Members of Congress cultivate legislative reputations as a means of signaling to their constituents what their representational priorities are and the type of work that they have been doing during their time in Congress. As shown in the last chapter, there are certain members of Congress that choose to build these reputations around serving disadvantaged groups in particular. While these reputations can vary in terms of which group on whose behalf a member chooses to act or exactly how central a particular disadvantaged group is to that reputation, there are a select group of members that make that choice consistently, across time and party lines. So, why do members do this? What factors contribute to a member’s choice to devote at least some notable portion of their legislative reputation to serving the disadvantaged?
In this chapter, I explore what drives members of Congress to form reputations as advocates for the disadvantaged. In particular, I investigate the influence that disadvantaged group size within a district, the district ambient temperature toward a group (the general group affect), and personal experience as a group member have on a member’s choice to cultivate a reputation as a disadvantaged-group advocate. To start, I will describe in greater detail the construction of these primary explanatory variables of interest. Next, I analyze their effects using a series of generalized ordered logistic regression models. In this analysis, I determine the impact of group size, ambient temperature, and other relevant variables on member reputation for different levels of disadvantaged-group advocacy, both individually and in a comprehensive model. Finally, I use these models to evaluate the role of the advocacy window, as described in Chapter 2.
4.1 Group Size
As discussed in Chapter 2, one of the primary hypotheses is that a large group presence in a district will make it more likely that a member will form a reputation around serving that group. This hypothesis is derived directly from a basic understanding of how a democratic republic should work; members of Congress are elected by people back in their districts, and thus it is their responsibility to work on their behalf. And as demonstrated in the previous chapter, this work frequently takes the form of developing a reputation around serving the groups in their district.
4.1.1 Measuring Group Size
Group size in a state or district is measured for each decade using state and district census totals. This is determined at the unit of district/state decade to consistently account for population shifts and changes in the number and size of districts over time. For nearly all groups, this is a straightforward measure of the percentage of group members in a state or district at the time of the census. Veterans are the percentage of civilian veterans over the age of 16, seniors are the percentage of residents aged sixty-five or older, immigrants are the percentage of foreign-born individuals, women are the percentage of the total population that identify as female, the poor are the percentage of individual residents whose income is at or below the poverty line, and racial/ethnic minorities are the percentage of residents in a district that do not identify as non-Hispanic whites.
Determining the size of the LGBTQ population in a state or district is considerably more complicated, and a perfect measure simply does not exist. The Census has never included any questions about sexual orientation or gender identity at the individual level, but it does provide data about the percentage of same-sex unmarried-partner households within a state or district. The way that the Census Bureau has determined this, however, has undergone important changes over the course of the last several decades. In 1990, the Census added the category of “unmarried-partner household” for the first time and also recorded whether that household contained an opposite-sex couple or a same-sex couple. Because same-sex marriage was not recognized at the national level until the 2013 US Supreme Court decision in United States v. Windsor, no formal Census report included same-sex married couples prior to the 2020 Census, and households that classified themselves as same-sex married couples had their relationship type changed to unmarried partner.Footnote 1
Using the percentage of unmarried-partner households in a district is certainly an imperfect proxy for the percentage of LGBTQ individuals, but the lack of a superior alternative has made it the preferred measure used by social scientists (e.g., Reference WawroWarshaw and Rodden, 2012; Reference HawkesworthHansen and Treul, 2015). For this same reason, I also utilize this measure as an approximation of the percentage of LGBTQ Americans in the state or district, but acknowledge that it has two important shortcomings.Footnote 2 First, by definition, this measure leaves out a count of LGBTQ individuals who are not in a same-sex coupled household and offers no insight into the percentage of transgender individuals or bisexual individuals partnered with a person of the opposite sex in a district. This undoubtedly results in an undercount relative to the actual percentage of LGBTQ individuals in a district. Second, the percentage of same-sex unmarried-partner households is likely to be highly correlated with the feelings of warmth or hostility toward a group. While LGBTQ individuals exist everywhere in the country, the likelihood of being part of an “out” same-sex couple that also feels comfortable enough to identify as such on a government form is assuredly higher in a less-hostile environment. I expand further upon the consequences of this expected strong correlation in Section 4.3 discussing the relationship between group size and group ambient temperature.
4.2 Ambient Temperature
Though most Americans do not conceptualize their political beliefs and decisions on strongly ideological grounds, a large percentage of individuals do use group identities and intergroup dynamics as a way of understanding politics (Reference Campbell, Converse, Miller and StokesCampbell et al., 1960; Reference Griffin and NewmanGreen et al., 2002; Reference MacDorman and MathewsLewis-Beck et al., 2008). Given this centrality of group-based political attitudes, intergroup dynamics within a district can play an important role in understanding the reputations members of Congress choose to form. In particular, feelings toward different disadvantaged groups can shape the likelihood that a member will work on their behalf.
A crucial component of a representative’s job is to act in accordance with their constituents’ wishes. Or, at the very least, for a member to weigh constituent preferences against other considerations like partisan pressures, individual beliefs, and national interest. Most conceptions of this representational relationship thus require that constituents be able to communicate their policy preferences to their member of Congress. For especially salient issues, particularly those coming up for a major vote, this communication can take place in a way that follows in line with these theoretical expectations. Under these circumstances, constituents are more likely to call or write in to congressional offices and show up at town halls to express their opinions, and members may even have access to internal or external polling indicating how the public feels about an issue. But when a member is considering a new policy proposal, or preparing for a vote on an issue that has not received a great deal of public attention, it is far less likely that members have a clear, specific sense of their constituents’ opinions to guide their actions.
Members of Congress have markedly less formal information while making legislative decisions than is commonly assumed (Reference CurryCurry, 2015). Members have ever-increasing demands on their time, but limited resources. When deciding whether or not to engage in a particular legislative action, they frequently must depend upon either their own intuitions or general cues from others about how important constituencies within a state or a district are likely to be impacted. In this low-information decision-making environment, a member’s trust in their own perceptions about their districts can take on paramount importance. Members of Congress pride themselves on how well they know their districts (Reference FennoFenno, 1978), and their perceptions of the subconstituencies within their districts impact their work within the institution (Reference MilerMiler, 2010). Even if members may not have all of the information they might wish they had, they likely do have a sense of how popular certain groups of people are back home in their district. Members do not need to have specific polling data on how constituents feel about a pilot program promoting minority-owned small businesses or bill to eliminate food deserts in poor communities to have a sense of whether or not there is political risk or political benefit to supporting such a piece of legislation.
This limited information environment and reliance on the perceptions of a groups’ relative popularity within a district influences not only member decisions on individual legislative actions but also their reputation formation more broadly. Making decisions based on the feelings of warmth or hostility toward a particular disadvantaged group can serve as a helpful shortcut that does not require the allocation of additional resources. Representatives must make choices about what issues they will devote their time and energy to working on, and each of these specialization decisions pulls resources away from other prospective issues. Members of Congress are risk-averse. If members feel that there is a potential for negative backlash from a considerable portion of their constituents if they were to work on behalf of a particular disadvantaged group, they will simply direct their reputation-building efforts in another direction. After members have taken into account the size of groups within their states or districts, considering the general feelings toward that group within their constituency is a reasonable heuristic when deciding whether or not to build a reputation as an advocate for that group. In the next section, I describe the operationalization of these feelings of warmth or hostility toward a particular group – what I refer to as the group ambient temperature.
4.2.1 Measuring Group Affect
For this analysis, I utilize feeling thermometer scores as a measure of how the average district resident feels about members of a particular disadvantaged group. A feeling thermometer is a commonly used survey tool that asks respondents to rate, on a scale from 0 to 100, how warmly they feel toward a particular societal group. Highly regarded and long-running political survey projects like the American National Election Study (ANES) have relied on feeling thermometer questions for decades, with some variation in the salient group identities included over time.
Feeling thermometers are a simple, readily understandable way for individuals to articulate how warmly they feel toward a particular group without needing much in the way of political knowledge or sophistication. For this reason, feeling thermometer ratings are preferable to other, more complex ways of attempting to determine group-specific representational preferences. Compared to more intricate undertakings like expressing a definitive opinion on a particular policy or political action, rating a group on a feeling thermometer is a much more manageable task for respondents across levels of political information. This makes it less subject to top-of-the-head influences and inconsistencies than other measures. As an example, a question asking individuals to identify which groups they most want their representative to work on behalf of or asking how they would want them to allocate their time is extremely complicated, and invites greater opportunity for respondents to misunderstand the question and its intent.
One potential downside to using feeling thermometers is the influence of social desirability bias. The very existence of social desirability bias as a possible issue, however, is also evidence that respondents have a keen understanding of the task set before them when presented with a feeling thermometer. Social desirability bias exists when respondents answer a question in a way that they feel is most likely to be acceptable and gain the approval of others (such as the person administering the survey) rather than in accordance with their genuine sentiment. As an example, in the United States, being racist is generally considered by society at large to be a bad thing, at least over the time period studied here. Thus, if an individual gives “Black Americans” a higher feeling thermometer rating than might actually be true out of a fear of being perceived negatively, that respondent clearly understands that the fundamental task of a feeling thermometer is to rate their own personal level of affinity toward members of different groups.
The range of disadvantaged groups I evaluate in this analysis runs the gamut in terms of expected social desirability bias. For instance, the poor and the LGBTQ community provoke very different levels of public support and affection, with the poor largely being well regardedFootnote 3 and the LGBTQ community being looked on with suspicion or outright animosity (particularly during the 1990s and 2000s). This results in different amounts of social pressure to positively evaluate each group. However, because this analysis focuses on differences across districts for each group rather than within-district differences across groups, it is only the relative positioning of each group in different areas of the country that is important. Additionally, because the presence of this bias would move the ratings in a consistent direction, artificially inflating the feeling thermometer scores, it actually provides a conservative test for the hypothesis that a higher ambient temperature will boost the likelihood of a member crafting a reputation around advocating on behalf of a disadvantaged group.
Using feeling thermometer scores has a number of practical advantages as well. Survey designers have been using feeling thermometers for decades, allowing for the analysis of the impact of district hostility on member behavior over a broader range of time than studies utilizing the more recent Cooperative Congressional Election Study (CCES) or the National Annenberg Election Study (NAES) data, which tend to focus on more policy-specific questions. The format and measurement of feeling thermometer questions have also been remarkably consistent over time, alleviating some of the difficulties of working with specific issue questions, where a topic might be asked about while it is salient and actively being pursued by Congress, but then is dropped or replaced as other issues gain more prominence. I generate estimates of state- and district-level feeling thermometer scores using multilevel regression with poststratification (MRP).
4.2.2 Estimating State and District Ambient Temperature
MRP modeling utilizes national public opinion data and regional (state- and district-level) demographic data to estimate the opinion of relevant population sub-groups, which are then weighted and summed for the geographic area of interest. In the remainder of this section, I provide a brief overview of the data and modeling techniques used to generate the ambient temperature estimates. Additional details on the benefits of MRP and the specific modeling formulations used in this project can be found in Appendix C.
To generate estimates of state- and district-level ambient temperatures, I use feeling thermometer data from the American National Election Study (ANES) times series data from 1992 to 2016. Feeling thermometer estimates are generated for each disadvantaged group in each of the three decades included in the scope of this project (the 1990s, the 2000s, and the 2010s) to account for changes over time. In each of these MRP models, the explanatory variables are relevant demographic data pulled from the decennial US Census and the US Religion Census, while the dependent variable is the group feeling thermometer score. The ANES includes feeling thermometer questions for a number of societal groups, and close approximates for others. The models estimating the average district and state ambient temperature toward racial and ethnic minorities, the poor, and seniors are a direct match with the feeling thermometer question in the ANES, while the models estimating feelings toward the LGBTQ community, women, immigrants, and veterans utilize proxies.Footnote 4 Respectively, these proxies are lesbians and gays, feminists/women’s libbers, illegal immigrants/illegal aliens, and the military.Footnote 5 Summaries of the estimated district feeling thermometer ratings by groups are given in Table 4.1.
Table 4.1 Summary of estimates for district feeling thermometer ratings by group
| Group | N | Mean | Std. Dev. | Min. | Max. |
|---|---|---|---|---|---|
| Seniors | 1,740 | 78.74 | 2.33 | 64.89 | 98.89 |
| Veterans | 2,175 | 72.66 | 3.88 | 54.14 | 85.80 |
| Poor | 2,175 | 69.35 | 2.96 | 56.91 | 87.05 |
| Women | 2,175 | 56.32 | 3.56 | 41.28 | 72.54 |
| Immigrants | 2,175 | 42.43 | 8.71 | 22.48 | 70.47 |
| Racial/Ethnic Minorities | 2,175 | 66.09 | 3.23 | 53.42 | 79.76 |
| LGBTQ | 2,175 | 47.83 | 7.82 | 25.47 | 77.83 |
Displayed are the average estimated values for the feeling thermometer scores across all congressional districts from 1992 to 2016. Estimates were generated using multilevel regression with poststratification. The estimate for racial/ethnic minorities is an average of the ambient temperature generated for each district for Black, Hispanic, and Asian Americans.
4.2.3 Interpreting Ambient Temperature Estimates
Before moving on to the application of these measures of district and state ambient temperature, it is necessary to say a word about how these estimates should be interpreted. First, as a general note, though the mean ambient temperatures for each group do follow approximately the same pattern as might be expected by how deserving of government assistance each group is broadly perceived to be, it is important to keep in mind that these two measures are not the same. It is possible, for instance, that people could have positive feelings toward a group, but also feel that it is not the government’s role to provide assistance. Conversely, people could have more negative personal feelings about the military, for example, while still feeling that the government owes particular benefits to those who served. Given this, it is expected that the ambient temperature toward a group can vary from state-to-state or district-to-district, independent of how generally deserving of government assistance the group is considered to be.
Second, and most crucial to appropriate interpretation of these estimates, the raw values for the means and standard deviations are not intended to provide a head-to-head comparison across groups. Particularly because of the use of proxy feeling thermometer scores for several of these groups, the estimated means may better represent the true latent ambient temperatures for some groups more so than others. With the exception of the use of “lesbians and gays” as a stand-in for the LGBTQ community (which could result in a slightly more positive rating by not activating biphobia or transphobia), it is likely that the proxy group measurements (described in the previous section) may result in a slightly lower average feeling thermometer score. However, because the research design of this project calls for the evaluation of the representation of each group individually, these deviations should not negatively impact the results. What matters is the relative difference within the same group’s ambient temperature across districts, not the absolute value comparison across groups. Similarly, given this research design, the different degree of variation in the ambient temperature measures is not a concern, as the effects of ambient temperature on reputation for group advocacy is evaluated independently for each group.
4.3 Group Size and Ambient Temperature: Related but Distinct Concepts
Before discussing and presenting the multivariate models exploring the factors that drive members of Congress to foster reputations as advocates of disadvantaged groups, this section takes a quick look at the relationship between the two primary independent variables – group size and ambient temperature – and establishes them as related but separate concepts, each with distinct reasons for inclusion in the models to follow.
Group size and ambient temperature are expected to be related concepts. Having frequent interactions with a group, or having a high percentage of group members within a district can boost the positive feelings toward that group. However, this relationship need not only work in a singular direction; for example, some research has shown that when white Americans are confronted with information about increasing racial/ethnic minority and/or immigrant populations, negative feelings toward those groups can actually increase (Reference BlumerBlumer, 1958; Reference Alba, Rumbaut and MarotzAlba, Rumbaut, and Marotz, 2005; Reference Craig and RichesonCraig and Richeson, 2014). Table 4.2 displays the correlations between district group size and district ambient temperature for each of the groups whose representation is being analyzed here.
Table 4.2 Correlations between district group size and group ambient temperature by disadvantaged group
| Seniors | Veterans | Poor | Women | Immigrants | Racial/Ethnic Minorities | LGBTQ |
|---|---|---|---|---|---|---|
| −0.0245 | 0.0640* | 0.4477* | 0.1635* | 0.5334* | 0.4287* | 0.6319* |
There is broad diversity across disadvantaged groups when considering the strength of the relationship between the size of group within a district and the general feelings toward that group. For the two groups that are broadly considered to be deserving of government assistance, seniors and veterans, there is little to no correlation between the two variables, with the correlation for seniors failing even to reach conventional levels of statistical significance. These results match with expectations – because of the high esteem these groups are held in, any variations in ambient temperature should not be related to the percentage of veterans and seniors in a district. Women also have a low correlation between ambient temperature and group size, best explained by the very small amount of variation in the percentage of women across districts. For the poor, immigrants, racial/ethnic minorities, and the LGBTQ community, there is evidence of a much stronger correlation between group size and ambient temperature. Given the higher levels of skepticism these groups face relative to groups like veterans or seniors, this relationship is expected and, as discussed above, likely is not purely linear. The correlation between the two variables for the LGBTQ community also reflects the nature of the creation of the group size variable itself, and how Census reporting for same-sex couples are likely themselves somewhat reflective of group ambient temperature. In no case, however, even for the LGBTQ community, is there evidence of even close to a perfect correlation between the two variables, as seen in Figure 4.1.
Figure 4.1 Average district ambient temperature and percentage of group members by district and disadvantaged group
Note: Figure displays maps of the estimated average group ambient temperature and group size by district for the 108th Congress. Groups included in the figure are racial/ethnic minorities, LGBTQ individuals, immigrants, veterans, seniors, the poor, and women.
Figure 4.1 shows the variation in the estimated ambient temperature and group size across all congressional districts in the continental United States for the 108th Congress (Jan 2003–2005). These maps showcase both the instances of overlap in the variation in ambient temperature and group size, and the discrepancies. When considering immigrants, for example, similar bright areas in Southern California and Southern Texas display the close relationship between the percentage of and general regard for immigrants in the area. Another clear example of similarities comes in the bright dot around San Francisco indicating both one of the higher percentages of LGBTQ individuals in the countries as well as one of the highest ambient temperatures. For groups like women, seniors, and veterans, however, a high level of variation in the shading between each of the two maps can be seen.
Though not present for all groups, the relatively high correlations between ambient temperature and group size for racial/ethnic minorities, immigrants, the poor, and LGBTQ individuals do have some ramifications for the multivariate models to follow. This multicollinearity will likely have the effect of slightly inflating the standard errors on these variables, making conventional statistical significance harder to achieve. Despite this, the strong theoretical reasons for including each of these measures, laid out in Chapter 2, necessitate that the models contain both of them. Group size is a straightforward accounting of the number of group members who comprise a member’s constituency, while ambient temperature measures how the district constituency tends to feel about a group writ large. To truly understand what drives members to make the reputational choices that they do, the impact of each of these variables must be accounted for. Including both of these variables is also necessary to evaluate the role of the ambient window, as is done toward the end of this chapter.
4.4 Modeling Reputation Formation
As discussed in detail in Chapter 3, reputation is coded as a four-category ordinal variable, ranking from Non-Advocates to Superficial Advocates to Secondary Advocates to Primary Advocates. In the remaining sections of the chapter, I analyze how group size, group ambient temperature, and other relevant explanatory variables impact the type of reputation members of Congress form. Specifically, in this chapter, I focus on the reputation formation of members of the House of Representatives, while the reputation formation of senators is considered in the following chapter. In each of the following sections, these relationships are modeled using generalized ordered logistic regression, or the partial proportional odds model, to account for both the ordinal character of the reputation variable and to allow for the explanatory variables to have differential impacts on primary, secondary, or superficial reputation formation.Footnote 6
4.4.1 Alternative Explanatory Variables
In addition to my primary variables of interest, group size, and group ambient temperature, I include a number of control variables: party affiliation, whether or not a member is in their first term in Congress, the partisan leaning of the district, and whether or not the member represents a district in the South. I also include decade fixed effects and cluster the standard errors by member. Controlling for the decade fixed effects accounts for specific time-bounded changes in what drives members to form reputations around serving disadvantaged groups. By clustering my standard errors by individual member, I avoid any artificially deflated standard errors that could arise from the same person remaining in the House across more than one Congress in the sample.
First and foremost, I control for the party of the representative, coded as a dichotomous variable.Footnote 7 Given the centrality of partisan concerns in the US Congress and the differences between the electoral coalitions of the Democratic and Republican Parties, understanding party-specific differences is crucial. As seen in the previous chapter, while both Democrats and Republicans do form these reputations for disadvantaged-group advocacy, the phenomenon is more common among members of the Democratic caucus. Thus, even with other factors accounted for, I expect that Democrats in Congress are going to be more likely to form a reputation around advocating for disadvantaged groups than Republicans, particularly for those groups for whom government intervention is generally viewed with more skepticism.
I include a dichotomous variable in the model indicating if a member is in their first term in the House for two reasons. First, members who have just won their first term in the House simply have not had enough time to develop a strong legislative reputation as an advocate for disadvantaged groups. Reputation building requires a pattern of behaviors and interactions, in which new members have not had the opportunity to engage. Second, practically speaking, Politics in America devotes less space to first-term members than to members that have at least one full congressional session under their belt, meaning there is also less available information that is written about them.Footnote 8 For these reasons, the coefficient on this variable is expected to indicate a strongly negative effect on whether a member has a reputation for disadvantaged-group advocacy.
I control for the partisan leaning of the district by including a measure of the two-party vote share that the Democratic presidential candidate received in the most recent presidential election. I expect that the more Democratic a district is, the more likely it is that a member will have a reputation for disadvantaged-group advocacy. Finally, I also control for regional effects by including a dummy variable for whether or not a district is located in the South. I define South as including the original former Confederate states, as well as states bordering the former Confederate states. These states include: Alabama, Arkansas, Delaware, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Texas, Virginia, and West Virginia. I expect that the political culture will generally make members representing districts within these states less likely to form a reputation around advocating for disadvantaged groups, and that these effects will be most acute for disadvantaged groups that are not highly regarded in the country at large.
4.5 Group Size, Ambient Temperature, and Member Reputation
In this section, I investigate the independent relationships between group size and then ambient temperature on member reputation, after taking into account all of the alternative variables listed above. Understanding each of these variables can provide valuable insight into how group size and group affect within a state or district influences the representational choices members of Congress make. Following from the basic expectations of republican governance, wherein an individual is elected by a set of constituents to act in accordance with their interests, it is expected that group size will have a positive relationship with member reputation. Members of the House of Representatives with a reputation for serving as a group advocate are more likely to come from a district with a large number of group members than a district with very little group presence. Given how inherent this is to the fundamental notion of representation, this should be true regardless of how deserving of government assistance a group is considered to be. Ambient temperature is also expected to have a generally positive effect on the likelihood of members forming reputations as group advocates. However, these effects should be the most apparent for groups that are considered to be less deserving of government assistance.
Tables 4.3–4.5 present the results of the generalized ordered logistic regression models with reputation for disadvantaged-group advocacy as the dependent variable of interest. Each group has three models (0, 1, and 2), which contain important information about how the independent variables can asymmetrically impact the likelihood of a member having a certain kind of reputation.Footnote 9 The dichotomous variable indicating whether or not a member was just elected to their first term in Congress is modeled as a having parallel proportional effects,Footnote 10 while all other explanatory variables are modeled with partial proportional effects. Model 0 demonstrates how group size and the other independent variables impact the likelihood of a member shifting from having no reputation for group advocacy at all to having some kind of reputation for advocacy. The likelihood of moving from no advocacy or superficial advocacy to secondary or primary advocacy is shown in Model 1, and the shift from one of the lower levels of advocacy to primary group advocacy is presented in Model 2. Breaking the models up in this way allows for differential impacts of the explanatory variables to account for the variation in representational levels. I will next evaluate all of the groups in turn to determine the impact of disadvantaged group size and ambient temperature on the reputation a member of Congress formulates.
4.5.1 Veterans and Seniors
The impact of the percentage of veterans and seniors in a district on member reputation is presented in Table 4.3. Across nearly all levels, group size has a positive and significant effect on the type of reputation a member forms. As expected, even for groups that are considered to be broadly deserving of government assistance across the United States, members representing districts with higher quantities of group members are more likely to form reputations around advocating on their behalf. For veterans, group size has a significant impact on all shifts across levels of reputation. Seniors exhibit the same pattern when moving from a non-advocate to an advocate and when shifting to develop a reputation as a primary or secondary advocate from non- or superficial advocacy. However, a higher percentage of seniors in a district did not significantly push members from the lower levels of advocacy into primary advocacy. Florida’s 5th district, for example, has consistently had one of the highest percentages of seniors in a Congressional district since the 1990s. Rep. Karen Thurman (D-FL5), who served in the 1990s, and Rep. Ginny Brown-Waite (R-FL5), who won the seat in the 108th Congress (during the 2002 elections), each developed reputations for superficial advocacy of seniors during their time in office.
Table 4.3 Group size, ambient temperature, and member reputation for advocacy for veterans and seniors
| Veterans | Seniors | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | 2 | 0 | 1 | 2 | 0 | 1 | 2 | 0 | 1 | 2 | |
| Group Size | 0.210 | 0.262 | 0.399 | – | – | – | 0.093 | 0.121 | 0.082 | – | – | – |
| 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.54 | |||||||
| Ambient Temperature | – | – | – | 0.09 | 0.17 | 0.18 | – | – | – | −0.02 | 0.06 | −0.10 |
| 0.00 | 0.00 | 0.05 | 0.66 | 0.44 | 0.57 | |||||||
| Republican | −0.466 | −0.814 | −1.148 | −0.157 | −0.425 | −0.262 | −0.830 | −1.123 | −1.733 | −0.721 | −0.929 | −1.565 |
| 0.02 | 0.02 | 0.13 | 0.41 | 0.20 | 0.63 | 0.00 | 0.00 | 0.03 | 0.00 | 0.00 | 0.05 | |
| Dem Pres Vote | 0.000 | 0.018 | 0.061 | 0.002 | 0.030 | 0.072 | −0.002 | −0.019 | −0.075 | 0.015 | −0.003 | −0.088 |
| 0.99 | 0.42 | 0.23 | 0.85 | 0.16 | 0.11 | 0.86 | 0.34 | 0.22 | 0.20 | 0.89 | 0.15 | |
| South | 0.185 | 0.590 | 0.512 | 0.028 | 0.363 | 0.544 | −0.118 | −0.215 | −1.550 | 0.025 | −0.199 | −1.586 |
| 0.38 | 0.19 | 0.38 | 0.91 | 0.46 | 0.42 | 0.53 | 0.49 | 0.13 | 0.91 | 0.63 | 0.10 | |
| 1990s | −1.319 | −1.126 | −1.649 | −0.346 | 0.322 | −0.048 | −0.217 | −0.334 | −1.227 | −0.578 | −0.770 | −1.605 |
| 0.00 | 0.01 | 0.03 | 0.25 | 0.57 | 0.96 | 0.11 | 0.24 | 0.05 | 0.00 | 0.02 | 0.02 | |
| 2000s | −0.938 | −0.671 | −1.594 | 0.217 | 0.941 | 0.618 | – | – | – | – | – | – |
| 0.00 | 0.09 | 0.02 | 0.33 | 0.01 | 0.28 | |||||||
| First Term | −1.056 | −0.920 | −1.020 | −0.834 | ||||||||
| 0.00 | 0.00 | 0.00 | 0.00 | |||||||||
| Constant | −3.669 | −6.611 | −11.207 | −8.595 | −17.650 | −21.992 | −2.495 | −3.143 | −0.973 | −0.719 | −7.104 | 8.840 |
| 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.01 | 0.00 | 0.01 | 0.68 | 0.80 | 0.26 | 0.56 | |
| N | 2,175 | 2,175 | 2,175 | 2,175 | ||||||||
| Wald’s Chi2 | 95.6 | 90.1 | 68.9 | 70.3 | ||||||||
| Pseudo-R2 | 0.0651 | 0.0388 | 0.0436 | 0.0414 | ||||||||
Different patterns are evident when examining the role of group ambient temperature. Warmer feelings toward veterans within a district have a positive and statistically significant impact on a member having a reputation around advocating for veterans, regardless of the level of advocacy. For seniors, however, group ambient temperature has no significant effect on member reputation. In the models including ambient temperature, the role of party affiliation is also different for seniors and for veterans. Once the ambient temperature for veterans is taken into account, members of Congress from both parties are equally likely to form reputations as advocates. This is different when considering members with reputations around advocating for seniors, where Democrats still have a statistically significant advantage. Given that both of these groups are considered to have high levels of deservingness of government assistance, this discrepancy is intriguing. This difference is likely attributable to the close ties between seniors’ issues and Social Security and Medicare, which in turn are also more closely bound to the Democratic Party.
When it comes to group size, for both veterans and seniors, Republicans are significantly less likely to form reputations as group advocates than are Democratic members. So, while a larger share of Republican members with reputations for disadvantaged-group advocacy are advocates for veterans and seniors relative to Democrats, Democratic members are still more likely to have formed these reputations on the whole. After other factors are accounted for, the partisan lean of the district and presence in the South do not have a significant impact on whether or not a member forms their reputation around advocating on behalf of veterans or seniors. Members of Congress were significantly less likely to have reputations for veterans’ advocacy in the 1990s and 2000s compared to the 2010s, which may be due to a recent acknowledgement of the challenges facing veterans returning from the wars in Iraq and Afghanistan.
4.5.2 LGBTQ and Racial/Ethnic Minorities
The LGBTQ community and racial/ethnic minorities are disadvantaged groups that are generally considered to be less deserving of government assistance, and it is expected that both the size of the group in the district and the group ambient temperature will make it more likely that a member will have a reputation around serving as a group advocate. As seen in Table 4.4, both group size and ambient temperature do have the expected positive and significant effect on member reputation. Unlike with seniors, where group size is not significantly related to the decision to serve as a primary group advocate and group affect has no significant determinative effect on reputation formation, decisions about advocacy for less highly regarded groups such as the LGBTQ community and racial/ethnic minorities are significantly related to these variables. When ambient temperature and group size are modeled separately, higher values of each increase the likelihood that a member will have a reputation for group advocacy. This is true across nearly all levels of advocacy.Footnote 11
Republican members of Congress are significantly less likely to form reputations around advocating for racial/ethnic minorities and the LGBTQ community. This is in accordance with the breakdowns by party shown in Figures 3.4 and 3.5 – considerably fewer Republicans hold reputations as advocates for racial/ethnic minorities and LGBTQ individuals than Democrats. Though less common, Republican representatives are not wholly immune to the effects of group size or ambient temperature. Ray LaHood (R-IL18), for instance, began to be known as a superficial advocate for racial/ethnic minorities (particularly those of Middle Eastern descent) in the 2000s, after the ambient temperature of his district increased by nearly a full standard deviation in the wake of redistricting.
Table 4.4 Group size, ambient temperature, and member reputation for advocacy for racial/ethnic minorities and the LGBTQ community
| LGBTQ | Race/Ethnicity | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | 0 | 1 | 0 | 1 | 2 | 0 | 1 | 2 | |
| Group Size | 1.824 | 3.805 | – | – | 0.051 | 0.064 | 0.054 | – | – | – |
| 0.01 | 0.03 | 0.00 | 0.00 | 0.00 | ||||||
| Ambient Temperature | – | – | 0.07 | 0.11 | – | – | – | 0.22 | 0.24 | 0.22 |
| 0.01 | 0.08 | 0.00 | 0.00 | 0.00 | ||||||
| Republican | −1.553 | −2.319 | −1.848 | −2.270 | −1.929 | −2.116 | −2.758 | −2.522 | −3.172 | −3.779 |
| 0.02 | 0.03 | 0.00 | 0.01 | 0.00 | 0.00 | 0.01 | 0.00 | 0.00 | 0.00 | |
| Dem Pres Vote | 0.057 | 0.092 | 0.045 | 0.059 | −0.027 | −0.043 | −0.041 | −0.013 | −0.016 | −0.021 |
| 0.12 | 0.04 | 0.20 | 0.27 | 0.11 | 0.05 | 0.07 | 0.38 | 0.43 | 0.40 | |
| South | – | – | – | – | −0.007 | 0.079 | 0.176 | 0.261 | 0.417 | 0.408 |
| 0.98 | 0.82 | 0.66 | 0.29 | 0.19 | 0.27 | |||||
| 1990s | 1.014 | 4.552 | 0.429 | 1.421 | 1.148 | 0.943 | 0.098 | 1.641 | 1.402 | 0.684 |
| 0.06 | 0.14 | 0.45 | 0.21 | 0.00 | 0.00 | 0.81 | 0.00 | 0.00 | 0.12 | |
| 2000s | −0.226 | 2.809 | 0.123 | 1.189 | 1.301 | 1.090 | 0.151 | 1.696 | 1.719 | 0.940 |
| 0.60 | 0.23 | 0.82 | 0.11 | 0.00 | 0.00 | 0.67 | 0.00 | 0.00 | 0.01 | |
| First Term | −1.348 | −1.411 | −1.882 | −1.705 | ||||||
| 0.08 | 0.06 | 0.00 | 0.00 | |||||||
| Constant | −7.956 | −14.778 | −9.861 | −14.065 | −2.867 | −3.683 | −3.752 | −16.683 | −18.497 | −17.114 |
| 0.00 | 0.00 | 0.00 | 0.00 | 0.01 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | |
| N | 2,175 | 2,175 | 2,175 | 2,175 | ||||||
| Wald’s Chi2 | 34.7 | 38.1 | 367.4 | 226.1 | ||||||
| Pseudo-R2 | 0.1889 | 0.1335 | 0.3048 | 0.1791 | ||||||
Region, however, has slightly different effects on the advocacy of racial/ethnic minorities and the LGBTQ community. When other explanatory variables are taken into account, whether or not a member represents a district in the South does not have a significant impact on forming a reputation as an advocate for racial/ethnic minorities. However, because there are zero members of Congress representing southern districts with a reputation for LGBTQ advocacy, the precise effects of region cannot be calculated, on account of perfect prediction.
There also exists an interesting difference between representation of the LGBTQ community and racial/ethnic minorities when considering the impact of the partisan leaning of the district. When group ambient temperature is taken into account, the partisan leaning of a district does not significantly impact reputation formation as an advocate of racial/ethnic minorities or LGBTQ individuals. But when group size is considered in place of ambient temperature, partisan leaning works in opposite directions for the LGBTQ community and racial/ethnic minorities. While partisan leaning does not have a significant effect on whether or not a member shifts from being a non-advocate to having a reputation as some level of group advocate, it does have significant impact on the move from non- or superficial advocacy to primary or secondary advocacy. After controlling for the size of the LGBTQ community in a district, members coming from more Democratic districts are more likely to have a reputation as a primary or secondary advocate. However, members from heavily Democratic districts are less likely to have a reputation as an advocate for racial/ethnic minorities once racial composition of a district is accounted for.
4.5.3 Immigrants and the Poor
Immigrants and the poor are groups that hold a complicated place in the American mind. Many people are highly sympathetic to the poor while simultaneously treating with great skepticism those who access our welfare systems, and immigrants hold dual roles in the national zeitgeist of both heroic ancestor and multicultural villain. For each of these two disadvantaged groups, district group size and ambient temperature have a significant, positive relationship with the likelihood of members having a reputation for group advocacy, as seen in Table 4.5. The single apparent exception to this comes when considering members with a primary reputation for serving the poor – group ambient temperature does not have a statistically significant effect on this final move up the ladder, implying that other conditions are the driving force for this last step. Major Owens (D-NY11) is an example of someone who made the decision to form a reputation as a secondary advocate for the poor during the 1990s, pushing for increases in the minimum wage and boosts for other social programs. His district fell just outside of the top 10th percentile for the percentage of people in poverty, but had the second highest ambient temperature toward the poor of any other district in the sample.
Across all levels of advocacy, Republican members of Congress are significantly less likely to have a reputation as an advocate for those in poverty. In contrast, this effect is less consistent for members with reputations as immigrant advocates. After immigrant ambient temperature is held constant, Republicans are not significantly less likely to be primary or secondary advocates for immigrants. Particularly considering the time period under consideration, this difference fits with the general perceptions of the party identities – despite some more recent changes, immigration had long been considered to be a fairly bipartisan issue. And once again, members in their first term are significantly less likely to have a reputation for disadvantaged-group advocacy.
Table 4.5 Group size, ambient temperature, and member reputation for advocacy for immigrants and the poor
| Immigrants | Poor | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | 2 | 0 | 1 | 2 | 0 | 1 | 2 | 0 | 1 | 2 | |
| Group Size | 0.111 | 0.149 | 0.214 | – | – | – | 0.067 | 0.094 | 0.107 | – | – | – |
| 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | |||||||
| Ambient Temperature | – | – | – | 0.09 | 0.13 | 0.16 | – | – | – | 0.06 | 0.13 | 0.07 |
| 0.00 | 0.00 | 0.06 | 0.05 | 0.00 | 0.31 | |||||||
| Republican | −0.635 | −0.570 | −2.518 | −1.044 | −0.551 | −1.380 | −1.136 | −1.818 | −2.037 | −1.385 | −2.217 | −2.624 |
| 0.04 | 0.22 | 0.00 | 0.00 | 0.36 | 0.51 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | |
| Dem Pres Vote | −0.079 | −0.080 | −0.213 | −0.014 | 0.016 | −0.017 | 0.021 | 0.016 | 0.002 | 0.012 | 0.002 | −0.001 |
| 0.00 | 0.04 | 0.02 | 0.56 | 0.74 | 0.86 | 0.04 | 0.22 | 0.96 | 0.25 | 0.91 | 0.97 | |
| South | −0.325 | −0.578 | −2.365 | −0.084 | 0.044 | −0.689 | −0.362 | −0.833 | −0.833 | −0.311 | −0.835 | −0.498 |
| 0.40 | 0.32 | 0.01 | 0.84 | 0.96 | 0.72 | 0.03 | 0.00 | 0.22 | 0.08 | 0.00 | 0.46 | |
| 1990s | −0.055 | 0.004 | 0.241 | 0.367 | 0.248 | 0.623 | 0.582 | 0.582 | −0.697 | 0.558 | 0.615 | −0.806 |
| 0.85 | 0.99 | 0.62 | 0.21 | 0.54 | 0.64 | 0.00 | 0.01 | 0.08 | 0.00 | 0.01 | 0.07 | |
| 2000s | −0.001 | 0.200 | −0.988 | −0.123 | −0.116 | −0.583 | 0.735 | 0.659 | 0.176 | 0.708 | 0.810 | 0.057 |
| 1.00 | 0.54 | 0.10 | 0.56 | 0.64 | 0.13 | 0.00 | 0.00 | 0.58 | 0.00 | 0.00 | 0.90 | |
| First Term | −2.000 | −1.621 | −1.222 | −1.212 | ||||||||
| 0.00 | 0.00 | 0.00 | 0.00 | |||||||||
| Constant | −0.020 | −2.174 | 3.013 | −5.697 | −10.510 | −10.221 | −3.039 | −4.204 | −4.506 | −5.446 | −11.104 | −7.545 |
| 0.99 | 0.32 | 0.50 | 0.00 | 0.00 | 0.30 | 0.00 | 0.00 | 0.04 | 0.01 | 0.00 | 0.16 | |
| N | 2,175 | 2,175 | 2,175 | 2,175 | ||||||||
| Wald’s Chi2 | 307.2 | 122.6 | 257.0 | 212.6 | ||||||||
| Pseudo-R2 | 0.2851 | 0.1166 | 0.1229 | 0.1011 | ||||||||
The negative effects on likelihood of reputation formation of coming from a southern district are particularly strong when considering advocates for the poor, and more mixed for those with reputations as advocates for immigrants. Partisan lean of a district also has a different effect for each of these two groups. When ambient temperature is accounted for in the model, partisan leaning does not have a statistically significant effect on reputation formation for either group. For the poor, more Democratic leaning districts are more likely to push members to form superficial reputations for working on behalf of the group, but do not have a significant effect on reputations for higher levels of advocacy. The pattern for immigrants, however, is more similar to that of racial/ethnic minorities – after controlling for the percentage of immigrants in a district, more Democratic-leaning districts are actually less likely to produce members with reputations as immigrant advocates.
4.5.4 Women
Women are particularly unique among the disadvantaged groups studied in this project, both because of their limited variation in group size across districts, and because their size is near or above a majority of individuals in all districts. Based on size alone, one could argue that building a reputation around advocating on behalf of women would be a winning strategy for all members of Congress, and yet very few members choose to foster such a reputation. As seen in Table 4.6, the percentage of women in the district has no statistically significant impact on the likelihood of a member having a reputation as a women’s advocate. Group ambient temperature does have significant effect on reputation formation, but the effect is nuanced. Higher ambient temperatures toward women have a significant impact on the shift from having a reputation as a non-advocate toward being an advocate at any level, even if superficial, but it is not a significant driving force behind the move to secondary or primary advocacy.
Table 4.6 Group size, ambient temperature, and member reputation for advocacy for women
| Women | ||||||
|---|---|---|---|---|---|---|
| 0 | 1 | 2 | 0 | 1 | 2 | |
| Group Size | 0.016 | −0.098 | 0.050 | – | – | – |
| 0.78 | 0.29 | 0.81 | ||||
| Ambient Temperature | – | – | – | 0.05 | 0.03 | 0.01 |
| 0.05 | 0.22 | 0.92 | ||||
| Republican | −0.826 | −1.342 | −2.135 | −0.766 | −1.277 | −2.253 |
| 0.00 | 0.00 | 0.01 | 0.00 | 0.00 | 0.01 | |
| Dem Pres Vote | 0.054 | 0.069 | 0.086 | 0.047 | 0.063 | 0.102 |
| 0.00 | 0.02 | 0.03 | 0.01 | 0.02 | 0.01 | |
| South | −0.554 | −1.054 | −1.612 | −0.550 | −1.039 | −1.613 |
| 0.05 | 0.03 | 0.18 | 0.05 | 0.04 | 0.17 | |
| 1990s | 0.704 | 0.272 | 0.062 | 0.702 | 0.306 | 0.374 |
| 0.00 | 0.33 | 0.93 | 0.00 | 0.30 | 0.50 | |
| 2000s | 0.631 | −0.001 | −0.041 | 0.535 | −0.040 | −0.042 |
| 0.00 | 1.00 | 0.93 | 0.01 | 0.88 | 0.93 | |
| First Term | −1.238 | −1.246 | ||||
| 0.00 | 0.000 | |||||
| Constant | −5.993 | −1.320 | −10.904 | −7.478 | −7.818 | −9.811 |
| 0.06 | 0.81 | 0.29 | 0.00 | 0.00 | 0.03 | |
| N | 2,175 | 2,175 | ||||
| Wald’s Chi2 | 131.6 | 121.5 | ||||
| Pseudo-R2 | 0.0901 | 0.0902 | ||||
Chapter 3 showed that among instances of disadvantaged-group advocacy for Democrats and Republicans, each party had equivalent percentages of their members with reputations as women’s advocates, but Democrats had a greater number of advocates overall. These descriptive statistics are born out here in the multivariate model. Across all of the models analyzed and all levels of advocacy, Republican members of Congress are significantly less likely to have reputations for advocacy on behalf of women. Other variables also had similar effects across models. First term members are less likely to have a reputation for advocacy, as are members from the South. These regional effects are statistically significant for the shift to superficial advocacy and secondary advocacy, but do not have a significant effect on the decision to form a primary reputation around advocating for women. Members also become increasingly likely to have a reputation as a women’s advocate the more Democratic the district that they represent.
4.6 The Relationship Between Group Size and Ambient Temperature
Generally speaking, with a few exceptions, the results above demonstrate that both group size and ambient temperature have an important and positive impact on whether or not a member has a reputation as a group advocate when each is considered separately. However, as discussed earlier, while group size and ambient temperature are related, they are nonetheless distinct concepts that are expected to work in different ways depending upon the circumstance. Thus, I next explore the overall effects of each of these variables when allowed to work together in a single model.
4.6.1 Veterans, Seniors, Racial/Ethnic Minorities, LGBTQ Individuals
Table 4.7 demonstrates that when combined within a single model, the general relationship between group size, ambient temperature, and reputation follow the same general pattern, regardless of whether a group is broadly considered to be highly deserving of government assistance or less deserving of government assistance. Across levels of reputation for advocacy, group size nearly always has a positive and statistically significant impact upon reputation formation. Once again, members with primary reputations for advocating on behalf of seniors present the only exception to this, as group size is not a significant determinant of which members choose to focus the bulk of their representational efforts on seniors. That this should represent the exception is not in defiance of expectations, because members should not fear negative reprisals from their constituents for investing such a large share of their efforts into serving a group considered to be so deserving of assistance, even if seniors do not themselves make up a sizeable portion of the district.
An example of the significant impact of group size on the likelihood that a member will form a reputation as a group advocate even for groups that are generally considered to be less deserving of government assistance can be seen in Rep. Jerrold Nadler (D). Throughout the 1990s, Nadler represented the 8th District of New York, covering western Manhattan and parts of Brooklyn. During that time, he consistently served as a secondary advocate for LGBTQ Americans. After the decennial redistricting, most of that territory was shifted into the 10th District, and the percentage of LGBTQ individuals increased by almost a full standard deviation. As the representative of the 10th District of New York in the 113th Congress, Nadler gained a reputation as a primary advocate, becoming a leader of the efforts to repeal Don’t Ask, Don’t Tell.
When group size is taken into account, ambient temperature is no longer seen to exert a significant effect upon the reputations members choose to form. This is not unexpected. While higher levels of group ambient temperature boost the likelihood of a member forming a reputation as a group advocate when considered independently, it should not overtake the effects of the size of a group. The other explanatory variables also largely perform as anticipated, with first termers and Republican members of Congress being less likely to have reputations as advocates for any of the four groups. Other district demographic factors such as whether or not a district is in the South and the partisan lean of a district generally have no significant impact on member reputation as advocates for these groups.
Table 4.7 Combined effects of group size and ambient temperature on member reputation for advocacy for veterans, seniors, LGBTQ individuals, and racial/ethnic minorities
| Veterans | Seniors | LGBTQ | Race/Ethnicity | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | 2 | 0 | 1 | 2 | 0 | 1 | 0 | 1 | 2 | |
| Group Size | 0.201 | 0.235 | 0.394 | 0.095 | 0.117 | 0.085 | 1.673 | 3.721 | 0.051 | 0.068 | 0.057 |
| 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.51 | 0.02 | 0.05 | 0.00 | 0.00 | 0.00 | |
| Ambient Temperature | 0.050 | 0.108 | 0.047 | −0.012 | 0.078 | −0.087 | 0.046 | 0.074 | 0.009 | −0.049 | −0.033 |
| 0.091 | 0.071 | 0.589 | 0.744 | 0.358 | 0.622 | 0.126 | 0.243 | 0.778 | 0.229 | 0.649 | |
| Republican | −0.466 | −0.819 | −1.173 | −0.775 | −1.006 | −1.643 | −1.532 | −2.316 | −1.927 | −2.116 | −2.749 |
| 0.02 | 0.02 | 0.11 | 0.00 | 0.00 | 0.05 | 0.02 | 0.03 | 0.00 | 0.00 | 0.01 | |
| Dem Pres Vote | 0.003 | 0.026 | 0.062 | 0.012 | −0.008 | −0.088 | 0.039 | 0.068 | −0.029 | −0.036 | −0.037 |
| 0.80 | 0.22 | 0.20 | 0.33 | 0.70 | 0.18 | 0.29 | 0.17 | 0.08 | 0.09 | 0.12 | |
| South | 0.039 | 0.322 | 0.363 | −0.017 | −0.317 | −1.632 | – | – | −0.019 | 0.128 | 0.213 |
| 0.86 | 0.50 | 0.55 | 0.94 | 0.39 | 0.12 | 0.94 | 0.71 | 0.59 | |||
| 1990s | −1.015 | −0.488 | −1.324 | −0.611 | −0.802 | −1.635 | 1.450 | 5.492 | 1.182 | 0.893 | 0.081 |
| 0.00 | 0.44 | 0.15 | 0.00 | 0.01 | 0.02 | 0.01 | 0.14 | 0.00 | 0.01 | 0.85 | |
| 2000s | −0.721 | −0.210 | −1.377 | – | – | – | 0.147 | 3.483 | 1.323 | 0.957 | 0.066 |
| 0.01 | 0.66 | 0.06 | 0.77 | 0.21 | 0.00 | 0.00 | 0.88 | ||||
| First Term | −1.051 | −0.897 | −1.321 | −1.892 | |||||||
| 0.00 | 0.00 | 0.08 | 0.00 | ||||||||
| Constant | −7.526 | −14.989 | −14.827 | −2.003 | −9.427 | 6.807 | −9.512 | −17.803 | −3.408 | −0.860 | −1.839 |
| 0.00 | 0.00 | 0.02 | 0.49 | 0.16 | 0.66 | 0.00 | 0.01 | 0.17 | 0.76 | 0.68 | |
| N | 2,175 | 1,740 | 2,175 | 2,175 | |||||||
| Wald’s Chi2 | 104.6 | 93.2 | 38.2 | 368.2 | |||||||
| Pseudo-R2 | 0.0678 | 0.0543 | 0.1954 | 0.3058 | |||||||
4.6.2 Women, Immigrants, and the Poor
Table 4.8 displays the results of the combined model evaluating reputations for serving women, immigrants, and the poor. For the groups that hold a more mixed position in the American psyche, such as immigrants and people in poverty, group size again has a positive and statistically significant impact on reputation for group advocacy. This can be seen through the actions of former Rep. Xavier Becerra (D-CA30/31/34). Originally known through much of the 1990s and 2000s as a primary advocate for immigrants, Becerra also began building other forms of advocacy into his reputation as redistricting moved him from a district of nearly 60 percent immigrants to one in which immigrants made up less than a majority.
Once group size is accounted for, ambient temperature does not have a statistically significant effect on member reputation for serving immigrants and the poor. Having a reputation as a women’s advocate, however, is not dependent upon group size. Ambient temperature does have positive and statistically significant impact on whether or not a member chooses to form a reputation as some kind of women’s advocate, though the choice to be a primary or secondary advocate is reliant upon other factors.
Table 4.8 Combined effects of group size and ambient temperature on member reputation for advocacy for women, immigrants, and the poor
| Immigrants | Poor | Women | |||||||
|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | 2 | 0 | 1 | 2 | 0 | 1 | 2 | |
| Group Size | 0.121 | 0.153 | 0.226 | 0.068 | 0.090 | 0.110 | 0.001 | −0.110 | 0.058 |
| 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.99 | 0.26 | 0.79 | |
| Ambient Temperature | −0.033 | −0.005 | −0.058 | −0.003 | 0.032 | −0.051 | 0.048 | 0.043 | 0.008 |
| 0.118 | 0.866 | 0.517 | 0.918 | 0.482 | 0.481 | 0.046 | 0.118 | 0.919 | |
| Republican | −0.664 | −0.407 | −2.809 | −1.136 | −1.815 | −2.067 | −0.765 | −1.286 | −2.173 |
| 0.03 | 0.45 | 0.01 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.01 | |
| Dem Pres Vote | −0.078 | −0.078 | −0.222 | 0.021 | 0.014 | 0.005 | 0.047 | 0.062 | 0.088 |
| 0.00 | 0.03 | 0.02 | 0.04 | 0.33 | 0.90 | 0.01 | 0.03 | 0.03 | |
| South | −0.356 | −0.595 | −2.608 | −0.358 | −0.903 | −0.597 | −0.550 | −1.050 | −1.617 |
| 0.35 | 0.28 | 0.02 | 0.05 | 0.00 | 0.39 | 0.06 | 0.04 | 0.17 | |
| 1990s | −0.292 | −0.160 | −0.450 | 0.575 | 0.633 | −0.749 | 0.713 | 0.257 | 0.091 |
| 0.40 | 0.72 | 0.67 | 0.00 | 0.01 | 0.08 | 0.00 | 0.39 | 0.90 | |
| 2000s | 0.051 | 0.157 | −1.110 | 0.728 | 0.773 | −0.041 | 0.541 | −0.094 | −0.132 |
| 0.84 | 0.65 | 0.08 | 0.00 | 0.01 | 0.93 | 0.01 | 0.74 | 0.77 | |
| First Term | −1.992 | −1.221 | −1.243 | ||||||
| 0.00 | 0.00 | 0.00 | |||||||
| Constant | 1.314 | −2.091 | 6.429 | −2.844 | −6.296 | −1.072 | −7.529 | −2.738 | −11.806 |
| 0.41 | 0.28 | 0.44 | 0.18 | 0.04 | 0.84 | 0.02 | 0.61 | 0.26 | |
| N | 2,175 | 2,175 | 2,175 | ||||||
| Wald’s Chi2 | 359.4 | 266.3 | 156.0 | ||||||
| Pseudo-R2 | 0.2895 | 0.1239 | 0.093 | ||||||
As with the representation of the groups discussed above, first term and Republican members are less likely to form any kind of reputation as advocates for women, immigrants, and the poor. Members coming from districts with a more Democratic lean are more likely to have reputations for advocating on behalf of women, but actually less likely to advocate for immigrants. Partisan leaning also has a statistically significant effect on the choice to form a reputation for at least some level of advocacy for the poor, but is not a significant determinant of primary or secondary advocacy. Members representing Southern districts are also less likely to form reputations as advocates for these groups, though it does not retain statistical significance across all advocacy levels.
4.7 Descriptive Representatives and Member Reputation
Thus far, this chapter has demonstrated that group size and group ambient temperature can play an important role in a member’s choices about the reputations they seek to foster, and which groups they do or do not wish to be known as an advocate for. But as discussed in Chapter 2, personal experiences – particularly, being a member of the disadvantaged group yourself – also contribute to these decisions, both by increasing the salience of issues facing that group and by providing an opening to be seen as a credible group advocate. In the next section, I first explore the impact of being a descriptive representative on the reputations for group advocacy that members form. Next, I take a closer look at how the effect of ambient temperature on reputation formation changes for members of Congress that are and are not descriptive representatives.
For these analyses, members are coded as being descriptive representatives of a group if they themselves can claim group membership. For women, racial and ethnic minorities, veterans, and LGBTQ members, this determination is straightforward. Members are considered to be immigrants if they themselves were foreign born and immigrated to the United States, while members with immigrant parents who personally were born in the United States are not included. Those who were first elected to the House at or above the age of sixty-five are considered to be descriptive representatives of seniors. This specific criterion is used in place of someone who has reached the age of sixty-five while serving to both acknowledge that reputations are something that are built over time and that having lived experience as a senior adult prior to entering into Congress may impact the representational choices a person may make once in the institution. This provides continuity with all other descriptive representatives in that experiences prior to serving in Congress are expected to make a difference.
Obviously, there are no members of Congress that are currently destitute – members of Congress tend to be much wealthier than the average American. However, I code a member as a descriptive representative if their CQ Politics in America profile describes them as having grown up in poverty (not working class, but poor), or if there was a time in their life that they relied on public assistance for those in poverty, such as welfare or food stamps. Given the popularity of “up by the bootstraps” narratives among those of both parties in Congress, members with such histories tend to place such experiences front and center in their life stories.
Table 4.9 shows the number of members who are themselves a member of a disadvantaged group who also serve as group advocates. To account for the relatively small number of descriptive representatives in the House of Representatives and ensure a sufficient number of cases across levels of advocacy, for these analyses I combine the categories of secondary advocates and primary advocates. Thus, in each of the models discussed below, there are only two sets of coefficients, rather than three. Model 0 is evaluating the relationship between the explanatory variables and the shift from reputations for non-advocacy to reputations for superficial, secondary, or primary advocacy, and Model 1 compares the impact of the same variables when comparing members with reputations for non-advocacy or superficial advocacy with primary and secondary advocates. Members with reputations as advocates for seniors are excluded from this portion of the analysis, as there are no members that were elected to the House when they were over the age of sixty-five that are secondary or primary advocates for seniors. I expand further on the significance of this lack of descriptive representatives devoting the bulk of their reputations to advocacy for seniors at the end of this section.
Table 4.9 Number of members serving as advocates of disadvantaged groups across descriptive and nondescriptive representatives
| Seniors | Veterans | Women | Poor | Immigrants | Racial/Ethnic Minorities | LGBTQ | ||
|---|---|---|---|---|---|---|---|---|
| Non-DR | Non | 1,897 | 1,437 | 1,785 | 1,587 | 2,032 | 1,781 | 2,130 |
| Superficial | 183 | 76 | 66 | 284 | 63 | 57 | 19 | |
| Secondary | 65 | 28 | 17 | 128 | 23 | 10 | 11 | |
| Primary | 11 | 11 | 1 | 39 | 24 | 3 | 1 | |
| Total | 2,156 | 1,552 | 1,869 | 2,038 | 2,142 | 1,851 | 2,161 | |
| DR | Non | 18 | 541 | 191 | 75 | 23 | 122 | 7 |
| Superficial | 1 | 54 | 47 | 28 | 2 | 69 | 1 | |
| Secondary | 0 | 20 | 42 | 17 | 6 | 81 | 6 | |
| Primary | 0 | 8 | 26 | 17 | 2 | 52 | 0 | |
| Total | 19 | 623 | 306 | 137 | 33 | 324 | 14 | |
Disadvantaged-group advocates in the 103rd, 105th, 108th, 110th, and 113th Congresses by disadvantaged group and member status as descriptive representatives.
4.7.1 Veterans, Racial/Ethnic Minorities, and the LGBTQ Community
The first sets of results in Table 4.10 display models including whether or not a member is themselves LGBTQ, a veteran, or a racial/ethnic minority alongside all of the previous explanatory variables. Here, while many of the patterns seen earlier for these groups that have higher and lower levels of perceived deservingness remain the same, there are a few additional insights that are important. Members who are themselves LGBTQ or a racial/ethnic minority are considerably more likely to form a reputation for group advocacy across both levels, even while group size remains an important determinant of member reputation. Reps. Albert Wynn (D-MD4) and Joe Crowley (D-NY7), for example, represented districts with nearly the same ambient temperature and percentage of racial/ethnic minorities in moderately liberal states. But while Wynn, a Black man, had consistently served as a superficial or secondary advocate for racial/ethnic minorities, Crowley, who is not a descriptive representative for racial/ethnic minorities, prioritized other groups. For veterans, however, being a descriptive representative is significantly related to having at least some part of your reputation built around advocating for veterans, but those who have taken on the highest levels of group advocacy are driven by the percentage of veterans in their districts.
Table 4.10 Descriptive representatives and member reputation for advocacy for veterans, LGBTQ individuals, and racial/ethnic minorities
| Veterans | LGBTQ | Race/Ethnicity | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | 0 | 1 | 0 | 1 | 0 | 1 | 0 | 1 | 0 | 1 | |
| Group Size | 0.195 | 0.233 | 0.195 | 0.237 | 2.010 | 5.072 | 1.974 | 5.058 | 0.015 | 0.033 | 0.016 | 0.032 |
| 0.20 | 0.00 | 0.00 | 0.00 | 0.01 | 0.07 | 0.02 | 0.07 | 0.01 | 0.00 | 0.01 | 0.00 | |
| Ambient Temperature | 0.047 | 0.105 | 0.020 | 0.062 | 0.059 | 0.189 | 0.060 | 0.189 | −0.018 | −0.070 | −0.039 | 0.091 |
| 0.114 | 0.084 | 0.529 | 0.301 | 0.065 | 0.106 | 0.069 | 0.108 | 0.712 | 0.184 | 0.413 | 0.528 | |
| Descriptive Representative | 0.663 | 0.504 | −5.725 | −12.130 | 4.380 | 6.351 | 9.064 | 5.756 | 3.339 | 3.257 | 0.221 | 14.324 |
| 0.00 | 0.12 | 0.07 | 0.17 | 0.00 | 0.00 | 0.11 | 0.87 | 0.00 | 0.00 | 0.97 | 0.16 | |
| Ambient Temp. & DR | – | – | 0.086 | 0.170 | – | – | −0.097 | 0.012 | – | – | 0.047 | −0.167 |
| 0.04 | 0.15 | 0.46 | 0.99 | 0.57 | 0.28 | |||||||
| Republican | −0.516 | −0.856 | −0.521 | −0.885 | −1.340 | −1.507 | −1.387 | −1.563 | −1.430 | −1.409 | −1.435 | −1.317 |
| 0.01 | 0.02 | 0.01 | 0.02 | 0.04 | 0.34 | 0.04 | 0.33 | 0.00 | 0.00 | 0.00 | 0.00 | |
| Dem Pres Vote | 0.004 | 0.028 | 0.004 | 0.028 | 0.011 | 0.013 | 0.013 | 0.014 | −0.016 | −0.022 | −0.017 | −0.019 |
| 0.73 | 0.19 | 0.74 | 0.18 | 0.72 | 0.85 | 0.68 | 0.84 | 0.32 | 0.30 | 0.29 | 0.37 | |
| South | −0.015 | 0.291 | −0.059 | 0.225 | – | – | – | – | −0.163 | 0.051 | −0.176 | 0.072 |
| 0.95 | 0.56 | 0.80 | 0.66 | 0.53 | 0.87 | 0.49 | 0.82 | |||||
| 1990s | −1.191 | −0.622 | −1.173 | −0.525 | 1.845 | 8.757 | 1.783 | 8.713 | 1.492 | 1.039 | 1.511 | 1.025 |
| 0.00 | 0.31 | 0.00 | 0.39 | 0.01 | 0.14 | 0.01 | 0.14 | 0.00 | 0.00 | 0.00 | 0.00 | |
| 2000s | −0.771 | −0.260 | −0.802 | −0.263 | 0.218 | 4.767 | 0.192 | 4.753 | 1.590 | 0.941 | 1.660 | 0.897 |
| 0.00 | 0.59 | 0.00 | 0.59 | 0.71 | 0.29 | 0.77 | 0.28 | 0.00 | 0.01 | 0.00 | 0.01 | |
| First Term | −0.953 | −0.953 | −1.643 | −1.523 | −2.501 | −2.512 | ||||||
| 0.00 | 0.00 | 0.10 | 0.12 | 0.00 | 0.00 | |||||||
| Constant | −7.413 | −14.894 | −5.372 | −11.746 | −9.304 | −24.508 | −9.377 | −24.508 | −2.181 | −0.344 | −0.833 | −10.974 |
| 0.00 | 0.00 | 0.04 | 0.01 | 0.00 | 0.05 | 0.00 | 0.05 | 0.50 | 0.92 | 0.80 | 0.24 | |
| N | 2,175 | 2,175 | 2,175 | 2,175 | 2,175 | 2,175 | ||||||
| Wald’s Chi2 | 90.6 | 101.8 | 89.2 | 141.6 | 493.3 | 512.9 | ||||||
| Pseudo-R2 | 0.0782 | 0.0816 | 0.3024 | 0.3033 | 0.4245 | 0.4258 | ||||||
Once again, first term and Republican members of Congress are less likely to form these reputations for group advocacy, though there is some small variation in the significance of the effect of party. This is particularly true for members with reputations as advocates for the LGBTQ community – once descriptive representatives are taken into account, these other variables that have been significant in prior models drop in significance. After the principal explanatory variables and decade-specific fixed-effects are taken into account, neither the partisan lean nor representing a district in the South has a significant effect on member reputation.
4.7.2 Women, Immigrants, and the Poor
Women, immigrants, and the poor are groups that do not fit as neatly into the categories of being generally considered to be of high or low deservingness of government assistance. The effects of descriptive representation, ambient temperature, and group size on member reputations as advocates for these groups are given in Table 4.11, and a few differences from the groups discussed above can be seen. Being an immigrant is not a statistically significant predictor of member reputation as an advocate for immigrants, while the percentage of immigrants and partisan leaning of a district do have a significant impact. Additionally, while Republican members are less likely to have reputations as advocates for immigrants at any level, partisanship is not a statistically significant factor separating non- and superficial advocates from secondary and primary advocates. But while members representing Southern districts are not any less likely to have reputations for serving as advocates for immigrants, Southern members are considerably less likely to build reputations as advocates on behalf of the poor, after group size and other variables are held constant. Partisan leaning of a district is also a significant source of separation between non-advocates and members with a reputation for some level of advocacy. After these district-specific factors are taken into account, group size, ambient temperature, and descriptive representation follow now-familiar patterns. Personal experience with poverty and the percentage of poor constituents in a district each have a significant impact on member reputation, while the significance of ambient temperature drops off.
Table 4.11 Descriptive representatives and member reputation for advocacy for women, immigrants, and the poor
| Immigrants | Poor | Women | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 1 | 0 | 1 | 0 | 1 | 0 | 1 | 0 | 1 | 0 | 1 | |
| Group Size | 0.123 | 0.155 | 0.127 | 0.160 | 0.066 | 0.089 | 0.064 | 0.088 | 0.098 | 0.009 | 0.089 | 0.008 |
| 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.16 | 0.94 | 0.20 | 0.94 | |
| Ambient Temperature | −0.041 | −0.020 | −0.046 | −0.029 | −0.005 | 0.027 | −0.017 | 0.025 | 0.016 | 0.016 | 0.045 | 0.077 |
| 0.078 | 0.532 | 0.059 | 0.381 | 0.861 | 0.580 | 0.588 | 0.625 | 0.581 | 0.646 | 0.182 | 0.136 | |
| Descriptive Representative | 0.084 | 0.235 | 5.377 | −7.473 | 0.574 | 0.606 | −10.239 | −1.740 | 2.746 | 3.370 | 6.340 | 9.151 |
| 0.90 | 0.73 | 0.05 | 0.04 | 0.02 | 0.04 | 0.05 | 0.78 | 0.00 | 0.00 | 0.03 | 0.02 | |
| Ambient Temp. & DR | – | – | −0.116 | 0.157 | – | – | 0.155 | 0.034 | – | – | −0.063 | −0.100 |
| 0.04 | 0.03 | 0.03 | 0.70 | 0.23 | 0.14 | |||||||
| Republican | −0.700 | −0.587 | −0.662 | −0.491 | −1.106 | −1.773 | −1.120 | −1.775 | −0.426 | −0.942 | −0.427 | −0.940 |
| 0.02 | 0.30 | 0.03 | 0.40 | 0.00 | 0.00 | 0.00 | 0.00 | 0.07 | 0.01 | 0.07 | 0.01 | |
| Dem Pres Vote | −0.080 | −0.087 | −0.082 | −0.088 | 0.022 | 0.015 | 0.022 | 0.016 | 0.050 | 0.053 | 0.050 | 0.054 |
| 0.00 | 0.03 | 0.00 | 0.03 | 0.04 | 0.28 | 0.04 | 0.27 | 0.01 | 0.05 | 0.01 | 0.05 | |
| South | −0.419 | −0.762 | −0.403 | −0.750 | −0.355 | −0.884 | −0.361 | −0.855 | −0.386 | −0.847 | −0.364 | −0.803 |
| 0.29 | 0.20 | 0.31 | 0.21 | 0.05 | 0.00 | 0.05 | 0.00 | 0.20 | 0.08 | 0.23 | 0.10 | |
| 1990s | −0.416 | −0.341 | −0.469 | −0.380 | 0.570 | 0.625 | 0.561 | 0.613 | 1.353 | 1.001 | 1.360 | 0.992 |
| 0.23 | 0.43 | 0.19 | 0.40 | 0.00 | 0.01 | 0.00 | 0.01 | 0.00 | 0.00 | 0.00 | 0.00 | |
| 2000s | 0.052 | 0.209 | 0.059 | 0.224 | 0.697 | 0.727 | 0.712 | 0.723 | 0.932 | 0.152 | 0.921 | 0.169 |
| 0.84 | 0.55 | 0.81 | 0.52 | 0.00 | 0.01 | 0.00 | 0.01 | 0.00 | 0.65 | 0.00 | 0.60 | |
| First Term | −1.983 | −1.965 | −1.206 | −1.203 | −1.771 | −1.771 | ||||||
| 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | |||||||
| Constant | 1.859 | −0.861 | 2.074 | −0.496 | −2.713 | −5.998 | −1.896 | −5.876 | −12.263 | −8.769 | −13.501 | −12.268 |
| 0.29 | 0.75 | 0.25 | 0.86 | 0.20 | 0.06 | 0.39 | 0.09 | 0.00 | 0.13 | 0.00 | 0.07 | |
| N | 2,175 | 2,175 | 2,175 | 2,175 | 2,175 | 2,175 | ||||||
| Wald’s Chi2 | 275.2 | 293.2 | 239.6 | 251.7 | 258.7 | 275.8 | ||||||
| Pseudo-R2 | 0.2894 | 0.2928 | 0.129 | 0.131 | 0.2521 | 0.2539 | ||||||
The unique situation of women among other disadvantaged groups is again revealed in these models – while group size in a district is not a driving force behind members forming reputations as women’s advocates, being a woman in Congress has an enormous influence. These effects can be clearly seen in the case of Rep. Barbara Lee (D-CA13). Lee has cultivated a reputation as a primary advocate for women, serving as a vocal defender of abortion rights and a strident opponent to any cuts in the funding of women’s health organizations such as Planned Parenthood. Lee was preceded as a representative of the 13th District of California by Rep. Pete Stark (D). Stark built a reputation as an advocate for liberal healthcare solutions, but did not focus his advocacy specifically on women’s health or other issues directly pertaining to women.
Once descriptive representation is taken into account, the ambient temperature toward women in a district no longer plays a statistically significant role in member reputation. Members coming from more Democratic districts are more likely to craft reputations around serving women, and while Republicans are less likely to have reputations as primary or secondary advocates, conventional levels of statistical significance are not quite met when evaluating the role of party on the choice to become a superficial advocate of women. This could indicate that once a member’s gender is taken into consideration, the choice to form a reputation as a superficial advocate of women is a slightly less partisan choice than that of other disadvantaged groups.
4.7.3 Interactive Effects of Descriptive Representation on Reputation
Next, I look more closely at the differences in the effect of ambient temperature on reputation for descriptive and nondescriptive members. To this point, ambient temperature has only occasionally (as in the case of women’s advocates) been seen to have a statistically significant effect on member reputation once group size is included in the model. But while it is not surprising that the size of the group within a district would have a much stronger pull on member reputation than group ambient temperature, I do expect that the marginal effects of changes in ambient temperature will have different effects for members who are themselves a descriptive representative of a disadvantaged group than those who are not.
I test this by including an interaction term between ambient temperature and whether or not someone is a descriptive representative, and then examining the predicted marginal effects on the probability of a member holding a reputation as a disadvantaged-group advocate. The models themselves can be seen in the second set of columns in Tables 4.10 and 4.11. Because these are generalized ordered logistic regression models, predicted probabilities are the best means of displaying the interactive effects of these variables on member reputation for group advocacy. These predicted effects can be seen in Figures 4.2 and 4.3. For each of these figures, all other variables in the models besides ambient temperature and the dichotomous measure for descriptive representatives take on the actual values that they have in the dataset, which allows for the specific marginal effects of ambient temperature to be isolated and estimated for descriptive representatives and nondescriptive representatives.
Figure 4.2 Predicted effects of ambient temperature for members with reputations as advocates relative to non-advocates for descriptive representatives
Note: Figures show the predicted marginal effects of ambient temperature on reputation for superficial, secondary, or primary advocacy relative to non-advocacy for members who are themselves descriptive representatives of the group and members who are not. Predicted marginal effects are calculated using Stata’s margins command for the Model 1’s containing interactions between descriptive representative and ambient temperature shown in Tables 4.10 and 4.11. All other variables are held to their observed values within the dataset. Clockwise from the top left, the groups whose representation is being analyzed are racial/ethnic minorities, LGBTQ individuals, women, the poor, immigrants, and veterans.
Figure 4.3 Predicted effects of ambient temperature for members with reputations as primary or secondary advocates relative to non- or superficial advocates for descriptive representatives
Note: Figures show the predicted marginal effects of ambient temperature on reputation for secondary or primary advocacy relative to non- or superficial advocacy for members who are themselves descriptive representatives of the group and members who are not. Predicted marginal effects are calculated using Stata’s margins command for the Model 1’s containing interactions between descriptive representative and ambient temperature shown in Tables 4.10 and 4.11. All other variables are held to their observed values within the dataset. Clockwise from the top left, the groups whose representation is being analyzed are racial/ethnic minorities, LGBTQ individuals, women, the poor, immigrants, and veterans.
Figure 4.2 shows the differences in the marginal predicted effects of ambient temperature on forming any level of reputation for advocacy of disadvantaged groups for members of the House who are themselves a part of that group, and members who are not a descriptive representative. One of the most interesting trends seen below is not just how the impact of ambient temperature on reputation varies depending upon the group being represented, but the differences in these predicted marginal effects for members who are and are not descriptive representatives. For nondescriptive representatives, the effects of changes in ambient temperature on the likelihood of having a reputation for group advocacy are extremely small, with slopes remaining close to constant. For members who are descriptive representatives, however, changes in ambient temperature have larger effects on reputations for group advocacy, with stronger effects for some groups than others.
For veterans, a group broadly considered to be deserving of government assistance, there is little change in the predicted effects across values in ambient temperature. The same is true for immigrants and LGBTQ members. Women in the House, though, have a clear and significant boost in the likelihood of having a reputation as a women’s advocate relative to men, but also do not see a clear shift in that likelihood as ambient temperature increases. For racial/ethnic minorities and those with a history of poverty, however, there is a clear increase in the likelihood of a member of the House who is a descriptive representative forming a reputation as some kind of group advocate – even if only a superficial advocate, as is true in the majority of cases – as the group ambient temperature of the district goes up.
The marginal effects of ambient temperature for descriptive and nondescriptive representatives when considering the difference between non- and superficial advocates and primary and secondary advocates are shown in Figure 4.3. These figures demonstrate that there are important differences in how descriptive representation and ambient temperature affects the likelihood of having a reputation that may include superficial group advocacy versus that of a reputation specifically for secondary or primary advocacy – those members that choose to cultivate a considerable amount of their legislative reputation around advocating for a disadvantaged group. The groups for which changes in ambient temperature were previously seen to do little to increase the likelihood of some level of advocacy on their behalf, regardless of whether or not a member was a descriptive representative, largely do see some effects of changes in ambient temperature when comparing those members with reputations for primary and secondary advocacy to all other members.
As before, members of Congress who are not themselves descriptive representatives experience only very small boosts in the likelihood of having primary or secondary reputations for group advocacy when ambient temperature increases. For members who are themselves part of the disadvantaged group, however, increases in the predicted effects are seen for nearly all groups. Veterans, immigrants, those with a history of poverty, and LGBTQ members see an increased likelihood of being primary or secondary group advocates at high levels of district group ambient temperature. Once again, members with reputations for advocating for women are the exception to this trend – even when only considering what sets primary and secondary advocates apart, changes in ambient temperature have very little effect once the large and clearly statistically significant boost that comes from being a woman is taken into account.
One of the most interesting trends that can be seen in Figure 4.3 is the effect of changes in ambient temperature on reputation as a primary or secondary advocate for racial and ethnic minorities. Here again, there is a clear and statistically significant difference in the effects of ambient temperature between descriptive representatives and nondescriptive representatives. However, in this instance, as ambient temperature increases, the likelihood that a member who is themselves a racial/ethnic minority will have a primary or secondary reputation for group advocacy actually decreases. This change in the direction of the predicted probabilities speaks to the uniqueness of the place of racial/ethnic minorities in American politics, even when compared to other disadvantaged groups. It implies that, unlike for other disadvantaged groups, members of Congress who are racial and ethnic minorities are more likely to serve as advocates for their group when the threat environment is high, rather than when it is low. I return to this point for further discussion in the conclusion.
4.7.4 The Advocacy Window
These figures showing the marginal predicted effects also allow for evaluation of the concept of the advocacy window introduced in Chapter 2. The advocacy window is the space between the level of representation that is expected based upon the size of the group in a district, conceptualized as the floor, and the high point of the amount of representation that would be tolerated given the feelings toward a particular group in a district, analogized to the ceiling. On their own, both group size and ambient temperature tend to have a positive, significant relationship with member reputation, as would be expected. But once these variables are evaluated within a single model, group size tends to take precedence, and ambient temperature loses significance. This indicates that, when making decisions about where to focus their reputations, most members tend to stay firmly rooted to the floor when considering the representation of disadvantaged groups.
A level of additional nuance is added, however, when descriptive representation is taken into account. Most other members tend not to shoot for the ceiling when creating their reputations as advocates, but descriptive representatives tend to be more likely to take advantage of an open advocacy window when it exists. This important difference in the effect of the advocacy window on members’ reputational choices is on full display in Figures 4.2 and 4.3. Particularly when considering members who build reputations as primary or secondary advocates for disadvantaged groups, there is a noticeable increase in the responsiveness of descriptive representatives to changes in the ambient temperature – the opening and closing of the advocacy window. Those who are themselves a member of a disadvantaged group tend to utilize this opening of the advocacy window to boost the level of group advocacy they choose to build their reputation around.
Members who are racial/ethnic minorities, however, have an important and unique response to these changes in the dimensions of the advocacy window. For these members, reputations for primary or secondary advocacy of racial/ethnic minorities are more likely to come in districts with a tighter advocacy window. This very likely speaks to the unmatched levels of suspicion and concentrated discrimination racial/ethnic minorities have experienced relative even to other disadvantaged groups. Under such circumstances, members who have themselves experienced this may be most likely to dedicate their reputation to serving racial/ethnic minorities when the threat is high – when they feel if they do not step up for their minority constituents, no one will. Only when the advocacy window is large can they relax, and focus more of their attention on other issues.
This is very different from the situation facing members who enter Congress as a senior, for instance. Zero members of Congress who won their first election to the House of Representatives at or after the age of 65 made the choice to craft a reputation as a primary or secondary advocate for seniors. Very clearly, there is not the same sense that if they do not take these actions, no one else will. Instead, they can feel free to focus their legislative reputations around whatever issues they please, knowing that there will always be other members willing to advocate for seniors.
4.8 Conclusions
The results from this chapter demonstrate the crucial roles that group size, ambient temperature, and descriptive representation have on the legislative reputations that members choose to cultivate. Both group size and ambient temperature tend to have a positive and significant effect on member reputation when considered independently, with group size serving as the more prominent force behind reputation formation when considered together. Descriptive representatives were also more likely to take on reputations as group advocates than other members, with the exception of those disadvantaged groups considered to be most highly deserving.
This chapter also introduced group ambient temperature as a new measure for group affect in a district. It explained the process through which estimates of how positively or negatively a district tends to feel toward a group on average can be generated, making use of the multilevel regression with poststratification procedure. Maps and summary measures were used to provide face validity for these ambient temperature estimates, which were then compared to the variations in group size across districts.
In sum, this chapter provides evidence that reputations for serving disadvantaged groups are rooted both in district demographics such as group size and partisan lean and in member characteristics such as party affiliation and personal experience as a member of a disadvantaged group. Decision-making on legislative reputations in the House of Representatives is well characterized by the concept of the advocacy window, wherein group size is considered to be the floor, or baseline level of representation that would be expected, and ambient temperature serves as a ceiling the members could reach for if they so choose. For most members, this is not a favored choice, preferring instead to root their reputation for advocacy on the floor, and preserve the remainder of their representational capital for other issues.
This trend is evidence of the avoidance behavior expected of a risk-averse member reacting to a potentially unpopular group in their district. Members want to maximize their electoral prospects, and thus are unlikely to expend more than the minimum amount of effort representing groups that have any potential to create a backlash. It is not the case that these members of Congress are doing no work to advocate on behalf of disadvantaged groups in their district, but, as a general rule, they are not maximizing the degree to which their reputations could reflect the advocacy work on their behalf. Descriptive representatives break this pattern by being more likely to increase the part of their reputation that they devote to group advocacy when their district provides them with a wider advocacy window. Racial/ethnic minority members are the exception to this trend, instead being more likely to cultivate a primary or secondary reputation for advocacy of racial/ethnic minorities in instances when ambient temperature is lower, and the advocacy window is reduced.
In the next chapter, I turn to the United States Senate. In it, I utilize a similar framework to determine what drives a senator to form a reputation as an advocate for the disadvantaged.
Open access