1 Bachelor Thesis
Title: The Power of Identification with Social Groups Fenja Elena Krause (s2073102)
Submitted: 18-12-2019 Word Count: 7993
Bachelor Project: Political Behavior: Can we trust Democracy to the Voters? Dr. Joshua Robison
Fall 2019
International Relations and Organizations Leiden University
2 Introduction
The influence of interest group lobbying on contemporary politics is omnipresent. Examples include industrialists like Charles and David Koch, who channel millions of dollars to right wing causes, as well as to subordinated groups that lobby grassroots organizations through the provision of biased information (Goldenberg, 2010).
One example to illustrate the heavy part interest groups play in politics took place in 1982. At this time both houses of Congress agreed that legislation for the withholding of taxes on interest from bank accounts and dividends from securities was needed in order to finally plug a major tax loophole related to unreported income. Just weeks later the House and the Senate changed their opinion completely and repealed the bill 382 to 41 and 94 to 5, respectively (Goldstein, 1999). What caused this rapid change of heart?
To answer that question one needs to look at the parties affected by the bill and what they did in the intermediate period. The passing of the bill would have involved a multi-billion dollar cost for the U.S.-American banking industry (Goldstein, 1999). Taylor (1983) summarized the following efforts to stop the bill from passing by the American Bankers Association (ABA) and the U.S. League of Savings as follows: “the hydrogen bomb of modern day lobbying [was dropped], an effort whose firepower was awesome, whose carnage was staggering. In one fell swoop down went the chairman of the Senate Finance Committee, down went the chairman of the House Ways and Means Committee, down went the Secretary of the Treasury, down went the president of the United States." (Taylor, 1983, p. A12).
The banking industry achieved this by targeting not the politicians directly, but their constituents. The banks sent out monthly statements and took out newspaper advertisements. The goal of this effort was to convince the public that the passing of the bill would have major negative consequences for them individually, as well as for their social group, so that they would contact congress in opposition of the new law. This proved successful: more than twenty-two million constituents communicated their dislike of the bill to their congressman, which eventually led to the law being shelved (Goldstein, 1999).
This is just one example, that shows how powerful lobbying groups are in functional democracies. In addition, the power of indirectly swaying politics through the constituents
3 rather than the politicians themselves, becomes evident. The attempts interest groups make to influence constituents are manifold, but one specific way in which they try to persuade the public is of special interest in this paper: the targeted appeal to identities.
Probably the most prominent example of this tactic is how it is used by the National Rifle Association (NRA) in the United States. A study by Lacombe (2019) demonstrates that the NRA tries to strategically develop and cultivate a gun owner identity, which can in turn be politicized through targeted propaganda, framing gun control policies as existential threats to gun owners. Lacombe (2019) further shows, that the efforts by the NRA are successful: the targeted group has adapted the identity offered by the NRA, and the language used by gun owners to publicly express their interests resembles the one used by the NRA.
This shows, that group identities, or “ideational resources”, are effectively used by interest groups such as the NRA to manipulate the public. Even though Lacombe’s findings are extensive, he suggests more general research focusing on other groups than the NRA, as this specific interest group is distinct in a way that it is also a provider of educational and social programs (Lacombe, 2019). In addition, he only shows how the gun owners adapted the identity and discourse advertised by the NRA, but not how this is reflected in actual political attitudes or decisions. My argument is that, following social identity theory, an individual’s sense of identity is shaped by their group belonging (Tajfel & Turner, 1979). This group identity can be used by interest groups to pursue their objectives. While Lacombe focused on one specific group which members already shared constitutive political attitudes, I assert that this tactic can also be used in less homogenous groups, which members do not necessarily share all the same political standpoints in the beginning. This due to the findings of Tajfel (1970), who concluded that group-identification can be invoked without references to
previously existing attitudes. If this is true, “ideational resources” are an even more powerful tool for interest groups and can possibly explain much more of their success in lobbying the public than previously assumed.
The research question of this paper is to what extent the attempt of interest groups to invoke strong in-group feelings through strategic framing, is more successful in convincing individuals of the lobbyist’s interests, than framing attempts without references to the individual’s group-identity.
4 To answer this question a survey and a quantitative analysis were conducted. The results were ambivalent: only if the individual truly identifies with the faith of his or her group, in-group frames seem to be more effective than identity-neutral frames. Males were differently effected by an in-group frame than females.
Theoretical framework
According to Democratic Theory, a legitimate representative government relies on the mass public to communicate its political attitudes to elite policy makers. Goldberg (2010) has shown, that to understand the nature of mass participation, one needs to understand the role of mass mobilization. If individuals do not speak up or organize themselves into groups, their voices will not be heard and hence not included in the democratic decision-making process.
Therefore the formation of groups representing specific interests is a basic prerequisite for any functional democratic system. It is important to mention that not all interest groups have the same power to assert influence. The party or party coalition that controls the government dictates which interest groups are more or less impactful. When a conservative government is in office in a democratic system, business groups usually have an advantage compared to labor unions and vice versa (Thomas, 2017). An additional factor that is decisive for a groups success is the relative size of organized opposition the group faces. Some groups have natural competitors, for example labor unions and industrial corporations, others, for example those fighting against child abuse, face little resistance. The trustworthiness of a group plays an important role as well (Halwey, 2014). What is true for all groups, following Schattschneider’s (1960) logic of mobilization and participation, is that the “choir leader” of any group has an exceptional amount of power over the choir. Goldberg (2010) uses this metaphor from the musical context to make his argument that interest groups and especially their leading figures, have an extraordinary amount of power over the public.
This is due to the fact, that voters and legislators alike have informational needs, which provide strategic opportunities for lobbying groups to strategically influence their beliefs in the lobbyist’s favor. The tactic of grassroots lobbying makes it possible for the interest group to fill such informational gaps with biased information in favor of their own interests (Dür, 2018). Therefore the public provides an extensive resource for interest groups that want to sway political outcomes. This corresponds to the findings of Kollmann (1998),
5 who investigated the primary targets of interest groups. The majority (56%) focused on the public to convey their messages. Here the aspect of group trustworthiness comes up again. According to Moran (2006), a group’s trustworthiness is a prerequisite if people are to learn, hence to satisfy their informational needs, from the groups public testimony.
The question arises, when and why interest groups make the decision to target the public. According to Dür (2018), this is mainly the case when the group assumes that they are unable to convince elite decision makers directly. By mobilizing and shaping public opinion, they are hoping to have an indirect impact on the politicians through their constituents. This is often done by the creation of campaign websites or online petitions, as well as more traditional means such as putting leaflets and campaign posters into circulation. Greenpeace for example, used all of these tactics in its opposition campaign of the Transatlantic Trade and Investment Partnership (TTIP) (Dür, 2018).
So far the literature tells us that interest groups can use the public’s “ignorance of politics”, to strategically shape public opinion and that they do so when the chances of swaying political decisions are higher through grassroots lobbying than through the direct targeting of political elites (Kinder, 1998). This is often done through various forms of traditional and new media outlets. The next aspect that has to be explained, is what tactics interest groups apply to assert this influence through the media.
Dür (2018) claims, that the usage of issue frames is a very successful way to influence an individual's opinion. According to Druckman (2004), framing refers to the choice to emphasize a specific aspect or consideration of an issue. Issue frames therefore “focus on qualitatively different but potentially relevant considerations” (Druckman 2004, p. 672) of an issue. For example, a speed limit on highways can be framed as a means to reduce traffic accidents or as an unwanted interference of the state into personal decisions, setting limits on individual freedom. To frame an issue in a specific way can change how it is perceived by the public and in turn trigger different political actions. Some interest groups use framing to invoke strong “in-group feelings”.
The largest public policy women organization “Concerned Women for America” offers an example of such a framing attempt: They oppose equal-sex marriage and especially the right of homosexuals to adopt children. The issue of whether or not it should be legal for
6 homosexual couples to adopt children can be framed in two ways: First, as a potential chance for millions of foster kids in the United States to get a new life with loving parents and as a statement that discrimination against homosexuals will no longer be tolerated. Second, as a threat to the “traditional family”, the marriage between man and woman, and as a burden for the adopted child, who will not grow up with a mother and a father. “Concerned Women for America” chose the latter frame and calls upon women to understand their special role in society, as the only ones who can truly fulfill the task of motherhood. This can be seen as an attempt to invoke strong “in-group feelings” amongst women, making them feel needed as a collective, while also claiming that their special role as mothers is threatened by homosexual couples. They also make a very clear distinction between the role of the wife and the husband in a family, claiming that both are important, but could never substitute each other. Hereby they are establishing an “outgroup” (men). For the proponents of a “traditional family”, the opponents of the movement constitute another “outgroup.” (Concerned Women for America, 2019). These concepts will be explained in the next paragraph.
The potential power in- and outgroup feelings can have in manipulating human beings, have been extensively studied. Many of those studies found that groups tend to be more competitive and hostile towards opponents than individuals, a phenomenon known as “the individual-group discontinuity effect” (Meier & Hinsz, 2004). Helevy, Bornstein and Sagiv (2008), add to this that groups are per se cooperative, but only when in-group gains can also be secured without outgroup losses. If this is not the case, a group becomes aggressive
towards the outgroup. Tajfel (1970) describes intergroup conflicts as an end in itself, which is inherently human and helps to release emotional tensions. However, where all these scholars agree, is that the feeling of belonging to a specific group is exceptionally powerful in guiding the decisions of people.
In addition, human beings have a desire for a positive self-concept, which leads them to always see their own social group in a positive light. Allport (1954) and Brewer (1999) concluded that because of these positive in-group feelings, the in-group is slightly more important for the formation of a group identity than the contrast to an outgroup.
A common obstacle in political mobilization is the common-action problem, which refers to a situation in which everyone would be better off if they cooperated with the other members of society, but refuse to do so because of conflicting interests on the individual
7 level, which discourages joint action. While Olsen (1965) argued that the only way to
overcome the common-action problem lays in the offering of material incentives in exchange for cooperation, Lacombe (2019) states that social and psychological benefits can have the same effect. Positive in-group feelings can be seen as such psychological benefits. Hansen (1985) and Wilson (1995) added that such incentives are especially effective when the group perceives their interests as threatened.
In sum, one would expect interest group’s attempts to pursue their ends by framing an identity to be most successful, when they are focusing on the cultivation of a strong, positive in-group feeling, which may or may not be contrasted with negative qualities of an outgroup. This due to the fact, that such in-group feelings offer psychological and social benefits to individuals, which offer higher incentives for political engagement. The informational needs of the politically uneducated, ordinary citizen can be strategically satisfied with biased information, framed in a way to invoke such an in-group-feeling. Strategical identity framing that invokes positive in-group feelings should therefore lead to political outcomes in the lobbyist’s favor. Interest groups decide to go public, when they perceive their chances to sway politics to be higher when targeting the masses than the political elite. To convey their
messages they use different media outlets.
The hypothesis, building on the theoretical framework, is that individuals who have been exposed to a frame that invokes a strong in-group feeling, are more likely to agree more strongly with the interest group’s political goals, than individuals who have been exposed to an identity-neutral frame.
8 Methods
The Sample
To answer the research question, a survey was conducted, the data from which was subsequently analyzed in a quantitative manner. For this the online survey platform “Qualtrics” was used. The respondents of the survey were European, Australian and U.S.-American citizens above the age of 18, because individuals of legal age with the power to vote are the main targets for lobbying efforts by interest groups. The countries stated above have a democratic system which is one of the main prerequisites for the successful operation of interest groups. One can therefore assume, that citizens of these countries are well aware of the power of their opinion and are therefore used to being exposed to a wide range of
opinionated discourse. It is important to note that this is a convenient sample which does not necessarily reflect the exact composition of society, especially because respondents came from different nations.
The Case
The case of interest for this survey experiment is the implementation of the “No Salary History Law” in the U.S.-State Illinois on September 29th, 2019. The bill is barring future employers and job recruiters to seek information about the payment candidates received in their past positions during job interviews. While mainly aimed at narrowing the payment gap between men and women by preventing female candidates from getting stuck in one low payment category throughout their careers, this law also triggers a conflict between a fair and open market and more income equality.
The passing of the law can therefore be framed either as a major breakthrough for all women collectively, or as a law that disregards the forces of the market, hereby making former career achievements inadmissible for job interviews and even putting hard working men at a disadvantage. This due to the fact that the contemporary discourse, which is once again reawakened through the passing of the bill, dictates that women are often treated unfairly in the workplace. While this is mainly true, it can also have the backlashing effect, that women are initially offered a higher salary than men, to avoid a bad reputation for the
9 firm. The risk for this to happen is even higher, when the prior payment of candidates is not taken into account. In addition, employers often ask about salary history because they try to avoid putting individuals, who are expecting a higher salary than available, through the entire application process (Frank, 2017). Due to the passing of the bill, both, men and women, will have to spend more time applying for jobs that turn out to be of no interest for them
financially. This means that while women might gain benefits from this law, as they might artificially be put in a new income category, men will lose out.
Men could therefore make the argument, that the bill only has disadvantages for them. The passing of the law can therefore be strategically framed in a way in which an in-group feeling between men is invoked, because they feel discriminated against. It has already been mentioned in the theory section that the existence of an outside threat by an outgroup (in this case women who want change) is especially effective for the political mobilization of groups.
On the other hand it is obviously also possible to frame the passing of the “No Salary History Law” in a way in which its importance for the fight against the discrimination of women is highlighted. The American interest group “Women Employed” is making use of such a frame, in an attempt to convince the public and especially women, of the positive effects of the law. To achieve this, the interest group states simple facts about gender inequality, while also calling for empathy amongst women.
The theoretical framework states that the ordinary citizen has informational needs, which interest groups try to satisfy with biased information to pursue their preferred ends. The case of the “No Salary History Law” in Illinois is therefore ideal to investigate the effects of filling such a knowledge gap with information framed in a way that appeals to an
individual’s identity. Due to the fact that the law was locally implemented in Illinois, respondents from other states or countries might not have heard of it before, meaning that a knowledge gap actually exists.
The Implementation/ The Survey
After being informed about the survey and asked for their consent, participants are randomly separated into a control group and a treatment group. Each individual, regardless of being in the treatment group or in the control group, is asked for their gender. Corresponding
10 to this choice, the respondent is either redirected to an in-group treatment for men or for women respectively, or, if in the control group, to the control condition for the corresponding gender.
In the treatment condition, respondents are exposed to a descriptive text that explains the scope of the “No Salary History Law”. Afterwards, gender relevant arguments are made. As stated in the theoretical framework, an identity frame is expected to be most successful at leading to political standpoints in the lobbyist’s favor, if it cultivates a strong, positive in-group feeling. Therefore the treatment texts offer frames that aim at invoking such a positive in-group feeling between either men or women, by showing the different advantages or disadvantages the law has for their gender group, while calling for empathy towards members of the own group. Both treatment texts include only pro-law (women) or con-law (men) arguments, respectively. The exact wording of the treatment texts can be found in the appendix.
The literature reviewed in the theory section also suggests that the negative contrast to an outgroup can strengthen the effect of a positive in-group feeling. Therefore the treatment contrasts the in-group (men or women) to an hostile outgroup (the opposite gender), who is posing a threat to a successful career of individuals belonging to the own group.
In light of the literature it is expected that the invoking of such in-group feelings offers enough psychological and social benefits so that respondent’s standpoints towards the law eventually strongly reflect the opinion expressed in the treatment text.
In the control condition, respondents also receive a descriptive text which firstly explains the scope of the law and then offers biased arguments as well. Again women only receive pro-law arguments and men only receive con-law arguments. This due to the fact that the question of interest is not whether a positive or a negative frame makes a difference on how the law is perceived, but the invoking of an in-group feeling through references to the group identity of the participant. The difference between the treatment and the control condition can therefore be found in the nature or the wording of the arguments. While the respondents in the treatment condition are exposed to arguments that aim at the
gender-identity of the respondent, respondents in the control condition read arguments that are gender neutral.
11 Because female individuals in both, the treatment and the control condition, receive only pro-law arguments, and men in both conditions receive only con-law arguments, it is expected that women will have a positive and men will have a negative outlook on the law, regardless of what condition they were in. However, in light of the literature, it is expected that survey respondents in the treatment condition, whose in-group loyalty has been called upon, will agree more with the arguments made in the text. In other words men in the treatment condition are expected to express a more extreme negative standpoint towards the law than men in the control group. Women in the treatment condition are expected to express a more extreme positive standpoint towards the law than women in the control group.
The Composition of the Treatment Texts
As mentioned in the last paragraph, the treatment text for women was mainly received from the interest group “Women Employed”. Women in the treatment condition are first made aware of the general scope of the law: “On September 29th, Illinois’ new “No Salary History Law” went into effect, banning employers from seeking information about the past wages of job applicants during the hiring process” (Women Employed, 2019, para. 1). All other respondents, unrelated to them being in the treatment or control condition, receive the exact same first sentence about the law in general. This due to the fact that not all survey participants have general knowledge about the law and the goal is not to measure the difference between being informed or not, but between receiving relevant or gender-neutral arguments. Afterwards the first gender-relevant pro-law argument is made, stating that the passing of the law is a major advance for gender pay equality. Because women are the ones who are suffering under the payment gap between the genders, this is clearly an argument that tries to invoke the gender-group-identity of women. To highlight this further, the female respondent is informed about the scope of the unfair treatment of women in the U.S. job market: “White women working full time make 77 cents for every dollar paid to white men…” (Women Employed, 2019, para. 1). Hereby, the outgroup threat mentioned earlier is introduced: women are threatened by men, who are wrongfully earning more.
Afterwards an appeal to empathy towards members of the own group (women) is made, to further strengthen a positive in-group feeling: “Questions about the salary history can lead to
us women being stuck in one low income category for the rest of our professional careers”
12 convince the reader that her (positive) opinion of the law is important, because the law might change the unfair conditions her own gender group is facing.
The treatment text for men was not received from a single source, but rather is a compilation of the different arguments opponents of the law make, phrased in a strategic way to invoke an in-group feeling amongst men. The treatment text for men is also substantially longer than the treatment text for women. This due to the fact that it is less obvious how gender, or more specifically being male, is of relevance when it comes to the negative aspects of the law. In order to make this connection as clear as possible, the different arguments had to be elaborated on in more detail.
Again, the characteristics of the law are explained first. Afterwards, two con-law arguments are made: first, that the law makes it harder for future employers to take former career achievements by men into account, second, that the passing of the bill bears the risk that women will have an advantage over men in job interviews, because the contemporary discourse dictates that women are always treated unfairly. Because employers try to avoid the fulfillment of this stereotype, they might initially offer women a higher salary than men. By making this argument, an attempt is made to invoke an in-group feeling through an outside threat: because of the feminist discourse, women might simply get rewarded for their gender, while men’s actual professional achievements might be underestimated.
It is then mentioned that the passing of the law means that the job search for both genders might become longer and more frustrating, while having no positive effect for men. In other words, while women might gain something from the new law, men lose out. This is again aimed at invoking an in-group feeling, because men as a collective will feel the negative effects of the law.
Afterwards an argument is made, that questions about the salary history of job candidates are unrelated to the concept of inequality, because both, men and women, were asked about their former income in the past. This is not related to inequality between the genders, but rather to the normal forces of the market.
To summarize the treatment conditions, the composition of the female treatment is as follows: first, general information about the law, second four gender specific pro-law
13 arguments: the law will 1. help with the persistent gender payment gap, 2. this is important because women still earn at least 23 cents less per dollar than men, 3. this is even worse if the individual is a woman of color 4. women need to be valued the same as men).
The treatment condition for men is composed in a similar way: first, general information about the law, second, four gender specific con-law arguments: the law will 1. make prior professional achievements by hard working men, shown by a high salary, irrelevant in job interviews, 2. due to the contemporary discourse women might get an advantage over men simply because of their gender, 3. job searches become longer and more frustrating, men will lose out and 4. equity means equal treatment, not interference into the market.
In addition to the treatment, all respondents are also asked to rate their standpoints towards the feminist movement, how much they identify with their biological gender and if they think that the law will have personal consequences for themselves. These questions were included to make it possible to control for such factors in the analysis, as Klar (2014) has shown, that individuals have more than one identity which might affect their political position.
After the treatment text, each respondent is asked to rate their support for the “No Salary History Law in Illinois” on a scale from 1 to 5, if 1 means complete support and 5 means complete opposition. Respondents who are not residents of Illinois are asked how much they think that such a law should be implemented in other states/countries as well.
The Composition of the Control Texts
In the control condition, candidates are asked the exact same questions, but they receive a different descriptive text about the “No Salary History Law”. This text also starts by informing the control candidates about the law in general.
Because the female candidates in the treatment condition receive a frame that only consists of pro-law arguments, female candidates in the control condition also only receive pro-law arguments. The difference between the conditions can be found in the nature of these pro-law arguments: while the arguments in the treatment text clearly make emotional
14 references to the gender of the respondent (“Questions about salary history can lead to us
women being stuck in one income category for the rest of our careers...”) (Women Employed,
2019, para. 1), the control text makes the arguments without such an emotional identity-appeal (“Proponents highlight the importance of the law to prevent people from being stuck in one income category throughout their careers.”).
The control condition for the male candidates mirrors the structure of the female version: because male candidates in the treatment condition receive a frame that only consists of con-law arguments, male candidates also only receive con-law arguments. Again, in the control condition, these arguments are made without any reference to the gender-identity of the respondent, while this identity is strongly highlighted in the texts in the treatment condition.
Coding of the variables
For the analysis the data was transferred into an SPSS-file and a dummy variable for each gender was created, indicating whether each individual female or male candidate was in the control or the treatment condition. A value of 0 corresponds to being in the control
condition, a value of 1 means the respondent was in the treatment condition. It is important to create different variables for both genders for all the measurements, because the goal is to compare females in the treatment condition with females in the control condition and males in the treatment condition with males in the control condition.
Two additional variables include information about how much opposition
the respective female or male respondents had for the concept of the law. As mentioned in the methods section, this was measured on a scale from 1 to 5, 1 standing for complete support and 5 standing for complete opposition. The different control variables were added to the file as well.
15 Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
opposition_men 52 1 5 2,46 1,146
opposition_fem 76 1 5 1,96 1,026
Valid N (listwise) 52
A total number of 128 respondents participated in the survey, 52 men and 76 women. The average level of opposition for men was 2,46, for women this average level was 1,96. The Statistical Tests
At first a simple linear regression between the binary variable indicating whether a respondent was in the treatment or the control condition and the variable measuring
opposition towards the law was conducted. The conditions of normality were met for all the variables. The condition of linearity is naturally met for binary variables, which can only take two values (0 and 1). No multicollinarity problem could be detected. In order to introduce control variables to the equation, a step-wise multiple regression and a two-way Anova were conducted as well.
For the females the expectations for the regression are as follows: The hypothesis states that individuals who have been exposed to a strong in-group frame, are more likely to agree with the framed message, than individuals whose group identity has not been invoked. If this hypothesis is correct, females in the treatment condition, who received a pro-law frame in correspondence to their gender-group, should have a more extreme positive opinion of the law than females in the control condition, who only received a simple, group-neutral pro-law frame.
As mentioned, a value of 1 in the grouping variable stands for being in the treatment condition and a number of 0 stands for being in the control condition. Because the opposition variable was measured on a scale on which a low number (starting at 1) stands for much support and a high number (ends at 5) stands for less or no support, a negative relationship (r) is expected. If the value in the grouping variable rises from 0 to 1 (meaning a candidate was in the treatment condition), the corresponding value in the variable measuring the opposition towards the law should drop (because a lower number stands for more support). The concept of “more support” needs to be understood in comparison to the support candidates have
16 towards the law in the control condition. In both conditions female candidates are expected to conceive the law as positive, ergo support it, but a statistical difference in the level of support is expected between the conditions
For the male respondents an effect in the opposite direction is expected. Again, individuals who were exposed to the treatment text were coded with the value of 1 on the grouping variable, individuals who were exposed to the control text were coded with a value of 0. As mentioned above, the variable measuring the opposition a respondent had towards the law was coded on a scale on which a low number (starting at 1) stands for much support and a high number (ends at 5) stands for less or no support. It is expected that male candidates in both conditions, treatment and control, will have negative viewpoints of the law. This due to the fact that both received a negative frame. However, if the hypothesis can be supported, respondents in the treatment condition who were exposed to a gendered frame in
correspondence to their own identity group (men), should have more extreme negative feelings towards the law, than respondents who were exposed to a text that did not call upon this identity.
Therefore, if the value in the grouping variable rises from 0 to 1 (meaning a candidate was in the treatment condition), the corresponding value in the opposition variable should rise as well, because a male respondent who was in the treatment condition is expected to have more negative feelings towards the law, than male respondents who were in the control condition.
17 The Analysis
For the female respondents a simple linear regression was calculated to predict opposition of the “No Salary History Law” based on being in the treatment or control condition:
Effect of being in the treatment group on opposition
Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 2,278 ,164 13,855 ,000 Being in the treatment condition -,603 ,227 -,295 -2,660 ,010
a. Dependent Variable: Opposition towards the law
A significant regression equation was found (p>0.05). Opposition towards the law decreased substantially by .603 on a scale from 1 to 5 for each unit increase in the binary independent variable (meaning a candidate was in the treatment instead of in the control condition). The expectation was fulfilled: for the female candidates, a statistically significant negative
relationship between the grouping variable and the opposition variable was detected. Females in the treatment condition, showed less opposition, or in other words more support, towards the law, than females in the control condition.
For the male respondents a simple linear regression was calculated to predict opposition of the “No Salary History” as well. An effect in the opposite direction as for the female candidates was expected.
Effect of being in the treatment group on opposition
Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 2,500 ,227 11,025 ,000 Being in the treatment condition -,077 ,321 -,034 -,240 ,811
18 No significant regression equation was found. Opposition towards the law decreased by .077 for each unit increase in the binary independent variable (meaning a candidate was in the treatment instead of in the control condition). It is interesting that the effect was not only not significant, but a trend in the opposite direction of what was expected was found: men in the treatment condition showed a trend of having less opposition towards the law, than men in the control condition.
In order to introduce a control variable to the equation, a step-wise multiple regression was conducted for both genders. The following table shows the results for the female
respondents. The first model includes the predictor that was controlled for: the degree to which the respondents identified with the feminist movement.
Effect of being in the treatment group on opposition after being feminist is controlled for (females)
Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 2,196 ,462 4,755 ,000
Identification with the feminist movement
-,072 ,136 -,061 -,528 ,599
2 (Constant) 2,611 ,469 5,565 ,000
Identification with the feminist movement
-,099 ,131 -,085 -,759 ,450
Being in the treatment condition
-,616 ,228 -,302 -2,704 ,009
a. Dependent Variable: Opposition towards the law
As one can see, the control variable had no significant explanatory power over the dependent variable. When the value of the variable measuring the identification with the feminist movement increased by one (meaning the female individual identified more with the feminist movement), the opposition of the candidate measured between 0 and 5 decreased by .072, which is not significant. The main predictor under investigation, whether or not the individual female respondent was exposed to the treatment or the control text, is still
19 beta value for the effect of the main predictor variable on the dependent variable slightly changed between the single linear regression and the multiple hierarchical regression. This is due to the fact that in practice variables are almost always correlated to a certain extent, which has an effect on the beta value when additional explanatory variables are added to the model. What is important to note is that the multicollinarity statistics showed that the correlation between the variables was not problematic (VIF< 4).
Despite the fact that the main predictor of whether a male respondent was in the control condition or not had no statistically significant explanatory power over the dependent variable, a multiple hierarchical regression was run for the male respondents as well. This due to the fact, that it is of interest whether the control variable whether or not a man identified with the feminist movement or not, might explain the variance on the dependent variable instead.
Effect of being in the treatment group on opposition after being feminist is controlled for (males)
Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 3,665 ,632 5,801 ,000
Identification with the feminist movement
-,429 ,218 -,270 -1,965 ,055
2 (Constant) 3,729 ,665 5,609 ,000
Identification with the feminist movement
-,433 ,221 -,273 -1,962 ,056
Being in the treatment group
-,108 ,315 -,047 -,341 ,735
a. Dependent Variable: Opposition towards the law
As one can see, whether a male respondent was a feminist or not had no explanatory power over the dependent variable. When the value of the variable measuring the
identification with the feminist movement increased by one (meaning that the respondents support for the feminist movement increased) the opposition towards the law decreased by -.429, an effect that is not (but almost) statistically significant (p=.055). The main predictor under investigation, whether or not the respondent was exposed to the treatment text or the control text, still has no statistically significant explanatory power over the dependent variable. The trend is still a negative relationship, meaning candidates in the treatment condition opposed the law less than individuals in the control condition (not significant).
20 Again, the correlation between the different predictor variables was statistically not
problematic (VIF< 4).
An important thing to notice is that because of the random assignment to the control and treatment condition, the effects of being feminist or not have less meaning and it is less surprising that they were found to be no significant predictors. This due to the fact that the randomizer is expected to have assigned roughly equal numbers of feminists and opponents of the movement to the two conditions, meaning that the effect of this variable on the
explanatory power of the model should not be too extreme. It is interesting however, whether or not an interaction effect between the grouping variable for the control and the treatment condition and the variable measuring identification with the feminist movement exists. In other words, were feminists more influenced by the different conditions than opponents of the feminist movement? Another variable that is of interest for an interaction effect is the one measuring whether the respondent thought that the law might have personal consequences for him- or herself. This question was asked after the treatment or control text was presented to each participant and might help to access how successful the treatment texts were at invoking an in-group feeling of identification with the own gender group. This due to the fact that an individual who expects the same positive or negative outcomes for himself, that were
explained to affect his gender group, shows signs of identification with the faith of his group. To make the interpretation of the following graphs easier, the variable measuring the extent to which an individual identified with the feminist movement was recoded: instead of having four different categories of identification with the feminist movement, individuals who either strongly or somehow agreed with the movement were coded with a value of 1,
individuals who either strongly or somehow disagreed with the movement where coded with 0.
A two-way Anova with an interaction effect was run, to investigate the questions stated above for both genders. For all following tests, the assumptions of independent observations, homogeneity and normality were met. All interaction effects were not statistically significant and have to be treated as trends which only enable a deeper
21 As one can see, there was a difference in the effect the treatment text had on feminists in comparison to the effect it had on opponents of the movement (the lines do not align). While both, feminists and opponents of the movement, showed lower degrees of opposition towards the law when they were in the treatment condition instead of in the control condition, this effect was stronger for feminists.
.
When it comes to the interaction effect between whether a female individual thought the law could personally effect herself and whether or not this individual was in the treatment group, a
difference was found as well. Individuals who thought that the law will affect themselves (the red line in the graph), reacted more strongly to the treatment text, than individuals who thought the law would have no effect on them.
22 The same interaction effects were also analyzed for the male respondents:
The graph on the left shows the difference in effect of the treatment text between feminists and
opponents of the feminist movement. What is interesting is that for
feminists (the red line), the treatment text had the opposite effect from what was expected: after they read a text that tried to invoke an in-group feeling amongst men, they opposed the law more (this corresponds to the results of the regression analysis). Opponents of the feminist movement however, were affected by the treatment text in the direction that was expected: they opposed the law more than the individuals in the control condition.
This graph shows the difference in effect of the treatment text between the
respondents that thought that the law might personally effect themselves (the red line) and those that thought that the law will have no direct effect on them (the blue line). Individuals who thought that the law will effect themselves reacted towards the treatment text as expected: they opposed the law more than the individuals in the control condition. Those who thought that the law will not affect themselves personally, showed the trend observed in the regression analysis: They opposed the law less in the treatment condition than in the control condition.
23 Interpretation / Reflection
In light of the findings of the analysis, no solid answer to the research question was found. While a significant effect was detected for the female candidates, supporting the hypothesis, the effect for the male candidates was not only not significant, but also contradicts the hypothesis. Instead of being convinced by the con-law arguments made in the gendered frame more than by the con-law arguments made in the gender neutral frame, respondents in the first condition showed a trend of having less opposition towards the law than in the latter condition. Even though this effect is not significant, the trend is still interesting and could possibly have different reasons:
First, being a man and insisting on gender related privileges is becoming less socially acceptable in contemporary Western societies (Matthews, 2017).The contemporary discourse dictates that women are oppressed. Therefore framing an issue in a way that invokes a strong in-group feeling amongst women who just call for justice, does not leave the individual woman feeling guilty for standing up for her gender group. Men on the other hand are increasingly taught that being “a good man” involves making compromises and never discriminating against women (Gilpin & Proulx, 2018).
Standing up for the interest of men (the in-group) is already difficult for the individual man, but opposing the interest of women (the outgroup) is socially unacceptable behavior. Therefore men might be less susceptible to the invoking of positive in-group feelings, because they were socialized to not be proud of their own gender group. Consequently, they might also be more reluctant to express an opinion against a law that was created to fight the gender gap which puts women in a disadvantage. In other words, because of the contemporary discourse, men might not dare to vote for their own interests, if this means an outgroup loss for women. In other words, they are simply less hostile towards the outgroup. This could also explain why men who identified with the feminist movement not only refused to stand up for their gender group, but even opposed the law more after an appeal to their identity as men was made. Feminists outnumbered non-feminists, therefore the overall effect contradicted the hypothesis.
24 The second reason could be that the male treatment text failed at invoking a strong in-group feeling amongst men. Men who thought that the law could have personal negative consequences for themselves, meaning they somehow identified with the faith of their gender group and the treatment was affective, showed a trend of opposing the law more when they were in the treatment condition compared to when they were in the control condition. The opposite effect could only be detected for men who did not think that the law would have negative effects for themselves (a sign that they did not identify with the faith of their gender group). More men belonged to the latter category and therefore the overall effect did not support the hypothesis.
For the females, the interaction analysis showed a trend of feminists reacting more strongly to the treatment text than opponents of the movement, even though identification with the movement alone was no significant predictor for opposition towards the law. If female respondents thought that the law might personally effect themselves, a sign that they truly identified with the faith of their gender group, the treatment text was more effective. These individuals outnumbered the women who did not believe in personal consequences, leading to the overall effect being significant.
Conclusion
The hypothesis for this paper was that individuals who have been exposed to a frame that invokes a strong in-group feeling, are more likely to agree more strongly with the interest group’s political goals, than individuals who have been exposed to an identity-neutral frame. It seems that if individuals truly identify with the faith of their gender group, the hypothesis can be supported. However, if this is not the case, an in-group-frame might even invoke the opposite reaction of what it was designed for. The research question, whether or not the invoking of strong in-group feelings is more successful in convincing individuals of lobbyists interests than identity-neutral frames, can therefore not be answered in a satisfactory way and seems to depend on the context and the identity in question.
While some interesting trends could be observed in this study, some of them even being statistically significant, I call for further studies using a less convenient and more representative sample. A case for which group membership is socially acceptable for both groups might improve the results as well.
25 References:
Allport, G., W. (1954). The Nature of Prejudice. Cambridge, MA: Addison-Wesley.
Brewer, M., B. (1999). The Psychology of Prejudice: Ingroup Love and Outgroup Hate?
Journal of Social Issues 55(3), 429–44.
Concerned Women for America. (2019, July, 15). Defense of Family. Retrieved November 5, 2019, from https://concernedwomen.org/issues/defense-of-family/
Druckman, J. (2004). Political Preference Formation: Competition, Deliberation, and the (Ir)relevance of Framing Effects. American Political Science Review, 98(4), 671-686.
Dür, A. (2018). How interest groups influence public opinion: Arguments matter more than the sources. European Journal of Political Research, 58(2), 514–535.
Frank, L. (2017, September 5). Why Banning Questions About Salary History May Not Improve Pay Equity. Retrieved November 17, 2019, from https://hbr.org/2017/09/why-banning-questions-about-salary-history-may-not-improve-pay-equity.
Gilpin, C. C., & Proulx, N. (2018). Boys to Men: Teaching and Learning About Masculinity in an Age of Change. Retrieved November 15, 2019, from
https://www.nytimes.com/2018/04/12/learning/lesson-plans/boys-to-men-teaching-and-learning-about-masculinity-in-an-age-of-change.html.
Goldenberg, S. (2010, October 13). Tea Party movement: Billionaire Koch brothers who helped it grow. Retrieved November 5, from
https://www.theguardian.com/world/2010/oct/13/tea-party-billionaire-koch-brothers.
Goldstein, K. M. (1999). Interest groups, lobbying, and participation in America. Cambridge: Cambridge University Press.
26 Halevy, N., Bornstein, G. & Sagiv, L. (2008). ‘‘In-Group Love’’ and ‘‘Out-Group Hate’’ as Motives for Individual Participation in Intergroup Conflict - A New Game Paradigm.
Psychological Science, 19(4), 405-411.
Hawley, K. (2014). Trustworthy Groups and Organisations. Retrieved November 17, 2019, from https://static.uni-graz.at/fileadmin/gewi-institute/Philosophie-Gewi/PDFs/hawley-groups-handout-.pdf.
Hansen, J. M. (1985). The Political Economy of Group Membership. American Political
Science Review, 79(1), 79–96.
Kinder, D. (1998). Communication and Opinion. Annual Review of Political Science, 1(1), 167-197.
Klar, S. (2014). Identity and Engagement among Political Independents in America. Political
Psychology, 35(4), 577-591.
Kollman, K. (1998). Outside lobbying: Public opinion and interest group strategies. Princeton, NJ: Princeton University Press.
Lacombe, M. J. (2019). The Political Weaponization of Gun Owners: The National Rifle Association’s Cultivation, Dissemination, and Use of a Group Social Identity. The Journal of
Politics, 81(4), 1342–1356.
Matthews, M. (2017, May 30). Should I feel guilty for being a man? Retrieved November 5, 2019, from https://www.smh.com.au/lifestyle/life-and-relationships/m03aboutlastnight-20170421-gvpkvp.html.
Meier, B.P., & Hinsz, V.B. (2004). A comparison of human aggression committed by groups and individuals: An interindividual-intergroup discontinuity. Journal of Experimental Social
27 Moran, R. (2006). Getting Told and Being Believed. In Lackey and Sosa (Eds.), The
Epistemology of Testimony, (pp. 272-306). Oxford, GB: Oxford University Press.
Olson, M. (1965). The Logic of Collective Action. Cambridge, MA: Harvard University Press.
Schattschneider, E. E. (1960). The Semi-Sovereign People: A Realist's View of Democracy in
America. New York, NY: Holt, Rinehart and Winston.
Taylor, P. (1983, July 31). The Death of Withholding, or How the Banker Won Big.
Washington Post, p. A12.
Tajfel, H. (1970). Experiments in Intergroup Discrimination. Scientific American, 223(5), 96– 102.
Tajfel, H. & Turner, J. (1979). An Integrative Theory of Intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The Social Psychology of Intergroup Relations. Monterey, CA:
Wadsworth.
Thomas, C. S. (2017, July 6). Interest group. Retrieved November 17, 2019, from https://www.britannica.com/topic/interest-group.
28 Appendix:
Male Treatment:
“On September 29th, 2019, Illinois' new "No Salary History Law" went into effect, banning employers from seeking information about the past wages of job applicants during the hiring process” (Women Employed, 2019, para. 1). The idea behind such a law is that questions about salary history could possibly lead to unequal wages between the genders.
If you as a man have earned a high salary due to promotions in a previous job your future employer is not going to be able to take that into account. Because the contemporary
discourse dictates that women are treated unfairly in the workplace they are more likely to get a higher salary offered than men to avoid a bad reputation for the firm (Tinsley & Ely, 2018, Harvard Business Review). In other words, due to the fact that women as a group are often discriminated against, the individual woman might get rewarded because of her gender instead of her achievements, even if no discrimination has taken place. Hereby the bill poses an unfair intervention into the market forces, giving women an advantage over us men.
The reason employers ask about salary history is to ensure they're not putting candidates through the interview process who they can’t afford. By this law interviewees will spend time applying for a job that pays them less. This means that while women might gain benefits from this law, as they will artificially be put in a new income category, we as men will lose out.
Equality means that all future employees are treated in the exact same way. In the past both, women and men were asked about their previous income, which had nothing to do with inequality. We as men can be trapped in one income category in the same way as women, because this is how the market works.
Equality is a noble concept for which most of us strive, but the solution to end discrimination against women cannot be to interfere in the market in a way that discriminates against us men.”
29 Female Treatment:
“On September 29th, 2019, Illinois' new "No Salary History Law" went into effect, banning employers from seeking information about the past wages of job applicants during the hiring process. It's a major advance for gender pay equality-the practice of factoring past wages into salary offers is a major contributing factor to persistent wage gaps. A woman working in the U.S. today reckons with a persistent wage gap that can rob her of economic stability. White women working full time make 77 cents for every dollar paid to white men. If she is a woman of color, the impact is all the more dire." (Women Employed, 2019, para 1). Questions about the salary history can lead to us women being stuck in one low income category for the rest of our professional careers.”
Female Control:
“On September 29th, 2019, Illinois' new "No Salary History Law" went into effect, banning employers from seeking information about the past wages of job applicants during the hiring process” (Women Employed, 2019, para. 1). Proponents highlight the importance of the law to prevent people from being stuck in one single income category throughout their careers. A low income can lead to individuals not being able to make ends meet and makes family planning and retirement harder.”
Male Control:
“On September 29th, 2019, Illinois' new "No Salary History Law" went into effect, banning employers from seeking information about the past wages of job applicants during the hiring process” (Women Employed, 2019, para. 1). Opponents of the concept argue that the law creates unnecessary regulation, which can have negative, discriminatory consequences because it will be harder to take former professional achievements into account. Furthermore information about the salary history of candidates is often times helpful in assessing the market, particularly for small employers. Searching for a job might become more stressful and time-intensive as well.
