1. Introduction
Scholars have noted the high degree of political homogeneity within the Black community—namely the high rates of Democratic Party identification among African Americans (Dawson, Reference Dawson1994). This political homogeneity, however, masks some of the political heterogeneity that exists among African Americans, including which issues make it onto the Black political agenda.Footnote 2 Previous research indicates that not all Black interests are deemed worthy of representation, as issues associated with the more marginalized members of the Black community tend to be ignored (Cohen, Reference Cohen1999; Harris, Reference Harris2012; Spence, Reference Spence2015; Lopez Bunyasi and Smith, Reference Lopez Bunyasi and Smith2019). Thus, this study offers an empirical test of whether Black state legislators are less likely to be responsive to a constituent affiliated with the more marginalized elements of the Black community. Specifically, we fielded an experiment to test whether an email associated with the National Association for the Advancement of Colored People (NAACP), would on average, have a higher response rate than an email related to Black Lives Matter (BLM), a movement associated with more marginalized African Americans. The results indicate that Black state legislators were, on average, as responsive to emails that referenced the NAACP, as those emails that referenced BLM. To our knowledge, this is the first field experiment to test for intragroup discrimination within the Black community.
Previous research has tested for and found evidence of intergroup discrimination by White legislators against Black constituents (Butler and Broockman, Reference Butler and Broockman2011). Also, studies focusing on intragroup dynamics have found evidence of a “qualified” linked-fate politics, whereby not all Black interests are deemed worthy of representation (Cohen, Reference Cohen1999; White, Reference White2007; Jefferson, Reference Jefferson2019; Lopez Bunyasi and Smith, Reference Lopez Bunyasi and Smith2019). There is evidence that some groups within the Black community are ignored altogether as mainstream Black public opinion shifts rightward (Tate, Reference Tate2010). Yet, there is also evidence that Black politicians might be more responsive to their Black constituents (Butler and Broockman, Reference Butler and Broockman2011) and are intrinsically motivated to represent Black interests (Broockman, Reference Broockman2013). This research, however, does not take into consideration that Black politicians might not represent all Black interests equally—building upon previous research, our field experiment tests for evidence of intragroup discrimination by Black state legislators.
Specifically, we test whether Black state legislators are more responsive to a constituent associated with the NAACP, an organization associated with more mainstream political tactics, relative to a Black constituent inspired by BLM, an organization associated with more confrontational, extra-institutional tactics. We hypothesize that Black politicians may have more of an affinity toward Black organizations that are perceived as working within the system—a system to which Black legislators belong. While our study does not find evidence of intragroup discrimination, it still makes a valuable contribution to the understanding of Black intragroup politics by demonstrating that in this instance, Black legislators were equally responsive to emails inspired by the NAACP versus emails inspired by BLM.
2. Black political representation
Black state legislators and Black political organizations are necessary vehicles through which the Black agenda reaches the larger policy space. As Burden (Reference Burden2007) illustrates, politicians are more likely to advance the interests of constituents with whom they share personal characteristics. Previous research also indicates that minority legislators advance the interests of racial and ethnic minority groups more than White members of their legislatures, via responsiveness to constituents, congressional roll call votes, and racial oversight hearings (Whitby, Reference Whitby1997; Butler and Broockman, Reference Butler and Broockman2011; Minta, Reference Minta2011; Broockman, Reference Broockman2013). Butler and Broockman (Reference Butler and Broockman2011), for example, found that constituent requests about voting received fewer responses from White legislators when the email was sent from a putatively Black alias as opposed to a putatively White alias. Furthermore, even when the email signaled the partisanship of the constituent, White legislators, Republicans, and Democrats alike, were less responsive to constituents with putatively Black names. The lack of responsiveness across party lines suggests that strategic partisan considerations were not at play, but rather, some White legislators might prefer to engage in discriminatory behavior. Conversely, Black legislators were found to be more responsive to the Black alias, which might be indicative of some form of racial group solidarity among African Americans.
Race not only affects the type of interests that legislators advance but also the intensity with which they pursue these interests on the floor of the legislature (Hall, Reference Hall1996). Broockman (Reference Broockman2013), for example, found that Black legislators were more motivated to represent the interests of African Americans, who resided both inside and outside of legislators' districts. Specifically, he found that while Black and non-Black politicians respond to an in-district Black alias seeking governmental assistance at approximately the same rate, Black politicians responded to out-of-district Blacks at a 20-point higher rate than non-Black politicians. His results indicate that even when electoral incentives are weakened, as indicated by the sender's location, the intrinsic motivation of Black legislators causes them to engage very differently with their Black constituents than their White counterparts. Ultimately, Black legislators not only exert more deliberate effort to advance Black interests, but arguably their efforts result in more substantive outcomes for Black Americans (although see Swain, Reference Swain1993).
Yet, despite previous studies indicating greater responsiveness on the part of Black legislators to their Black constituents (Butler and Broockman, Reference Butler and Broockman2011; Broockman, Reference Broockman2013), these aggregate results could potentially be masking some intragroup differences. There are circumstances under which Black legislators may not be motivated to represent the interests of all Black constituents equally. Bias does not always manifest itself as differential treatment between racial groups but can also occur within racial groups. Thus, bias can be exhibited within a racial group such that relatively advantaged members of a marginalized group can further marginalize less advantaged members of the marginalized group in question through a process that Cohen (Reference Cohen1999) refers to as “secondary marginalization.”
Secondary marginalization entails relatively advantaged members of a marginalized group regulating the behavior, attitudes, and access to resources of less advantaged members of the marginalized group. Therefore, the racial group solidarity that was exhibited in the 2013 Broockman study, for example, may not be extended to all segments of the African American community, particularly those who are less advantaged, such as Black people who are queer, those with criminal histories, or are otherwise disadvantaged on other axes of marginalization, that might be deemed less “respectable.” Since the BLM movement was spearheaded by three queer, millennial women, who explicitly reject respectability politics, we hypothesize that Black legislators might limit the access to resources of Black people who are affiliated with BLM.Footnote 3
Furthermore, as (Smith, Reference Smith1996) highlights, the political incorporation of Black leaders—the shift from protest to politics—has had many negative consequences for Black Americans seeking to advance their interests in the larger political system. He states that “the results of incorporation are that Blacks have lost the capacity to effectively press their demands on the system and that the system has consequently responded to their demands with symbolism, neglect, and an ongoing pattern of cooptation” (Smith, Reference Smith1996, 19). Smith (Reference Smith1996) underscores that the integration of Black leaders into government positions may have actually weakened the ability of Black politicians to address problems of race in the post-civil rights era. In short, Black politicians may not be able to address “Black interests” in the post-civil rights era, because they are a part of the very governmental system that Black activists and communities are challenging, often through Black political organizations.
Ironically, many Black politicians made their entrance into politics through participation in the Black organizations that have historically played an important role in advancing Black political interests, including the National Urban League, the NAACP, and the Black church. Some of these organizations have endorsed candidates, mobilized voters, sponsored extensive advertising campaigns, and otherwise have done the work that party committees perform in partisan races. Yet, not all Black organizations are likely to be perceived equally. Even though both the NAACP and BLM, for example, were founded in response to race-based violence, for example,—the NAACP in 1909 after a deadly race riot in Springfield, IL, and BLM in 2014 after a spate of shootings of unarmed Black civilians—we suspect that Black legislators might react differently to the NAACP than they do to BLM because of differences in the organizations' respective histories and approaches to politics.
The NAACP, for example, has gained prominence by working within the system, waging legislative battles that have sometimes spanned decades. Conversely, BLM has relied on more extra-institutional tactics, such as “die-ins” and protests that were able to be quickly organized via online mass mobilization. Thus, Black state legislators may privilege the NAACP, an organization that has a long history of working within the system. In fact, political elites who have been critical of the BLM movement have suggested that the BLM movement would be more successful if it emulated earlier African American movements that did not engage in radical confrontation (Tillery, Reference Tillery2019).
Differences in the leadership structure of the two organizations might also drive Black legislators to treat constituents associated with BLM differently than constituents associated with the NAACP. BLM, for example, characterizes itself as a “leader-full” organization which means that leadership is decentralized, whereas the NAACP has a more hierarchical, top-down structure, characterized by a central leader or president. Black legislators may prefer to work with a group that has a clearly identified leader and power broker. In fact, some Black legislators have even simultaneously held leadership positions in government and the NAACP, with Kwesi Mfume being arguably the most prominent example.Footnote 4
In addition, the NAACP routinely endorses political candidates, as opposed to BLM, an organization that has been known to openly protest and confront elected officials. The BLM network, for example, refused to endorse any 2016 presidential candidate, instead pledging to continue with protests and interruptions during the campaign season. According to co-founder, Alicia Garza, “Sometimes you have to put a wrench in the gears to get people to listen,” which stands in stark contrast to the NAACP, who endorsed Secretary of State Hillary Clinton, the establishment candidate in the 2016 presidential race. Black legislators may prefer to work with an organization that might mobilize its members to vote for them via an endorsement since previous research indicates that endorsements help Black candidates secure electoral victory (Benjamin, Reference Benjamin2017).
The express purpose of BLM is to emphasize that all BLM, regardless of gender, sexuality, immigration status, style of dress, or previous encounters with the criminal justice system. In other words, the lives of Black people whose social performances do not align with predefined norms of civility matter, and not just those Black people whose behavior comports with dominant norms (Obasogie and Newman, Reference Obasogie and Newman2016). According to the BLM website, “We [BLM] are expansive. We are a collective of liberators who believe in an inclusive and spacious movement. We also believe that in order to win and bring as many people with us along the way, we must move beyond the narrow nationalism that is all too prevalent in Black communities. We must ensure we are building a movement that brings all of us to the front [italics added].Footnote 5”
Thus, BLM is offering a critique of some of the dominant institutions in the Black community, arguing that many spaces within Black communities are not inclusive of all Black people. They do not explicitly name any institutions, but it is plausible that BLM is critiquing institutions such as the Black church, and civil rights organizations, including the NAACP. Previous research has shown these institutions to be less than inclusive (Cohen, Reference Cohen1999; Griffin, Reference Griffin2000). Older Black movement organizations, such as the NAACP, have often privileged more “respectable” African Americans as spokespersons for the civil rights movement, rather than Black people whose morality could be called into question.Footnote 6 Thus, Black politicians may find their interests more aligned with the NAACP, an organization that has traditionally worked within the system, as opposed to BLM, an organization that has in effect called for the dismantling of said system.
These types of divisions between Black organizations are nothing new to mainstream Black politics. It is also worth noting, for example, that there are parallels between the contemporary divisions between the NAACP and BLM, and divisions that existed between the NAACP and the Student Nonviolent Coordinating Committee (SNCC) during the height of the civil rights movement. While SNCC and the NAACP, had the same goal of advancing the Black agenda for political advancement, the two organizations maintained a rocky relationship (Bond Reference Bond2000). Morgan and Davies (Reference Morgan and Davies2012) highlight that amidst the growing efficacy of the sit-in movement “there was a prevailing suspicion in the new movement that the voice of age, particularly as represented by the black establishment, was the counsel of caution” (Morgan and Davies, Reference Morgan and Davies2012, 9). Ella Baker, the former NAACP activist, also cautioned student activists at the Shaw University conference against association with the Black establishment, as it would limit the scope of the organization's work. In other words, Baker recognized that the “non-institutionalized,” disruptive tactics SNCC sought to employ would threaten the status quo that groups such as the NAACP were incentivized to maintain. This is similar to the dilemma that Black politics faces today.
3. Respectability politics
Differences in the NAACP's and BLM's approach to “respectability politics,” might also influence the responsiveness of Black legislators. Black legislators might prefer to engage constituents who are associated with organizations that are deemed as more “respectable,” such as the NAACP. There is a long history within the Black community of a “politics of respectability,” whereby Black elites chastise and police the behavior of other African Americans, many of whom Black elites perceived as needing guidance and correction. Higginbotham (Reference Higginbotham1993), who coined the term “respectability politics,” shows in her account of the Woman's Convention (W.C.) of the National Baptist Convention how middle- and upper-class Black women sought to inculcate in poor Black migrants from the South, “temperance, industriousness, thrift, refined manners and Victorian sexual morals” (14–15). According to Higginbotham, the W.C. emphasized the reform of individual behavior as a goal in itself, but also as a strategy for reform of the entire structural system of American race relations. Black elites reasoned that by Black people ascribing to dominant, mainstream norms, they could prove themselves worthy of full inclusion in American society.
The politics of respectability also has a history of being embraced by some of the foremost Black politicians and thinkers (Gaines, Reference Gaines1996). As Harris (Reference Harris2012, 102) reminds us, prominent Black leaders such as W.E.B. DuBois and Booker T. Washington, “were united in their view that the habits of ordinary Black folk needed self-correction and supervision in order for the race to progress,” although they differed on the means of achieving Black progress. Respectability politics also featured prominently in the civil rights movement, including ideas about which people were deemed as reputable spokespersons for the movement, as well as what constituted appropriate dress for sit-in participants (jackets for men and dresses for ladies) (Schmidt, Reference Schmidt2018).
Respectability politics continues to be an important and hotly debated aspect of contemporary Black politics. Recent debates about the tactics of BLM—a movement that explicitly eschews respectability politics—have exposed fractures that are entrenched in traditional, mainstream Black politics. While these divisions are nothing new to members of the Black community, they have recently received more attention in mainstream media. For example, in an op-ed that appeared in the Washington Post in 2015, ordained minister and author, Barbara Reynolds was very critical of the BLM movement, noting that most people who were activists in the 1960s, “admire the cause of these young activists but fundamentally disagree with their [Black Lives Matter's] approach.” She drew a sharp comparison between BLM and the civil rights movement by saying, “Trained in the tradition of Martin Luther King Jr., we were nonviolent activists who won hearts by conveying respectability and changed laws by delivering a message of love and unity.” In other words, Reynolds identifies, “conveying respectability,” as an integral part of producing social change.
In a similar vein, Harvard Law School Professor, Randall Kennedy, contends that respectability politics is necessary to “improve our [Blacks’] chances of surviving and thriving” in the United States. Finally, President Obama also appears to embrace respectability politics in his critique of BLM. While speaking at a town hall event in London on April 23, 2016, he said, “[The Black Lives Matter movement is] really effective in bringing attention to problems…Once you've highlighted an issue and brought it to people's attention and shined a spotlight, and elected officials or people who are in a position to start bringing about change are ready to sit down with you, then you can't just keep on yelling at them” (Shear and Stack, Reference Shear and Stack2016). If in fact, the view of political elites is that BLM movement leaders are intent on “yelling” at elected officials, then it stands to reason that elected officials might be less receptive to BLM, relative to other movement organizations without a reputation for embracing disruptive tactics.
4. Hypotheses
We hypothesize that the email from the NAACP-inspired alias will receive a higher response rate than both the control and BLM treatment email (H1). This hypothesis is informed by the fact that BLM has explicitly eschewed respectability politics. In contrast, the NAACP claims the distinction of being “the oldest and boldest,” which is a nod to them being the oldest civil rights organization in the United States and also the boldest. Despite claiming to be the “boldest” however, it is again worth noting that relative to other civil rights organizations that existed during the civil rights movement, such as the Student Nonviolent Coordinating Committee (SNCC) and the Congress on Racial Equality (CORE), the NAACP was less radical in its tactics (Haines, Reference Haines1984; Morris, Reference Morris1984; McAdam, Reference McAdam1999).
We also expect a lower response rate to the control email relative to the NAACP email. The control email does not mention inspiration from any organization. Thus, the control email may draw less of a response from legislators because it does not signal any connection to the Black political establishment. Our hypotheses more formally stated are as follows:
H1: The response rate will be higher for NAACP-inspired emails relative to the control condition.
H2: The response rate will be lower for BLM-inspired emails relative to the NAACP condition.
5. Methods
To test our hypotheses, in 2017, we fielded an experiment on a sample of approximately 400 Black state legislators, which is considerably large, given that many states in the United States do not have any or very few Black legislators in their legislature. Every legislator in our sample was sworn in as of January 2017. We collected the emails of Black state legislators through state government websites, which provide both the emails and pictures of legislators serving in the state legislature. State legislatures who did not provide the emails of state legislators on their websites were not included in the study because we had no means of emailing them. As a result, 10 states were excluded from the sample. An additional nine states were also excluded from the sample because they did not have any Black state legislators.Footnote 7 In total, 31 states were included in the study. For this experiment, Black state legislators included both Black state representatives and Black state senators. Many states included in this experiment also have Black legislative caucuses. Therefore, in the cases where the states had Black legislative caucuses, we checked the membership rolls of these caucuses, which helped to ensure that we correctly gathered the emails of all Black legislators in each state legislature. We also collected relevant demographic information, including the legislators' gender, party, and incumbency status. Details about the sample are outlined in Table 1.
Table 1. Descriptive statistics of the sample
All emails were sent from the alias, “Deshawn Robinson,” purposely selected to signal Blackness. Fryer and Levitt (Reference Fryer and Levitt2004) show that there are virtually no White people with the name “Deshawn.” The 2010 Census data also lists Robinson as a top surname among Black Americans (Comenetz, Reference Comenetz2016). Therefore, this experiment was intentionally designed to present Black state legislators with an opportunity to advance the interests of a member of his or her racial group. “Deshawn” described himself in the email as a college student who wanted help getting more involved in politics. Black state legislators were randomly assigned to one of three virtually identical emails save for whether the email mentioned the NAACP, BLM, or no organization at all. Random assignment was determined through the use of a random number generator. We did not block on any of the legislators' characteristics, such as party or gender. Random assignment ensured that each state legislator in our sample had an equal chance of receiving one of the three fictitious emails. Therefore, any differences in the responses to the emails can be attributed to the email, and no systematic differences in the treatment groups at the onset of the experiment.
In one condition, the fictitious college student expressed interest in getting more involved in politics because the political activities of BLM inspired him. In contrast, in another condition, he was inspired by the political activities of the NAACP. There was also a control condition in which the fictitious college student expressed interest in getting involved in politics without any mention of inspiration from an organization. This control condition provides an appropriate baseline for testing whether Black legislators were generally unresponsive to any email from the putatively Black college student, or if they were particularly responsive or unresponsive to an email from a student inspired by either BLM or the NAACP.
In conducting the experiment, we had several ethical considerations. First, our use of deception, the use of a fictitious alias, and email when contacting legislators was not a decision that we took lightly. We wanted to test, however, whether Black legislators were marginalizing other African Americans, who some might deem less “respectable.” It is also plausible that some African American constituents may be penalized because of their association with an organization that engages in more confrontational tactics. Therefore, we felt that the benefits of what we would learn by conducting the study would outweigh any of the potential drawbacks associated with the use of deception. Also, of note is that while we would have liked to have a more direct test of responsiveness to the two organizations, for ethical reasons, we could not design the experiment such that our fictitious Black constituent was a member of the organization. Nevertheless, our analysis still enables us to detect whether Black legislators exhibited bias toward an individual who referenced inspiration from BLM versus the NAACP.
We also took steps to minimize any harm that our experiment might cause. Therefore, we have maintained the anonymity of the legislators to ensure that our experimental results are not used to tarnish the reputation of any one legislator. It would also be irresponsible and unethical to report the results in a manner that revealed the behavior of an individual legislative office. Furthermore, from a scientific perspective, we learn far more by reporting the average response to a given treatment or email. We also cannot observe all potential outcomes for any given legislator, and thus we do not know how they would have responded to the other treatments. We can only make average comparisons across groups of legislators. Finally, we were also conscious of the burden that an email request could place on legislators' time. Thus, the email subject was deliberately designed such that legislators could answer the inquiry with one or more opportunities for civic engagement in one email response. Examples of the treatments follow:
6. Results
Data collection occurred for 2 weeks in March 2017Footnote 8. Approximately 21% of the 382 emails that we sent to Black state legislators received a response. While we expected that emails associated with BLM would be less likely to receive a response, the response rate to emails in the BLM condition was the highest at 25%. The response rate for the NAACP condition was the lowest at 19%, while the response rate was approximately 21.5% in the control condition. All results are shown in Table 2.
Table 2. Response rate and sample size by experimental condition
Subsequently, we used a Tukey multiple comparison test to determine whether the difference in the average response rate across the three conditions was statistically significant. These results are displayed in Table 3. As indicated in Table 3, these differences did not attain traditional levels of statistical significance, but this lack of a difference may be a result of insufficient power. Power analysis indicates that at a response rate of 21.5% in the control condition, a treatment effect of 10% should be expected to require 305 participants per group (power = .8, α = .05). Given that our differences between groups were weaker, and our sample smaller (n = 130 per group), the lack of a significant difference may be in part due to insufficient statistical power.
Table 3. Tukey's multiple comparison of means test
Note: Entries are the results from Tukey's multiple comparison of means test.
It is also worth noting that the average length of reply from legislators was 61 words, which are about two to three sentences. Thus, the responses were concise. The median length of reply from legislators was about 40 words, which is somewhat less than the median of 50 words that Broockman (Reference Broockman2013) reports but still arguably within the realm of a standard response. For example, one legislator enthusiastically responded, “Hello Deshawn, We will need to meet as Session begins to slow down. I have included NAACP [President] in this email to seek guidance on getting more involved. Let us stay in touch” (33 words). In contrast, an email that responded to the BLM-inspired alias was much shorter, stating, “Deshawn, I admire your ambitions of wanting to lead but questions your motives” (13 words). It is unclear what the legislator meant by questioning the alias' motives because he did not elaborate. However, we include these emails not as evidence of a systematic pattern of responses to the NAACP emails relative to the BLM emails, but rather for qualitative richness.
We estimate that the legislator's staff, including legislative aids, chiefs of staff, and interns, responded to the alias' email 36% of the time in this experiment (based on email signature and the content of the response). It is worth noting, however, that we have no way of verifying who actually wrote and sent the email response.Footnote 9 Finally, the sample of legislators was 52% male and 48% female. Also, 62% of legislators were from the South and 77% were state representatives, rather than senators. In general, legislators' responses were polite and helpful.
Aside from conducting Tukey's multiple comparison test, we also estimated a logistic regression model, where the dependent variable was whether the legislator responded to the email. We report the results in Table 4. The results of the model indicate that controlling for legislator traits such as gender, incumbency status, residing in the South, and the professionalism of the legislator does not alter the null finding.Footnote 10 Specifically, Table 4 indicates that the likelihood of the alias in the BLM condition receiving a response to her email is 26% higher relative to the alias in the NAACP condition, but this difference is not statistically significant. Also, the likelihood of the alias in the control condition receiving a response to her email was 6% higher relative to the alias in the NAACP condition, but again, this difference was not statistically significant. Thus, we do not find evidence of Black state legislators engaging in intragroup discrimination based on the organization that was referenced in the alias' email. Also, of note, is that male legislators were more likely to respond to the email relative to female legislators (p < .05). Given that the alias Deshawn is a Black male, this supports existing evidence on the influence of descriptive representation, such as race or gender, on substantive representation (Minta Reference Minta2011). Conversely, the legislators' incumbency status was associated with a decline in responsiveness. This finding is consistent with previous research that suggests that elected officials who face less competition are less responsive to their constituents (Griffin, Reference Griffin2006). Finally, the percent Black population in the state was also associated with a statistically significant decline in responsiveness. However, it is worth noting that this decline was not substantively meaningful, given that the odds ratio associated with this variable was very small (.01).
Table 4. Likelihood of responding to email by experimental condition (baseline = NAACP condition)
Note: Entries are odds ratios from a logistic regression model were responding to the email is the dependent variable. Standard errors in parentheses. The model controls for legislator gender (male = 1), residing in the south (south = 1), and incumbency in years. The model also controls for the percentage of the state's population that is Black and the Squire index of legislative professionalism.
*** p < .05, ** p < .01, * p < .001.
Overall, the results of this study are interesting but should be interpreted with caution. This study was underpowered, and thus it is unclear whether the lack of difference in the response rate is because legislators did not engage in intragroup discrimination, or if we simply did not have sufficient power to pick up on any differences in behavior. One result, however, that is not contingent upon having sufficient statistical power is the overall response rate. The results indicate that the legislators in the sample were generally unresponsive, with an overall response rate of 21%. A response rate of 21% is normatively troubling because it suggests that a young, Black college student, who is interested in getting involved in politics is unlikely to have her email acknowledged. Furthermore, the response rate of 21% is notably lower than response rates found in similar audit studies (Butler and Broockman, Reference Butler and Broockman2011; White et al., Reference White, Nathan and Faller2015), which had response rates of 57 and 71%, respectively.
We think that several factors might be at play. One, the email request from the alias asking to “get more involved in politics” was purposely vague. Given that there was not a specific constituency request, such as requesting information about voter identification laws or where to vote, legislators may have been reluctant to respond to such a vague request, regardless of the constituent's source of inspiration. The incentive to help a student, “get more involved in politics” is weaker than the clear, electoral incentive associated with a request regarding voting. Helping a college student become more politically engaged may be less likely to lead to a direct benefit for the legislator driving them to be generally unresponsive to those requests. Two, it is also plausible that despite our use of deception, as audit studies of this type have grown in popularity, legislators may have detected this and thus opted not to respond. Finally, since our study excluded non-Black legislators, we have no context to assess whether the response rate of Black legislators was low relative to legislators of other races.
Also, of note is that our decision to present the alias as a college student and not as a member of the respective organizations allowed us to tap into legislators' perceptions regarding the organizations more accurately. Identifying the alias as a college student removes the external social pressure that may exist for the legislator to respond to a member of one of the two political organizations. Legislators may fear damaging relationships with a Black political organization by not responding to a member, which may have potentially negative electoral consequences. By portraying the alias as an ordinary college student, this experiment more authentically operationalizes the feasibility of political incorporation for external groups, such as BLM. Nevertheless, given that younger voters have poor turnout rates (Plutzer, Reference Plutzer2002), the alias' identification as a college student may have influenced legislators not to respond. However, waves of protests on college campuses close to the time of data collection in March 2017 may have indicated to the legislator that youth, especially college students, seek to be more involved in both electoral and non-electoral forms of political participation.
Finally, though this research utilized deception, future research should continue to employ deception without debriefing, with overwhelming caution, and only in situations where it is necessary. For the experiment, deception was necessary to understand the actual perceptions of Black legislators toward two prominent Black movement organizations, the NAACP and BLM. The deception was coupled with an experimental design such that external social pressure and the potential electoral benefits were weakened. We argue that this is an important context in which to study legislator behavior and the potential for intragroup discrimination. Future research should explore intragroup differences within the Black community with sufficient power. In addition, scholars could test for intragroup discrimination with respect to a more specific request, such as applying for an internship or requesting information about an opportunity to participate in politics, such as attending a legislative session or hearing.
In conclusion, our study is the first audit study, to our knowledge, to test for secondary marginalization and intragroup discrimination among African Americans. While previous studies have documented how higher status African Americans can marginalize other African Americans (Cohen, Reference Cohen1999; White, Reference White2007; Jefferson, Reference Jefferson2019; Lopez Bunyasi and Smith, Reference Lopez Bunyasi and Smith2019), such intragroup dynamics are typically very difficult to quantify with respect to the behavior of elected officials. Thus, we see our study as a modest attempt in that direction. We find that Black state legislators were just as responsive when they were responding to an email sent from a BLM-inspired constituent, as when the email came from an NAACP-inspired constituent, of course with the caveat that the study was underpowered. Our conclusion is based on one study, at one point in time. Therefore, we are not making a blanket claim about whether Black legislators engage in secondary marginalization or intragroup discrimination, more broadly, or in different contexts. Yet despite our null result and the limitations of our study, this study is still interesting, timely, and relevant to scholarship on Black politics. We move beyond tests for intergroup discrimination with an experiment that explores intragroup dynamics within the Black community.
Acknowledgments
The authors thank the anonymous reviewers, the participants at the 2018 University of Michigan Emerging Scholars Conference, and the participants at the 2019 Meeting of the National Conference of Black Political Scientists for their valuable feedback.
