Introduction
Approximately half of US citizens are employed in manual labor or service-based jobs, yet only 6% of state legislators and 2% of US representatives have previously been employed in a working-class occupation (Carnes Reference Carnes2013). The effects of America’s white-collar government are clear—wealth inequality has dramatically increased in the last half-century with the top 1% of Americans becoming increasingly wealthy while workers’ wage earnings have stagnated. Scholars find that US policy advantages the rich while ignoring the interests of working-class and poor Americans (Bartels Reference Bartels2016; Gilens Reference Gilens2012; Miler Reference Miler2018; Persson and Sundell Reference Persson and Sundell2023).Footnote 1 One potential reason that policy reflects the interests of the rich is the drastic overrepresentation of wealthy Americans serving in political office (Carnes Reference Carnes2013).
The primary explanation for why workers are underrepresented in American legislatures is that structural biases exist in American elections that prevent working-class candidates from emerging and successfully running for elected office (Carnes Reference Carnes2018). While a growing literature examines how and why American elections disproportionately disadvantage candidates from a working-class background (Carnes Reference Carnes2018; Treul and Hansen Reference Treul and Hansen2023), it is also necessary to consider how class-based electoral bias is related to workers’ ability to effectively govern. The current literature suggests that if workers’ numerical representation in legislatures increases, better policy representation for working-class Americans will likely follow (Carnes Reference Carnes2013, Reference Carnes2018; Mansbridge Reference Mansbridge1999). And based on the policy priorities of working-class lawmakers this appears to be true—working-class legislators are more likely than white-collar legislators to introduce and vote for pro-worker policies (Carnes Reference Carnes2013). This logic, however, is contingent on the assumption that workers perform equally or better than white-collar lawmakers once elected to legislative office. If workers are ineffective lawmakers, their policy preferences are unlikely to be successfully legislated into law. In this article, I examine whether a class-based effectiveness gap exists in American legislatures.
Throughout America’s history, some political leaders have advanced arguments suggesting that, if elected, working-class lawmakers would be ineffective at carrying out the duties and responsibilities of a legislator. Alexander Hamilton, writing in the Federalist Papers, suggests that workers are less politically skilled than individuals working in white-collar jobs and are, therefore, less suitable for political office (Hamilton et al. Reference Hamilton1788). More recently, President Trump publicly stated in reference to his selection of cabinet secretaries that he preferred to appoint rich cabinet secretaries because he “just didn’t want a poor person” in the position (Calfas Reference Calfas2017). If these arguments are correct and white-collar lawmakers are better suited to govern, we may expect working-class legislators to be less effective lawmakers than their white-collar colleagues. To date, however, there is no empirical evidence suggesting that working-class legislators are less effective lawmakers than white-collar legislators.
I argue, in contrast, that working-class legislators should perform equally or better than white-collar lawmakers. I develop a theory of class-based electoral selection that links class-based discrimination in elections to legislators’ performance in office (Anzia and Berry Reference Anzia and Berry2011). Working-class candidates face class-based biases in elections that make it more difficult to emerge and successfully win elective office (Carnes Reference Carnes2018). Given that white-collar candidates do not face similar biases in elections, working-class candidates must work harder and develop skills to overcome these barriers. As a result, the workers who do win elected office are qualified, hard-working candidates capable of effective lawmaking.
My data set pairs pre-legislature occupational data for more than 14,000 unique state legislators (Makse Reference Makse2019) with Bucchianeri, Volden, and Wiseman’s (Reference Bucchianeri, Volden and Wiseman2024) state legislative effectiveness scores (SLES). The data set in total includes more than 50,000 legislator-term specific observations from 49 states for more than 30 years (1988–2017). Of these 50,000 legislator-term specific observations, approximately 3,500 are working-class legislators. A simple comparison of means shows that, on average, white-collar legislators are marginally more effective than working-class legislators. However, in a regression model with controls, this relationship disappears, suggesting that working-class and white-collar legislators are equally effective lawmakers.
Studying legislators’ performance in office, particularly for underrepresented groups, is necessary for several reasons. First, empirically evaluating the legislative effectiveness of underrepresented groups empowers scholars to address discriminatory arguments that these groups are in some way less capable than majority groups. My analysis provides some evidence against these discriminatory arguments; I find no evidence that white-collar legislators are more effective lawmakers than workers. Indeed, my findings stand in contrast to arguments that suggest workers are less suitable for political office because of their occupational background (Hamilton et al. Reference Hamilton1788). Second, examining the effectiveness of working-class lawmakers may have implications for the substantive representation of working-class Americans. Though I do not directly test whether working-class lawmakers are more likely to pass legislation that would benefit working-class constituents, the fact that workers are equally as effective as white-collar lawmakers certainly suggests that workers have the lawmaking skills necessary to substantively represent working-class Americans.
The legislative effectiveness of underrepresented groups
There is an extensive literature that seeks to conceptualize, measure, and analyze legislative effectiveness in the US Congress and US state legislatures (Bucchianeri, Volden, and Wiseman Reference Bucchianeri, Volden and Wiseman2024; Hitt, Volden, and Wiseman Reference Hitt, Volden and Wiseman2017; Matthews Reference Matthews1959; Stacy Reference Stacy2020; Volden and Wiseman Reference Volden and Wiseman2014, Reference Volden and Wiseman2018; Volden, Wiseman, and Wittmer Reference Volden, Wiseman and Wittmer2013; Weissert Reference Weissert1991). Volden and Wiseman define legislative effectiveness as “the proven ability to advance a member’s agenda items through the legislative process and into law” (Volden and Wiseman Reference Volden and Wiseman2014, p. 18). Bucchianeri, Volden, and Wiseman’s (Reference Bucchianeri, Volden and Wiseman2024) SLES measure legislative action throughout the lawmaking process (sponsorship, action in committee, action beyond committee, a bill passing one chamber, and a bill becoming law). This measure comprehensively describes legislators’ lawmaking efforts at each stage of the lawmaking process. Legislative effectiveness scores are used to analyze institutional and individual-level factors, and how the intersection of both, shapes legislators’ effectiveness. Scholars have primarily analyzed the legislative effectiveness of two underrepresented groups—women and Black legislators—in the US Congress (Volden and Wiseman Reference Volden and Wiseman2014; Volden, Wiseman, and Wittmer Reference Volden, Wiseman and Wittmer2013).
Volden and Wiseman (Reference Volden and Wiseman2014) find that women legislators, when in the minority party, are more effective than male legislators. In the majority party, however, women are equally effective as male legislators (Volden, Wiseman, and Wittmer Reference Volden, Wiseman and Wittmer2013). Women legislators are particularly effective at the consensus-building portions of the lawmaking process such as committee and floor action (Volden and Wiseman Reference Volden and Wiseman2014; Volden, Wiseman, and Wittmer Reference Volden, Wiseman and Wittmer2013). Volden et al. attribute women’s increased effectiveness at consensus-building stages of the legislative process to behavioral differences between genders—women are more collaborative than their male colleagues (Volden, Wiseman, and Wittmer Reference Volden, Wiseman and Wittmer2013).Footnote 2
Black legislators are less effective than White legislators when Democrats are in the majority party. However, they are equally as effective as White legislators when Democrats are in the minority party (Volden and Wiseman Reference Volden and Wiseman2014). Volden and Wiseman (Reference Volden and Wiseman2014) theorize that this is a result of Black legislators developing a more specialized legislative agenda. Existing scholarship on the legislative effectiveness of women and Black legislators suggests that underrepresented groups are not uniformly less effective than majority groups. Instead, their effectiveness is uniquely shaped by the interaction between legislative institutions and their descriptive identities. Therefore, while considering the legislative effectiveness of women and Black legislators may be theoretically useful, workers’ effectiveness will likely be, in part, unique.
There is limited scholarship directly analyzing the relationship between legislators’ social class backgrounds and their performance in legislatures. While no existing work uses legislative effectiveness scores (Bucchianeri, Volden, and Wiseman Reference Bucchianeri, Volden and Wiseman2024; Volden and Wiseman Reference Volden and Wiseman2014) to analyze how legislators’ social class backgrounds are related to their legislative effectiveness, scholars have used other measures of legislative productivity. The findings are mixed among the existing literature but broadly suggest that working-class legislators are not uniformly less effective than white-collar legislators. Carnes, in his book White-Collar Government (Carnes Reference Carnes2013), examines the legislative entrepreneurship of workers in the US Congress in the context of economic policy. He finds that workers sponsor and cosponsor more economic legislation than white-collar legislators, and pass economic policy at equal rates as their white-collar colleagues.Footnote 3 Likewise, Carnes and Lupu (Reference Carnes and Lupu2016b) examine the relationship between legislators’ educational backgrounds and their performance in office and find that legislators pass the same number of bills regardless of their educational background. Finally, in an examination of underrepresented groups in state legislative leadership positions, Hansen and Clark (Reference Hansen and Clark2020) find that workers are equally as likely to be represented in state leadership positions as white-collar legislators. While state legislative leadership positions are not a direct test of legislators’ effectiveness, existing work suggests that legislative leaders are among the most effective lawmakers (Volden and Wiseman Reference Volden and Wiseman2014).
I provide a more robust analysis of class-based legislative effectiveness in two ways. First, using SLES allows me to define legislative effectiveness as a lawmaker’s effectiveness at all stages of the legislative process rather than only sponsorship, cosponsorship, or final passage votes. Importantly, this allows me to observe workers’ actions at less visible stages of the lawmaking process such as actions in committee and on the floor. Second, I analyze state legislators rather than US representatives, where only 2% of representatives have previously been employed in working-class occupations (Carnes Reference Carnes2013). I conduct my analyses at the state legislature level because the percentage of working-class representation is higher than in the US Congress (6% rather than 2%). Additionally, given that there are far more state legislators than US representatives, the raw number of working-class representatives is much higher, offering more variation in levels of representation across time to examine.
Class-based electoral bias and legislative effectiveness
I argue that working-class lawmakers are equally or more effective than white-collar lawmakers due to class-based electoral selection effects. Building on existing work, I develop a theory of class-based electoral selection that links class-based discrimination in elections to legislators’ performance in office (Anzia and Berry Reference Anzia and Berry2011). This theory suggests that because workers face prejudice and discrimination during legislative elections as a result of their class identity, they are forced to work harder and develop the necessary skills to win elections. As a result, the working-class candidates who win elections become effective lawmakers.Footnote 4
Electoral selection effects occur when political candidates face biases in elections as a result of their identity (Anzia and Berry Reference Anzia and Berry2011; Ashworth, Berry, and de Mesquita Reference Ashworth, Berry and de Mesquita2024). Existing research suggests that women, non-white, working-class, and LGBTQ+ candidates disproportionately face electoral obstacles that increase the difficulty of winning elections (Anzia and Berry Reference Anzia and Berry2011; Bateson Reference Bateson2020; Carnes Reference Carnes2018; Piston Reference Piston2010; Wagner Reference Wagner2019). As a result, candidates from underrepresented groups must work harder than majority groups to win elections. There are at least three causes of electoral selection effects (Anzia and Berry Reference Anzia and Berry2011; Ashworth, Berry, and de Mesquita Reference Ashworth, Berry and de Mesquita2024). First, voters may be biased toward a given social group, requiring that candidates who identify with these groups be exceptionally qualified to secure electoral support. Second, candidates may perceive that voters are biased against them, even if they are not. If this is the case, only the most ambitious and qualified candidates will emerge and enter the electoral arena. Third, political elites and gatekeepers may be biased against certain social groups, forcing candidates who identify with these groups to work harder to gain elite support during their campaigns.
Existing scholarship suggests that the type of selection effects candidates face varies across identity groups.Footnote 5 Prejudice and discrimination from political elites and gatekeepers—rather than voter bias or perceived bias—is likely the primary obstacle preventing working-class candidates from winning elections (Carnes and Lupu Reference Carnes and Lupu2016a; Carnes Reference Carnes2018; Griffin, Newman, and Buhr Reference Griffin, Newman and Buhr2020; Hoyt and DeShields Reference Hoyt and DeShields2021; Treul and Hansen Reference Treul and Hansen2023). Evidence from survey experiments suggests that voters are not biased against working-class candidates relative to white-collar candidates. Carnes and Lupu (Reference Carnes and Lupu2016a) find that respondents in the United States, Britain, and Argentina viewed hypothetical working-class candidates as “equally qualified, more relatable, and just as likely to get their votes” (Carnes and Lupu Reference Carnes and Lupu2016a, p. 832). Other work suggests that voters rate workers as more “warm” relative to white-collar candidates (Hoyt and DeShields Reference Hoyt and DeShields2021). On the other hand, voters rate white-collar candidates as less honest and less caring than working-class candidates (Griffin, Newman, and Buhr Reference Griffin, Newman and Buhr2020). Likewise, working-class candidates do not see themselves as unqualified for elected office. When asked whether they had ever thought about running for elected office, working-class respondents reported a similar level of political ambition as white-collar respondents (Carnes Reference Carnes2018). Similarly, working-class Americans are equally as likely to feel qualified to run for office as white-collar Americans (Carnes Reference Carnes2018).
Workers do experience discrimination during campaigns from political elites and electoral gatekeepers. Carnes (Reference Carnes2018) finds that party leaders view the working class as less viable political candidates, often citing their difficulty to fundraise and win elections (Carnes Reference Carnes2018, p. 110). As a result, party leaders are less likely to recruit and support working-class candidates in legislative races. Relatedly, without the financial support of party leaders, working-class candidates struggle to fundraise in elections.Footnote 6 This dual resource and recruitment burden makes it extremely difficult for workers to enter the electoral arena, and even more difficult to win the race.
As a result of class-based discrimination, working-class candidates and white-collar candidates experience a very different electoral environment (Carnes Reference Carnes2018). Working-class candidates must work harder than white-collar candidates to receive the same electoral outcome (Carnes Reference Carnes2018). In doing so, I argue that workers develop and refine skills that promote effective lawmaking once in office (Anzia and Berry Reference Anzia and Berry2011). To date, there is little work directly testing mechanisms that may explain why electoral selection effects produce effective lawmakers (Anzia and Berry Reference Anzia and Berry2011). I suggest two plausible mechanisms that may explain why class-based electoral selection produces effective working-class lawmakers.
First, one way class-based electoral selection may produce effective lawmakers is that less qualified working-class candidates will lose elections. Working-class lawmakers who do not overcome class-based electoral barriers—whether it be because they were not qualified candidates or because the electoral barriers were insurmountable—will lose their election. Given that white-collar candidates face fewer electoral obstacles than working-class candidates, they may be more likely to win elections even if they are less qualified than working-class candidates. Workers’ performance in primary versus general elections lend some evidence in support of this expectation. Treul and Hansen (Reference Treul and Hansen2023) find that workers underperform white-collar candidates in primary elections; however, workers are equally as likely as white-collar candidates to win general elections (Carnes Reference Carnes2018). One explanation for this seemingly inconsistent trend may be that less qualified working-class candidates are weeded out during primary elections, while qualified workers perform equally as well as their white-collar opponents in general elections. This selection process results in only the most effective working-class candidates winning elections and gaining representation in legislatures.
Second, in addition to hollowing out the working-class candidate pool, I expect class-based biases in elections to meaningfully influence the working-class candidates who do win elections. The working-class candidates who do successfully win elections have experienced the effects of class bias in elections, and are aware of the effort that is required to outperform their white-collar colleagues. There is reason to expect that workers will carry this over-performance mindset with them as they begin working in legislatures. Put differently, class-based biases exist within American political institutions, and so similar to the electoral stage of the political process, workers must work harder than their white-collar colleagues in legislatures to accomplish their goals (Carnes Reference Carnes2013, Reference Carnes2018).
Working-class candidates face electoral biases that white-collar candidates do not face. The result of this electoral selection effect is that only the most qualified and capable workers win elected office and gain representation in legislatures. As a result, I hypothesize that working-class lawmakers should be equally or more qualified than their white-collar colleagues.
H1 (Class-Based Legislative Effectiveness): Workers are equally or more effective lawmakers than white-collar legislators.
Data and measurement
To test this hypothesis, I pair pre-legislature occupational data for more than 14,000 unique state legislators (Makse Reference Makse2019) with Bucchianeri, Volden, and Wiseman’s (Reference Bucchianeri, Volden and Wiseman2024) SLES. The data set includes SLES for 51,929 legislator-term-specific observations for 49 states from 1987–2017.Footnote 7 Of these observations, 3,572 (or 6.8% of my sample) were previously employed in a working-class occupation.Footnote 8
SLES are constructed similarly to legislative effective scores (LES) used to measure effectiveness in the US Congress (Volden and Wiseman Reference Volden and Wiseman2014; Volden, Wiseman, and Wittmer Reference Volden, Wiseman and Wittmer2013). SLES, like LES, captures the weighted average of a legislator’s actions throughout five stages of the lawmaking process: bill introduction, action in committee (AIC), action beyond committee (ABC), passing one chamber (PASS), and becoming law (LAW) (Bucchianeri, Volden, and Wiseman Reference Bucchianeri, Volden and Wiseman2024). Therefore, these scores evaluate effectiveness throughout the entirety of the legislative process rather than simply analyzing roll-call votes. Additionally, SLES are weighted to reflect the substance and significance of legislation. Commemorative and symbolic legislation influences a legislator’s effectiveness score less than substantive and significant legislation. Bucchinaeri, Volden, and Wiseman (Reference Bucchianeri, Volden and Wiseman2024) calculated SLES by scraping the legislative history of every bill available on state legislative websites. Bill data are available for some states (Maine, South Carolina, New Hampshire, Texas, and Pennsylvania) dating back to the 1980s. The legislative history of every bill for every state (except Kansas) is included in the data set after 2003.Footnote 9
To operationalize social class, I use pre-legislature occupational data (Makse Reference Makse2019). I consider legislators to be working-class if their most recent pre-legislature occupation was in construction, office or clerical work, public safety, retail and service, a skilled trade, or as semi-skilled or unskilled laborers. My definition of workers most closely resembles Makse’s (Reference Makse2019) definition; however, unlike Makse, I do not consider “transportation professions” which include pilots, railroad engineers, and air traffic controllers to be working-class. My definition of working class differs from Carnes’ (Reference Carnes2013) definition in that I consider contractors engaged in blue-collar work and public safety professionals to be working-class. I use the occupational backgrounds of legislators to operationalize social class because it is arguably the best predictor of individuals’ income and social status (Matthews Reference Matthews1954; Hout Reference Hout2008, cited in Carnes Reference Carnes2012), and it has become convention in the study of social class (Barnes, Beall, and Holman Reference Barnes, Beall and Holman2021; Carnes Reference Carnes2013; Makse Reference Makse2019; Nickelson and Jansa Reference Nickelson and Jansa2023). A complete list of working-class and white-collar occupations can be found in the Supplementary Appendix (A.1).
One downside of Makse’s (Reference Makse2019) occupational data is that lawmakers’ occupational histories are limited to their most recent pre-legislature occupation. If a lawmaker’s most recent occupation before being elected is in real estate, they are coded as white-collar. Conversely, if a lawmaker’s most recent occupation prior to being elected is a retail worker, they are coded as working-class. This data cannot distinguish legislators who worked in a working-class occupation prior to working in a white-collar occupation. For example, a legislator who worked as a retail worker for 5 years before transitioning into a job in real estate is coded as white-collar. Ideally, I would have complete occupational histories for every state legislator. I would then be able to analyze how legislators’ occupational histories influenced their effectiveness within legislatures. Perhaps legislators who entered the workforce as working-class but transitioned into a white-collar job have a higher or lower effectiveness score than legislators who entered the workforce in a working-class job and remained in a working-class job until their election.
Unfortunately, a data set including the occupational histories of state legislators from all state legislatures across the time series of my data does not exist.Footnote 10 While this is a limitation within my data, I argue that defining a legislators’ social class by their most recent pre-legislature occupation is a reasonable test for my theory. I argue that working-class lawmakers should be equally or more effective than white-collar lawmakers because of class-based biases in elections. These biases will likely be most pronounced for working-class lawmakers currently employed in a working-class occupation during their campaign (rather than a formerly working-class candidate employed in a white-collar occupation during their campaign). For example, given that party leaders discriminate against the working class when recruiting potential candidates, this discrimination will be most pronounced for working-class candidates who are currently employed in a working-class occupation rather than candidates who were previously employed in a working-class occupation. Thus, if class-based electoral selection effects influence working-class and white-collar legislators’ effectiveness differently, this effect will be most pronounced for candidates whose most recent occupation was a working-class job. Therefore, while operationalizing legislators’ social class as their most recent pre-legislature occupation does not comprehensively describe their occupational history, it is a reasonable test for my theory.
I condition on several covariates that likely influence legislators’ effectiveness. First, I control for the percentage of working-class legislators within a legislature term to ensure that the estimated relationship is not dependent on the proportion of workers represented within a legislature. Second, given that legislators hold multiple social group identities that could confound the relationship between social class and legislative effectiveness, I control for demographic covariates such as race, gender, and party identification.Footnote 11 Third, I include chamber-specific covariates that differ between legislators, such as seniority, vote share, majority party status, governor’s party, leadership positions, and polarization. Fourth, I control for covariates that differ across state legislatures such as professional and term limits. Finally, I include state and term fixed effects to control for variation specific to each state legislature and term.Footnote 12
Are workers effective lawmakers?
I first analyze the mean effectiveness score of white-collar and working-class state legislators. Figure 1 plots the mean SLES for both white-collar and working-class legislators. White-collar legislators, on average, have a mean SLES of .01. Working-class legislators, on average, have a mean SLES of −.02. After plotting the average effectiveness score for both white-collar and working-class legislators against the entire range of the dependent variable (−3 to 9), it becomes clear that a class-based effectiveness gap of 0.03 in the raw data is substantively small.
Figure 1. Average SLES of Working-Class and White-Collar Legislators.
The substantively small difference in means between working-class and white-collar legislators’ effectiveness entirely disappears in a regression model. I estimate ordinary least square (OLS) regression models with “Worker” as the independent variable and “SLES” as the dependent variable. As column 6 in Table 1 shows, all else equal, workers are, on average, 0.028 times less effective than white-collar legislators. However, the magnitude of the relationship is small and not statistically significant.Footnote 13 The dependent variable ranges from −2.9 to 9.9, indicating that an effectiveness gap of 0.028 is substantively small. Put differently, the difference in working-class and white-collar lawmakers’ effectiveness is approximately 2.8% of a standard deviation.
Table 1. Working-class and white-collar legislators are equally effective lawmakers
Notes: t statistics in parentheses.
<*>
$ p<0.05 $
, <**>
$ p<0.01 $
, <***>
$ p<0.001 $
.
I also estimate workers’ effectiveness at each stage of the lawmaking process. Table 1 (columns 1–5) shows that the worker coefficient is positive, though small in magnitude and not statistically significant at each of the five stages of the lawmaking process. Figure 2 displays the effectiveness gap between working-class and white-collar legislators at each stage of the lawmaking process—bill introduction, action in committee (AIC), action beyond committee (ABC), passing chamber (PASS), and becoming law (LAW). A point estimate greater than zero indicates that workers are more effective than white-collar legislators. Likewise, a point estimate below zero indicates that white-collar lawmakers are more effective than working-class legislators. To observe any variation away from zero, the Y-axis must be set to a substantively small range (0.0025 to −0.0025) of the dependent variable, suggesting no meaningful difference between working-class and white-collar legislators’ effectiveness. Importantly, the absence of a class-based effectiveness gap is independent of workers’ numerical representation in the legislature. Working-class lawmakers are equally as effective as white-collar lawmakers regardless of the percentage of workers represented in the legislature. This relationship also holds in various state legislative institutional arrangements.Footnote 14 The evidence from Figure 2 and Table 1 is consistent with my expectation that workers are no less effective than white-collar lawmakers.
Figure 2. Working-Class and White-Collar Legislators Are Equally Effective Lawmakers.
A relationship between a set of variables that is not statistically significant, however, does not necessarily have a negligible effect (Rainey Reference Rainey2014). A regression coefficient may be statistically indistinguishable from zero for reasons other than the absence of a relationship between a set of variables. For example, a small sample size can result in large error estimates that might make a large coefficient not statistically significant (Rainey Reference Rainey2014). To ensure that the effectiveness gap between working-class and white-collar lawmakers is indeed negligible, I follow the advice of Rainey (Reference Rainey2014) and (1) define a contextually specific negligible legislative effectiveness score (−m and m) and (2) use a 90% confidence interval to examine whether the estimated confidence interval falls within the zone of negligibility (−m and m).Footnote 15
I define the zone of negligibility as an effect that ranges between −0.075 and 0.075. It is important to carefully define how I selected this range. Rainey (Reference Rainey2014) advises that scholars define negligibility in a way that is contextually specific to their data. This means that scholars are tasked with evaluating their data and determining at what level effects are no longer substantively meaningful (Rainey Reference Rainey2014). The SLES variable ranges from −2.9 to 9.9. Therefore, an effect (and error estimates) that falls within the range of −0.075 and 0.075 is only 15% of a standard deviation. I argue that an effect that falls within this range does not meaningfully explain any variation in legislators’ effectiveness. Put differently, if two lawmakers’ SLES differ by only 0.075, on average, their actions in all five stages of the lawmaking process look very similar. I plot estimates and their confidence intervals and analyze whether they fall within this zone (−0.075 and 0.075). If the 90% confidence intervals fall within the zone of negligibility, this suggests that the null results indicate a negligible effect (Rainey Reference Rainey2014).
Figures 3 and 4 plot the estimated relationship between legislators’ class backgrounds and their effectiveness using clustered standard errors, bootstrapped standard errors, and median regression.Footnote 16 The solid black line labeled “estimate” is the estimated relationship in row one, column six of Table 1. Given that state legislative data is particularly likely to have clustered groups and heavy-tailed distributions, I replicate my results using clustered and bootstrapped standard errors and median regression. I use clustered standard errors and bootstrapped standard errors to ensure that the grouped nature of the data does not produce unmodeled correlations that result in a downward bias in standard error estimates (Harden Reference Harden2011). Additionally, I replicate the OLS results using median regression to ensure that the heavy-tailed error term does not produce inefficient estimates (Harden and Desmarais Reference Harden and Desmarais2011).
Figure 3. Negligible Class-Based Effectiveness Gap.
Figure 3 shows that all four estimates and confidence intervals are similar in magnitude and fall within the zone of negligibility. This means that, across four different models, the 90% confidence intervals only include estimates within the zone of negligibility. To better contextualize the negligible relationship between legislators’ class background and their effectiveness, Figure 4 plots the same data as Figure 3 over the entire range of the SLES variable. Figures 3 and 4 collectively show that the absence of a class-based effectiveness gap is precisely estimated and negligible when considering the range of the dependent variable.
Figure 4. Negligible Class-Based Effectiveness Gap Across Entire Range of SLES.
To further clarify the precision of the null effect, the lower bound of the 90% confidence interval for the worker coefficient is less than 1% of the range of the SLES variable. Even more striking, the largest possible value of the worker coefficient is less than 1% of the committee chair coefficient, which is the largest coefficient observed in the model. Therefore, consistent with the findings in Table 1, I conclude that there is no meaningful gap between working-class and white-collar legislators’ effectiveness in my sample. These findings support my expectation that workers will be no less effective lawmakers than white-collar legislators.
Conclusion
Political elites have long argued that, if elected, working-class politicians would be less effective at governing than white-collar politicians (Hamilton et al. Reference Hamilton1788; Calfas Reference Calfas2017). I analyze the legislative effectiveness of 14,000 state legislators and find no evidence that working-class legislators are less effective lawmakers than white-collar legislators. Indeed, I provide evidence of a negligible relationship between working-class and white-collar legislators’ effectiveness. White-collar and working-class legislators are equally effective throughout each stage of the lawmaking process.
I argue that workers perform equally as well as white-collar legislators once in office because they face class-based discrimination in elections. Class-based electoral biases create incentives for working-class candidates to work harder—both in terms of effort and skill development—than white-collar candidates. As a result, less qualified working-class candidates lose elections and the workers who do win elective office are effective lawmakers.
Future research should continue to explore the legislative behavior of working-class lawmakers. Two future directions are particularly relevant in light of the findings reported in this paper. First, scholars should examine the policy areas in which working-class lawmakers are most effective.Footnote 17 If workers prioritize labor and economic policy, we can be more confident that the descriptive representation of working-class lawmakers leads to the substantive representation of workers’ policy preferences (Carnes Reference Carnes2013). Second, future work should consider whether workers’ effective lawmaking is related to their performance in elections. If effective working-class lawmakers are electorally rewarded for their legislative performance, this may suggest that working-class lawmakers’ performance in office is not a primary cause of the numeric underrepresentation of working-class lawmakers.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/spq.2024.8.
Data availability statement
Replication materials are available on SPPQ Dataverse at https://doi.org/10.15139/S3/WFVTCW (Lollis Reference Lollis2024).
Acknowledgments
The author thanks Craig Volden, Alan Wiseman, and the Center for Effective Lawmaking for sharing data. I also thank Todd Makse for sharing his data on lawmakers’ professional backgrounds. Justin Kirkland and Craig Volden provided helpful comments on drafts of this manuscript.
Funding statement
The author received no financial support for the research, authorship, and/or publication of this article.
Competing interest
The author declares no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Author Biographies
Jacob M. Lollis is a Ph.D. Candidate at the University of Virginia.
Open access
