mandated insurance coverage for applied behavior analysis services (Unumb, 2015). Indiana was the first state to pass autism insurance legislation in 2001 and it took until 2019 when Tennessee became the 50^th state to mandate coverage (Bernhard, 2019). Securing coverage in every state is directly attributable to the nationwide advocacy effort and the impact of parents pressuring their legislators to provide this coverage. Although these state insurance mandates were associated with a 16% increase in the number of BCBAs from 2003 to 2017 (McBain et al., 2020), Zhang and Cummings (2020) found that the number of certified ABA providers fell short of the BACB recommended maximum caseload guidelines in 49 states and the District of Columbia, indicating a continuing shortage of providers.

A second factor increasing awareness of applied behavior analysis was the establishment of licensure for behavior analysts. Beginning with Nevada and Oklahoma in 2009, a total of 37 states have enacted licensure laws regulating the practice of behavior analysis (BACB, 2022) and legislation is pending in several other states. This is significant for several reasons. First, licensing is an important consumer protection tool, guaranteeing at least some minimum standards of quality and safety (Farronato et al., 2020). Second, the adoption of continuing education requirements helps to ensure that practitioners maintain competency in their field of practice (Curran et al., 2019). Third, licensure provides a mechanism for addressing complaints and taking disciplinary action against practitioners (Van Horne, 2004). Fourth, the lack of licensure can limit opportunities for reimbursement (Bruni & Lancaster, 2019). In fact, as the establishment of behavior analyst licensure began to occur, some insurers initially attempted to deny reimbursement for applied behavior services if the provider was not licensed, even in states where licensure had not yet been established.

Given some of the challenges and the relative lack of attention historically by behavior analysts, should behavior analysis have a role in public policy and policymaking? From a process perspective, because laws, regulations, and ordinances can be seen as written statements of interlocked contingencies that control individual behavior (Todorov, 2005), it would appear that behavior analysis has a role in shaping public policy. From a personal perspective, we are likely all critical of public policies from time to time and most of us have ideas about how they could be better (Hill & Varone, 2021), suggesting that behavior analysts may also have a role to play as well. In fact, given the increase in behavioral studies and interventions in public policy globally (Straßheim, 2020) and that translational behavioral research will remain a critical component of effective policy development and implementation (Anderson-Carpenter et al., 2023, p. 1), behavior analysts may find that creating policy-relevant conceptual analyses and shaping policy is uniquely reinforcing (Fawcett et al., 1988).

One viable area for future research would be a differential analysis of the effectiveness of the policymaking processes within various state legislatures. For example, given the diversity in composition and ideology, are there different functional IBCs across states or cross-functional IBCs within different state legislatures? Do implicit or overt political biases affect the interpersonal interactions among the members (Ardila-Sánchez et al., 2020a)? Do changes in partisan control change the contingencies and culture of a legislative body? Are there noticeable differences in the IBCs and the interlocked behaviors themselves in the 16 states that have term limits restricting the number of years a legislator can serve? Does seniority affect voting and legislative priorities in term limited versus non-term limited states?

Outside the United States, behavior analysts should study and analyze the structural and cultural differences between the scenarios presented herein and the legislative bodies within their own country. For example, how do the British Parliament and Israeli Knesset differ and do the differences have an impact on legislator voting? Are the Canadian Parliament and provincial legislatures analogous to Congress and our state legislatures? What is the effect of an open or closed list voting system on legislators and the contingencies influencing their voting patterns? Does compulsory voting in various counties affect voters and the actual votes that they cast? The answers to each of these questions will help to identify the relevant metacontingencies in each of these governmental systems.

As noted earlier, approximately 85 countries have introduced gender quotas, which typically bring more women into legislatures and greater attention to the interests and priorities of women as a group (Clayton, 2021). Given the growing body of evidence demonstrating gender differences in legislator responsiveness (Thomsen & Sanders, 2020; Wiener, 2021), information use and processing (Nownes & Freeman, 2019), committee assignment requests (Carroll, 2006) and constituent requests (Butler et al., 2022), how do changes in gender composition affect the IBCs and metacontingencies at work in legislative bodies? What is the impact of other potentially relevant demographic variables (e.g., partisan distribution, first-term versus incumbents) on these metacontingencies?

Likewise, since citizen voting is a requisite antecedent to legislator voting, do legal changes such as automatic voter registration (Morris & Dunphy, 2019), voter identification (Hajnal et al., 2017), reenfranchisement of persons with felony convictions (Morse, 2021) and long lines at the polls (Pettigrew, 2021) only influence turnout? Can they also affect the outcome of the election and, ultimately, the nature of the votes cast during a subsequent legislative session?

Although voters and the legislators they elect are the primary policymakers in our system, behavior analysts should also look to themselves and culturo-behavior science (CBS) as vehicles to advance public policy. In fact, the growing interest in the study of cultural issues can be largely attributed to several initiatives within the profession itself, including the publication of the journal, Behavior and Social Issues, the creation of the Behaviorists for Social Responsibility (BFSR) special interest group (SIG) within ABAI, think tanks on cultural studies, the creation of a culture-behavior science verified higher education course sequence (ABAI, n.d.), and an ABAI specialized conference on cultural matters (Malott, 2021).

These efforts generated interest and support from behavior analysts from around the globe, particularly from social activists and applied researchers who were more interested in solving “real-world problems” than the abstract understanding of the underlying principles of the basic science (Cihon et al., 2021). The research to date has included conceptual analyses as well as empirical studies focusing on a broad range of contemporary issues, including social and environmental justice (Mattaini et al., 2020), global warming (Alavosius & Houmanfar, 2020), environmental sustainability (Gelino et al., 2021), homelessness (Switzer & Rakos, 2022), and violence against women (Amorim et al., 2022). The growing interest in this area should lead to further advances in CBS and may open other pertinent areas for study, such as public health, racism, and gun violence.

Another area where behavior analysts can have an impact is in addressing diversity, equity and inclusion (DEI), both within the field and beyond. Despite the increasing number of consumers with diverse cultural backgrounds (Fong et al., 2017) and the inclusion of cultural responsiveness and diversity in the most recent Ethics Code for Behavior Analysts (BACB, 2020), behavior analysis was “late to the game of cultural change related to diversity and inclusion” (Rehfeldt et al., 2023, p. 3). DataUSA (n.d.) reported that 5,497 degrees were awarded in applied behavior analysis in 2021, reflecting a 438% increase over a 5-year period. Given this expansion, behavior analysts recognized the importance of reaching the next generation of practitioners by promoting cultural awareness and cultural training in graduate programs (Fong et al., 2017) and the need to incorporate DEI within courses in behavior analysis (Cirincione-Ulezi, 2020; Hollins et al., 2023). For example, Hilton et al. (2021) described a series of interventions implemented within an ABA department to improve students’ cultural competency, but the effort ultimately went beyond one academic department. Diversity initiatives became a major priority and over 50 courses are offered across 19 departments at Endicott College, which also instituted a requirement that students to take at least two DEI designated courses during their academic career.

Unfortunately, at the same time that training programs in behavior analysis and other disciplines were seeking to increase student awareness and understanding of DEI, 40 bills were filed in 22 states which would prohibit colleges from having diversity, equity, and inclusion offices or staff; ban mandatory diversity training; prohibit institutions from using diversity statements in hiring and promotion; or prohibit colleges from using race, sex, color, ethnicity, or national origin in admissions or employment (Chronicle of Higher Education, 2023). Although only seven of these bills became law in 2023, one can expect a greater number being filed next year, in light of the U.S. Supreme Court’s recent decision outlawing race-conscious admissions in Students for Fair Admissions, Inc. v. President and Fellows of Harvard College.

Although the focus in this article has been on policymaking within the legislature, the aforementioned Supreme Court decision illustrates how judges in the judicial branch also can make policy, both at the state and federal level. Perhaps because of this power, every year state legislators file “court curbing” bills that threaten to “restrict, remove or otherwise limit the Court’s power” (Clark, 2011, p. 19).

Why do legislators seek to curb the authority of judges? Some attribute it to “activist judges” substituting their personal preferences for the “will of the people” (Ringhand, 2007) when they overturn a prior judicial decision or a law passed by the legislature. Others suggest that court-curbing may simply be driven by discontent with the policy implications of a court’s decision (Bartels & Johnson, 2020).

Despite the fact that state courts decide far more cases than federal courts, most of the research and commentary focused on the U.S. Supreme Court and less is known about judicial decision making among state court judges (Catalano, 2022). Fortunately, there is a growing body of research regarding the volume of court curbing legislation and its impact at the state level.

For example, Leonard’s (2022) analysis of 1,253 court curbing bills filed in state legislatures between 2008 and 2016 revealed that 78% never made it out of committee and appeared to simply be position-taking, enabling legislators to demonstrate to their constituents that they disagree with a judicial decision. Furthermore, the data indicate that electoral retaliation for court curbing is very small, especially after one accounts for partisanship (Driscoll & Nelson, 2021), suggesting that there is little risk for legislators to file a court curbing bill, particularly if it strengthens the support of their constituents.

Because American voters do not appear to punish legislative incumbents for court curbing, would the same be true for “activist judges?” Wenzel et al. (1997) studied judicial activism among state supreme courts and found that the most activist courts were in states where judges were elected and courts in states where judges were appointed were least likely to exhibit activist behavior. This may be an example of the classic “chicken or the egg” metaphor. Are citizens more likely to vote for activist judges or are judges more likely to overrule prior decisions or legislation because of the will of the people? This question, together with an analysis of the various IBCs at work in judicial policymaking, should be addressed in future research.

Recognizing that many readers may have no interest in pursuing research on legislative structures or operations, what can an individual behavior analyst do to affect public policy? There are several possibilities and options. First, you should become a member or become more actively involved with a local professional or trade association (e.g., ABAI affiliate chapter, APBA affiliate, Autism Speaks, Council of Autism Service Providers). Most have a legislative affairs committee where you can learn about relevant bills or proposed changes that could affect you or your organization, such as licensure, reimbursement rates, reciprocity). As relevant bills are filed, you can provide testimony, either in person at a public hearing or by submitting written comments to the legislative committee where the bill has been assigned.

Second, remember that your vote matters. So, be sure that you are registered to vote in your community and take full advantage of this opportunity. Also, if you don’t know who represents you in the state legislature, find out by entering your address at https://www.usa.gov/elected-officials/. Then, reach out to your legislators offices and introduce yourself to them and their staff. You may become a valuable resource to them on behavioral issues and, as a constituent, your input and feedback may influence their votes on issues that are important to you.

Third, if you want to have a direct impact on public policy, then you should consider serving in some capacity, whether on the legislative committee or the board of your local association, applying to be appointed to the board that oversees licensure in your state, or even running for the legislature, especially if your state holds formal sessions for a relatively short time each year. Remember, the state legislature is an important part of our “representative” democracy and who can best educate the members of the legislature or represent the interests of behavior analysts than a behavior analyst? In short, there are a number of options, depending on your availability and level of interest.

Regardless of the road you take to try to influence policy, remember Skinner’s first principle not formally recognized by scientific methodologists: “When you run onto something interesting, drop everything else and study it” (Skinner, 1956, p. 223).

Funding

The author did not receive support from any organization for the submitted work and has no relevant financial or nonfinancial interests to disclose.

Compliance with Ethical Standards

Data Availability

Data sharing is not applicable to this article as no datasets were generated and the work proceeded within a theoretical framework.

Ethics

The work reported in this article did not involve collection of data from participants. Therefore, no institutional review was required, nor was there a need to obtain informed consent.