How naming and shaming affects human rights perceptions in the shamed country

Abstract

Although individual citizens perceive the human rights conditions in their country differently, existing research on human rights and public opinion has tended to ignore the possible impact from international sources of information. This article builds upon previous research on human rights, public opinion, and international organizations by arguing that citizens will perceive the human rights conditions in their country more negatively when their country is shamed by the international community for human rights violations. This hypothesis is tested using both observational and experimental methods. I use multilevel modeling and survey responses from both the Gallup International Millennium Survey and the World Values Survey to show that individual human rights perceptions are negatively related to the adoption of resolutions by the United Nations Commission on Human Rights shaming a survey respondent’s government for human rights abuses, as well as the number of human rights-focused international nongovernmental organizations with members or volunteers living within the survey respondent’s country. These results are supplemented by an original, web-based experiment on human rights perceptions in the United States and India in which survey respondents are exposed to an experimental manipulation modeled on press statements from Amnesty International shaming the survey respondent’s country for human rights abuses.

Keywords

human rights naming and shaming public opinion

In every region of the world, governments abuse the basic human rights of their citizens through the use of torture and violence and by stifling freedom of speech, assembly, and religion, among other things. But there is reason for optimism. Today there are much stronger international norms and laws regarding human rights, and the naming and shaming of abusive governments has become more prevalent, although scholars disagree on whether it is effective (Franklin, 2008; Krain, 2012), or not (Bob, 2005; Hafner-Burton, 2008).

This article does not focus on abusive governments, but on individual people. As with any political issue, individuals will perceive the human rights conditions in their country differently. I argue that the naming and shaming of abusive governments negatively affects individual human rights perceptions within the shamed country by providing citizens of the shamed country with more negative information regarding the local human rights conditions. I use both observational and experimental methods to show that this is the case. I use multilevel modeling and survey responses from both the Gallup International Millennium Survey and the World Values Survey to show that the citizens of shamed countries have more negative perceptions of their country’s human rights conditions, even when controlling for multiple objective measures of the conditions within their country (including the PTS, CIRI Physical Integrity Index, and CIRI Empowerment Rights Index). I also run an original, web-based experiment (using Amazon’s Mechanical Turk) on human rights perceptions in the United States and India that constitutes a more direct test of the relationship between individual exposure to naming and shaming and human rights perceptions.

Although I use the bulk of the article to discuss the relationship between naming and shaming and human rights perceptions, in the conclusion I elaborate upon how a relationship between naming and shaming and public opinion could fit into the larger story of rebellion and repression.

Shaming and perceptions

On 4 May 2012, Amnesty International publicly shamed the Iranian government for imprisoning a political activist, claiming that the ‘nine-year jail term for a prominent human rights lawyer is another nail in the coffin for freedom of expression and association in Iran and should be quashed immediately’ (Amnesty International, 2012). Similarly, in 2007 the United Nations Human Rights Council (UNHRC) passed a resolution shaming the government of Myanmar for its violent repression of peaceful protests and demonstrations. Naming and shaming has been used as a tool against abusive governments by international organizations (IOs), international nongovernmental organizations (NGOs), and other countries. But who is the intended audience of that shame? Certainly the UNHRC resolution shaming the government of Myanmar was noticed by that country’s government. But it was also seen by other governments as a signal to modify their behavior toward Myanmar. Lebovic & Voeten (2009), for example, argue that the adoption of public resolutions by the United Nations Commission on Human Rights (the predecessor to the UNHRC) provided the necessary information and political cover for international financial institutions to adopt sanctions against shamed governments. The authors find that the adoption of a UNCHR resolution caused a significant reduction in foreign aid to the shamed government and, perhaps more interestingly, that objective measures of the human rights conditions within the shamed country had no effect on foreign aid, suggesting a strong institutional influence upon government behavior (Lebovic & Voeten, 2009: 79).

But what about the general population of shamed countries? Carlson & Listhaug (2007) use survey data from the Gallup International Millennium Survey (GIMS) to examine the micro- and macro-level determinants of human rights perceptions, including a survey respondent’s gender, age, and education level, and their country’s regime type and GDP per capita, among other things. They find that human rights perceptions are negatively affected by the number of human rights violations within a survey respondent’s country (as determined by objective experts), suggesting that ‘the concept of human rights may share similar meanings in the minds of citizens and experts for at least some aspects of human rights values’ (Carlson & Listhaug, 2007: 465). The authors also find that respondents who believe that their country is governed by the will of people or who are highly educated tend to have better perceptions of the local human rights conditions, presumably because the highly educated ‘might be expected to come from the most privileged social classes’ (Carlson & Listhaug, 2007: 473) and might therefore experience fewer violations in the first place.

Previous analyses of individual human rights perceptions have included independent variables reflecting only internal sources of information (Holsti, 1996; Carlson & Listhaug, 2007). For example, the regime type of a survey respondent’s country can be determined solely by looking within that respondent’s country. The same is true for an individual’s personal characteristics, including their gender, age, and education level. When limiting their analyses to internal sources of information, scholars implicitly assume that individuals are unaffected by external sources of information, or by world politics more generally. The 2011 Arab Spring, however, demonstrates that individuals are indeed aware of international events, and that such events can change an individual’s perceptions of both what is normal and what is possible within their own country – often referred to as the ‘CNN Factor’ (Livingston & Eachus, 1995). I suspect that the information used by individual citizens to form their perceptions of the local human rights conditions comes from both internal sources and international sources.

Previous research on naming and shaming has itself been limited, concentrating solely on the response to that shame by the targeted government (Hafner-Burton, 2008; Franklin, 2008) and by other governments toward the targeted government (Lebovic & Voeten, 2009). What scholars have therefore produced is a body of research in which the individual-level outcome (human rights perceptions) has been explained using only individual- and country-level variables, and an international-level outcome (naming and shaming) that has only been linked to country- and international-level outcomes of interest (bilateral aid, government repression, etc.). Given that scholars still disagree on whether naming and shaming affects government behavior (Franklin, 2008; Krain, 2012) or not (Bob, 2005; Hafner-Burton, 2008), it is useful to consider what other consequences naming and shaming may have within the shamed country, because if those consequences affect government behavior then it might be necessary to control for them in order to isolate the effect, whether direct or indirect, from naming and shaming upon government behavior.

Previous research has already shown that individual perceptions can be influenced by international organizations. For example, Chapman (2009) shows that the authorization of foreign policy decisions by the UN Security Council often increases public support for those decisions. Grieco et al. (2011) use an experimental analysis to show that institutional endorsements can increase public support for military force within the survey respondent’s country. Previous research has also implied that a link between naming and shaming and human rights perceptions should exist. Cavallaro & Brewer (2008) argue that international courts are more likely to rule on human rights cases when they believe that doing so will help galvanize the local population, implying that citizens are – to some extent – aware of those courts’ decisions (although not directly testing whether this is the case). Hafner-Burton (2008: 693) claims that the ‘global publicity from NGOs, the news media, or the UN could have the accidental side effect of providing incentives for groups to orchestrate acts of violence large enough to attract the spotlight’ and that ‘governments sometimes react to these security challenges by repressing human rights even further’, suggesting that naming and shaming might actually increase human rights violations in the shamed country because its citizens are made aware of that shame.

Whether individual people are its intended audience or not, naming and shaming does provide individuals with information about the local human rights conditions. It stands to reason that that information will have some impact on the formation of individual human rights perceptions.¹ Hayes & Guardino (2011) show that US public opinion regarding the 2003 invasion of Iraq was affected by the appearance of foreign diplomats on US mass media. The authors find that self-identified Democrats were especially affected by information coming from foreign voices due to their ‘individual-level partisan-ideological predispositions’ (Hayes & Guardino, 2011: 832). Although criticism from foreign diplomats is not functionally equivalent to international naming and shaming, it is indeed similar, and its effect on US public opinion suggests that external sources of information such as naming and shaming could affect public opinion. Davis, Murdie & Steinmetz (2012) have already shown that naming and shaming by international nongovernmental organizations (INGOs) with a human rights focus has a negative impact on aggregate public opinion within the shamed country. I suspect that the same relationship will be found at the individual level as well, and for naming and shaming more generally. Because naming and shaming provides negative information about a country’s human rights conditions, I expect that – all else being equal – naming and shaming will have a negative effect upon individual human rights perceptions within the shamed country. In other words, I expect that individuals will update their beliefs to reflect more negative attitudes when they are exposed to more negative information.

It is possible that some individuals may have a more nuanced reaction to negative information. For example, Nyhan & Reifler (2010) find evidence of a ‘backfire effect’, where survey respondents who are provided information that runs counter to their predisposed beliefs find that their resolve in those beliefs is strengthened. Similarly, citizens who hold their government in high esteem may find that their support for the government is strengthened whenever their country is criticized by the international community. It is also possible that the specific source of international shame might affect an individual’s reaction. For example, survey respondents in Iran or Cuba might take criticism from the United States as a badge of honor. On average, however, I expect that citizens will perceive local human rights more negatively when they receive negative information regarding the local human rights conditions. Future research could explore in more detail whether an individual’s political ideology, patriotism, and attitude towards other countries opens up the possibility of a positive reaction to negative information.

Turning to specific sources of naming and shaming, Lebovic & Voeten (2006) find that the adoption of a UNCHR resolution shaming a country’s human rights conditions causes a significant reduction in foreign aid to the government of that country. Similarly, I suspect that the adoption of a UNCHR resolution will have a negative effect on individual perceptions of the local human rights conditions. I make the following hypothesis:

Hypothesis 1: The adoption of a UNCHR resolution shaming a survey respondent’s country for human rights violations has a negative effect on the respondent’s perception of the human rights conditions within their country.

Although the United Nations is the largest and most visible organization to take a stand on human rights worldwide, there are a sizable number of smaller international nongovernmental organizations (INGOs) that are predominantly human rights focused and that broadcast their criticisms of countries’ human rights records within public forums (Murdie & Davis, 2012). I expect that such criticism will have a similarly negative effect on individual human rights perceptions within the shamed country. I make the following hypothesis:

Hypothesis 2: The public shaming of a survey respondent’s country for human rights violations by international human rights organizations has a negative effect on the respondent’s perception of the human rights conditions within their country.

Not only do human rights focused INGOs publish their criticisms in public forums, they send volunteers to those countries that are experiencing especially high volumes of human rights violations. It stands to reason that the presence of human rights focused INGOs within a country should result in the dissemination of negative information regarding the conditions in that country, which should have a negative effect upon individual perceptions within that country. I make the following hypothesis:

Hypothesis 3: The presence of human rights focused INGOs within a survey respondent’s country has a negative effect on the respondent’s perception of the human rights conditions within their country.

Observational analysis

Previous research has shown that there is a relationship between naming and shaming and objective measures of human rights conditions (Hafner-Burton, 2008), and between objective measures of human rights conditions and individual human rights perceptions (Carlson & Listhaug, 2007). I isolate the effect from naming and shaming upon individual human rights perceptions by using human rights perceptions as the dependent variable and naming and shaming as the independent variable, and controlling for objective measures of the human rights conditions within each country.

Dependent variables

I measure human rights perceptions using survey responses from multiple surveys. First, the Gallup International Millennium Survey (GIMS), which gauged opinions on a wide variety of issues, but most importantly for the purposes of this study, on individual perceptions of the human rights conditions within survey respondents’ countries (Gallup International, 2000). Because the survey was collected entirely within the year 1999, I sacrifice temporal variation and assemble a cross-sectional dataset using GIMS responses. The dependent variable is coded 1 if a respondent believes that human rights are not generally respected in their country, 2 if partially respected, and 3 if fully respected.

Next, I use the World Values Survey (WVS, 2009). The WVS was collected across multiple years, but since there are no instances of human rights perceptions being measured in the same country across multiple years, there is no real temporal variation in the dataset that I assemble using WVS survey responses, even though different countries were collected in different years. The WVS asks survey respondents to rate the human rights conditions in their country on a 4-point scale. The dependent variable is coded 1 if a respondent believes that there is no respect for individual human rights nowadays, 2 if there is not much respect, 3 if there is some respect, and 4 if there is a lot of respect.²

Independent variables

I use three different measures of naming and shaming, all of which I expect to be negatively correlated with human rights perceptions. First, I employ the data by Lebovic & Voeten (2006) on the adoption of UNCHR resolutions publicly shaming a survey respondent’s country. UNCHR resolutions were used specifically to shame governments for violating some part of the Universal Declaration of Human Rights. The variable is coded 1 if the UNCHR passed a resolution shaming a survey respondent’s country during the year prior to their survey response being collected; 0 otherwise. In other words, the variable is lagged one year behind the dependent variable, giving survey respondents enough time to have been made aware of the resolution. For the GIMS dataset, in which all survey responses were collected during the year 1999, the variable is coded 1 if the UNCHR passed a resolution shaming a survey respondent’s country during the year 1998. For the WVS dataset, which was collected in different years for different countries, the variable is coded 1 if the UNCHR passed a resolution shaming a survey respondent’s country during the year prior to the year that the WVS collected survey responses within that country.³

Next, I employ the variable HRO Shaming from the events dataset by Murdie & Davis (2012). The variable measures the number of instances in which an international nongovernmental organization (INGO) with a human rights focus criticized a survey respondent’s country and then broadcast that criticism in a public forum (specifically, Reuters) during the year prior to that person’s survey response being collected. Some examples of the organizations included are the African Institute of Human Rights, Central America Human Rights Committee, International Committee for the Defence of Human Rights in Iran, Union of Arab Jurists, and World Habeas Corpus.

Finally, I employ the variable HRO Presence, which indicates the number of human rights focused INGOs that had members and/or volunteers living within a survey respondent’s country during the year prior to their survey response being collected (Smith & Wiest, 2005). Although HRO Presence does not measure a specific form of naming and shaming, INGOs with a human rights focus are more likely to send members or volunteers to countries that are experiencing human rights violations. Therefore, the information that they provide about the local human rights conditions is more likely to be negative than positive, and their presence should result in more negative information being provided to the local population regarding their country’s human rights conditions. Therefore, I expect that the variable HRO Presence will be negatively related to individual human rights perceptions.

I use multiple objective measures of the human rights conditions within survey respondents’ countries, all of which I expect to be negatively correlated with human rights perceptions. This expectation is derived from previous research (Carlson & Listhaug, 2007) as well as the rather intuitive sensibility that, all else being equal, more human rights violations should result in a more negative view of the local human rights conditions, despite the inevitable variation in what the citizens of different countries believe constitutes a human rights violation.

First, I use the Political Terror Scale (PTS).⁴ The variable is measured on a discrete scale from 1 to 5, with 1 indicating governments that fully respect citizens’ rights, and 5 indicating governments that employ a broad use of political terror and egregious violations of human rights (Gibney, Cornett & Wood, 2007). Next, I use two variables from the CIRI dataset on human rights (Cingranelli & Richards, 2010), both of which I inverted so that higher scores indicate higher levels of human rights violations. The CIRI Physical Integrity Rights Index (CIRI Physical) reflects the levels of torture, political imprisonment, extrajudicial killing, and disappearances committed by a government during a given year, and is measured on an inverted, discrete scale from 0 to 8, with 0 indicating full government respect for those rights, and 8 indicating no government respect for those rights. The CIRI Empowerment Rights Index (CIRI Empowerment) reflects the levels of freedom of speech, freedom of religion, freedom of assembly and association, etc. granted to a country’s citizens during a given year, and is measured on an inverted, discrete scale from 0 to 14, with 0 indicating full government respect for those rights, and 14 indicating no government respect for those rights. Although the CIRI Empowerment measure is markedly different than the PTS and CIRI Physical measures on a conceptual level, I expect that the relationship between the shaming variables and human rights perceptions will be robust to the inclusion of any of the objective measures of human rights conditions, regardless of any conceptual differences between them. For all three measures (PTS, CIRI Physical, and CIRI Empowerment), I use the score assigned to a survey respondent’s country during the year prior to each survey response being collected.⁵

Control variables

I control for several variables at both the respondent level and country level that might affect human rights perceptions and/or have been utilized in previous studies (Carlson & Listhaug, 2007; Anderson, Regan & Ostergard, 2002).

The GIMS and WVS ask survey respondents for their age, gender, and education level. I control for these variables in case there is a discernable difference in human rights perceptions between respondents who are male or female, young or old, highly educated or not. In both surveys, gender is recorded as 1 if the respondent is female, 0 if male. Age and education are measured differently in each survey. The GIMS measures education on a discrete scale from 1 to 4, with 1 indicating no education and 4 indicating a university degree, and age on a discrete scale from 1 to 7, where 1 indicates that the survey respondent is very young and 7 indicates that the survey respondent is very old. The WVS measures age in years, and education on a discrete scale from 1 to 8, with 1 indicating that the respondent inadequately completed an elementary education and 8 indicating a university degree. Carlson & Listhaug (2007) find that human rights perceptions are positively related to education and age, and negatively related to whether an individual is female.

I control for the level of political development that survey respondents perceive within their country. For GIMS survey respondents, I control for whether they believe that their government reflects the will of the people (yes/no). Carlson & Listhaug (2007) find a strong, positive correlation between responses to this question and human rights perceptions. For WVS survey respondents, I control for their opinion of the political system in their country, measured on a discrete scale from 1 to 10, with 1 indicating bad and 10 indicating good. I also control for an objective measure of political development: the winning coalition size in a survey respondent’s country during the year prior to their survey response being collected (Bueno de Mesquita et al., 2003).⁶

Although scholars have produced conflicting evidence regarding the relationship between economic development and human rights perceptions (Carlson & Listhaug, 2007) or the lack thereof (Anderson, Regan & Ostergard, 2002), economic development has also been linked with decreases in government repression (Poe, Tate & Keith, 1999). Therefore, I control for the level of economic development within survey respondents’ countries in my analysis of human rights perceptions. It is operationalized in the GIMS dataset as the logged GDP per capita in a survey respondent’s country during the year prior to the survey being taken (measured in thousands of US dollars). It is operationalized in the WVS dataset as the extent to which survey respondents are employed. The variable is measured as a discrete scale from 1 to 8, with 1 indicating full employment and 8 indicating unemployment. Although GDP per capita and individual employment are markedly different variables, they are similar enough that they should each account for some of the effect that economic development has upon government repression (Poe, Tate & Keith, 1999).

Finally, I control for the level of press freedom in a survey respondent’s country during the year prior to their survey response being collected. The variable is constructed using the annual Freedom of the Press reports by Freedom House (2000). It is measured on a discrete scale from 1 to 100, with higher values indicating higher levels of press freedom.⁷

The spatial-temporal domains for the datasets were determined by the total availability of the dependent and independent variables. The GIMS dataset has approximately 20,000 observations spanning 24 countries. The WVS dataset has approximately 30,000 observations spanning 26 countries.⁸ Descriptive statistics for the GIMS and WVS datasets, as well as a complete breakdown of the dependent variables by country, are provided in the online appendix.

Multilevel modeling

The basic statistical model predicting human rights perceptions is as follows:

\begin{aligned} P e r c e p t i o n s = α + b_{1} * O b j e c t i v e m e a s u r e o f c o n d i t i o n s \\ + b_{2} * N a m i n g a n d s h a m i n g \\ + b_{3} \dots b_{N} * C o n t r o l s + ϵ \end{aligned}

The model is specified separately for the GIMS and WVS datasets. Perceptions refers to individual survey responses regarding the human rights conditions in the respondent’s country. Objective measure of conditions refers to one of three expert-based measures of human rights conditions (PTS, CIRI Physical, and CIRI Empowerment). Naming and shaming refers to one of three measures of the naming and shaming of a respondent’s country during the year prior to their survey response being collected.

Because the datasets are hierarchically nested (with citizens nested inside countries), I modify Equation 1 using multilevel modeling techniques. The results from the two approaches (fixed effects OLS and multilevel modeling) tend to be similar, although multilevel modeling decreases the likelihood of committing a Type I error. Furthermore, the techniques used in estimating multilevel models have become more advanced over the past decade, allowing researchers to ‘build models that capture the layered structure of multilevel data, and determine how layers interact and impact a dependent variable of interest’ (Steenbergen & Jones, 2002: 218).⁹ I take the approach outlined in Steenbergen & Jones (2002) and allow the regression parameters for the micro-level variables (respondents’ gender, age, etc.) to vary across countries. The regression parameters for the macro-level variables remain fixed. See Equation 2:¹⁰

\begin{aligned} {P e r c e p t i o n s}_{i j} = α_{i} + b_{1} * O b j e c t i v e m e a s u r e o f c o n d i t i o n s \\ + b_{2} * N a m i n g a n d s h a m i n g \\ + b_{3 j} \dots b_{N j} * M i c r o - l e v e l c o n t r o l s \\ + b_{4} \dots b_{N} * M a c r o - l e v e l c o n t r o l s + ϵ_{i j} \end{aligned}

Perceptions $_{i j}$ is the micro-level dependent variable measured at the survey respondent level, with respondent i nested within country j. The regression parameters for the micro-level independent variables are allowed to vary across countries (as identified by the j subscript on each parameter). This equation reflects the fixed nature of the macro-level variables, whose values do not vary across respondent units within each country and whose regression parameters should therefore remain fixed across each macro-level unit j. MLwiN software has made the estimation of such models relatively painless (version 2.00; Rasbash et al., 2000). I estimate the statistical model in Equation 2 separately for each survey, using every combination of naming and shaming (UNCHR resolutions, HRO Shaming, and HRO Presence) and every objective measure of human rights conditions (PTS, CIRI Physical, and CIRI Empowerment). I provide the results in Tables I– III.

Results

The results in Tables I–III provide strong support for Hypotheses 1 and 3, and mixed support for Hypothesis 2. The coefficient for the variable UNCHR Resolution is negative and statistically significant across all six model specifications in Table I – across multiple surveys and multiple objective measures of human rights conditions – suggesting that survey respondents are affected not only by their actual conditions, but by the international community putting their conditions into perspective and telling them that their conditions have worsened. The results for the variable HRO Shaming are inconsistent across the two surveys, showing a negative relationship with individual human rights perceptions in the GIMS dataset, and a positive relationship in the WVS dataset (see Table II). The coefficient for the variable HRO Presence is negative and statistically significant across five of the six model specifications in Table III.

Table I.

Multilevel analysis of human rights perceptions and UNCHR resolutions

	GIMS			WVS
Political terror scale $_{t - 1}$	−0.159***			−0.013
	(0.029)			(0.044)
CIRI physical $_{t - 1}$		−0.065***			0.156***
		(0.010)			(0.034)
CIRI empowerment $_{t - 1}$			0.021*			0.024
			(0.009)			(0.013)
UNCHR resolution $_{t - 1}$	−0.297***	−0.204**	−0.225**	−0.396***	−0.417***	−0.457***
	(0.073)	(0.071)	(0.076)	(0.084)	(0.078)	(0.091)
Population $_{t - 1}$	0.071***	0.077***	0.035	0.166***	−0.008	0.141***
	(0.018)	(0.018)	(0.018)	(0.024)	(0.040)	(0.019)
GDP per capita $_{t - 1}$	−0.112***	−0.108***	−0.085***
	(0.024)	(0.024)	(0.025)
Respondent: unemployed? $_{t}$				−0.001	−0.001	−0.002
				(0.002)	(0.002)	(0.002)
Winning coalition $_{t - 1}$	−0.002	−0.114	−0.219	−0.139	−0.160	−0.051
	(0.127)	(0.127)	(0.146)	(0.128)	(0.117)	(0.127)
Press freedom $_{t - 1}$	−0.003	0.001	0.004	−0.003	0.004*	0.000
	(0.002)	(0.001)	(0.002)	(0.002)	(0.002)	(0.003)
Respondent gender $_{t}$	−0.000	−0.001	0.000	0.012	0.014	0.016
	(0.011)	(0.011)	(0.011)	(0.026)	(0.026)	(0.026)
Respondent age $_{t}$	0.002	0.000	0.001	0.002	0.002	0.002
	(0.006)	(0.005)	(0.006)	(0.001)	(0.001)	(0.001)
Respondent education $_{t}$	0.015	0.015	0.017	0.001	0.001	0.001
	(0.009)	(0.009)	(0.010)	(0.005)	(0.005)	(0.006)
Respondent: Will of the people $_{t}$	0.209***	0.210***	0.211***
	(0.020)	(0.021)	(0.020)
Respondent: Political system approval $_{t}$				0.091***	0.090***	0.091***
				(0.008)	(0.008)	(0.008)
Constant	2.019***	1.552***	1.643***	−0.628	1.272**	−0.610*
	(0.312)	(0.320)	(0.364)	(0.337)	(0.493)	(0.308)
N	18,474	18,474	18,474	26,553	26,553	26,553

Maximum likelihood estimates using MLwiN 2.00. Standard errors are shown in parentheses. ***p $<$ 0.001, **p $<$ 0.01, *p $<$ 0.05.

Turning to the control variables, I find no consistent relationship between the objective measures of human rights conditions and human rights perceptions. I had expected the regression parameters for all three measures to have negative coefficients, reflecting the rather intuitive sensibility that an increase in human rights violations should result in more negative human rights perceptions. But the results for all three measures are inconsistent across the two surveys. For example, the Political Terror Scale is statistically significant and negatively related to human rights perceptions in the GIMS dataset (Tables I–III, Model 1), but statistically insignificant in the WVS dataset (Tables I–III, Model 4). Similarly, the CIRI Physical Integrity Rights Index is negatively correlated with human rights perceptions in the GIMS dataset (Tables I–III, Model 2), but positively correlated in the WVS dataset (Tables I–III, Model 5). This suggests one (or all) of the following: (1) expert-based measures cannot adequately predict human rights perceptions in survey respondents’ countries because they cannot account for the shift (perhaps cultural?) in local perceptions of what constitutes a human rights violation. Previous literature had suggested otherwise, claiming that ‘the concept of human rights may share similar meanings in the minds of citizens and experts’ (Carlson & Listhaug, 2007: 465); (2) the information provided by naming and shaming could have more effect on human rights perceptions than the information provided by personal experiences, or by what citizens experience and see around them (in other words, context matters);¹¹ or (3) the existing data on human rights violations and human rights

Table II.

Multilevel analysis of human rights perceptions and HRO shaming

	GIMS			WVS
Political terror scale $_{t - 1}$	−0.135***			0.010
	(0.031)			(0.044)
CIRI physical $_{t - 1}$		−0.062***			0.187***
		(0.011)			(0.035)
CIRI empowerment $_{t - 1}$			0.026**			−0.013
			(0.010)			(0.012)
HRO shaming $_{t - 1}$	−0.284*	−0.134	−0.449***	0.035*	0.030	0.032
	(0.137)	(0.132)	(0.149)	(0.018)	(0.018)	(0.019)
Population $_{t - 1}$	0.074***	0.079***	0.049**	0.144***	−0.050	0.159***
	(0.019)	(0.018)	(0.019)	(0.025)	(0.041)	(0.021)
GDP per capita $_{t - 1}$	−0.075**	−0.074**	−0.052
	(0.028)	(0.027)	(0.028)
Respondent: unemployed? $_{t}$				−0.002	−0.00	−0.002
				(0.003)	(0.003)	(0.003)
Winning coalition $_{t - 1}$	−0.158	−0.137	−0.460*	0.140	0.173	0.086
	(0.192)	(0.175)	(0.179)	(0.141)	(0.120)	(0.142)
Press freedom $_{t - 1}$	−0.001	0.002	0.006*	−0.000	0.008*	−0.002
	(0.002)	(0.001)	(0.002)	(0.002)	(0.003)	(0.003)
Respondent gender $_{t}$	0.003	0.002	0.004	0.029	0.032	0.028
	(0.011)	(0.011)	(0.011)	(0.027)	(0.025)	(0.027)
Respondent age $_{t}$	−0.001	−0.002	−0.001	0.002	0.002	0.002
	(0.006)	(0.005)	(0.006)	(0.001)	(0.001)	(0.001)
Respondent education $_{t}$	0.012	0.012	0.014	−0.001	−0.000	−0.001
	(0.010)	(0.010)	(0.010)	(0.005)	(0.005)	(0.005)
Respondent: Will of the people $_{t}$	0.208***	0.207***	0.208***
	(0.021)	(0.022)	(0.021)
Respondent: Political system approval $_{t}$				0.092***	0.092***	0.093***
				(0.009)	(0.009)	(0.009)
Constant	1.631***	1.217***	1.199***	−0.682	1.377**	−0.695*
	(0.363)	(0.344)	(0.395)	(0.370)	(0.495)	(0.333)
N	16,730	16,730	16,730	24,193	24,193	24,193

Maximum likelihood estimates using MLwiN 2.00. Standard errors are shown in parentheses. ***p $<$ 0.001, **p $<$ 0.01, *p $<$ 0.05.

perceptions are not sophisticated enough for an observational analysis. The GIMS and WVS do not, after all, ask survey respondents whether they have personally experienced any human rights violations.

Experimental analysis

Because the GIMS and WVS surveys do not specifically ask survey respondents whether they are aware of any naming and shaming directed towards their country, the observational analysis does not constitute a direct test of the causal mechanism (that human rights perceptions are affected by individual exposure to naming and shaming). The statistical significance of the binary variable indicating the adoption of a UNCHR resolution could be the result of something else that those countries receiving a UNCHR resolution have in common (for example: regime type, economic development, or religious homogeneity). In other words, without more direct evidence of a link between naming and shaming and individual perceptions, it is unclear whether naming and shaming appears to have a significant effect because it actually affects perceptions or because it serves as a proxy for something else not included within the model (e.g. omitted variable bias). Moreover, without any temporal variation in the dependent variable, I run the risk of conflating the overall attitude toward shamed governments with the change in human rights perceptions induced by shaming. Indeed, if countries are selected for shaming because their populations are at odds with the government, then it is possible that the reverse relationship exists. This problem is ameliorated – to some extent – by the inclusion of the objective measures of human rights conditions, which strongly affect the countries that are selected for shaming. Ultimately, however, the lack of temporal variation in the

Table III.

Multilevel analysis of human rights perceptions and HRO presence

	GIMS			WVS
Political terror scale $_{t - 1}$	−0.106***			0.003
	(0.032)			(0.047)
CIRI physical $_{t - 1}$		−0.065***			0.219***
		(0.011)			(0.042)
CIRI empowerment $_{t - 1}$			−0.002			−0.037**
			(0.010)			(0.014)
HRO presence $_{t - 1}$	−0.218**	−0.257***	−0.324***	−0.088	−0.124	−0.182*
	(0.077)	(0.072)	(0.082)	(0.067)	(0.073)	(0.072)
Population $_{t - 1}$	0.111***	0.128***	0.110***	0.151***	−0.112*	0.198***
	(0.027)	(0.027)	(0.029)	(0.027)	(0.056)	(0.031)
GDP per capita $_{t - 1}$	−0.022	−0.017	0.013
	(0.035)	(0.034)	(0.036)
Respondent: Unemployed? $_{t}$				−0.003	−0.003	−0.003
				(0.003)	(0.003)	(0.003)
Winning coalition $_{t - 1}$	−0.039	−0.071	−0.358*	0.142	0.183	−0.045
	(0.148)	(0.121)	(0.150)	(0.150)	(0.119)	(0.146)
Press freedom $_{t - 1}$	0.001	0.004	0.005*	0.001	0.008*	−0.002
	(0.002)	(0.002)	(0.002)	(0.002)	(0.003)	(0.003)
Respondent gender $_{t}$	−0.004	−0.004	−0.003	0.034	0.034	0.033
	(0.012)	(0.012)	(0.012)	(0.030)	(0.027)	(0.032)
Respondent age $_{t}$	−0.003	−0.004	0.003	0.002	0.002	0.002
	(0.006)	(0.006)	(0.007)	(0.001)	(0.001)	(0.002)
Respondent education $_{t}$	0.006	0.004	0.006	−0.003	−0.003	−0.004
	(0.009)	(0.010)	(0.010)	(0.005)	(0.005)	(0.005)
Respondent: Will of the people $_{t}$	0.211***	0.209***	0.212***
	(0.023)	(0.023)	(0.023)
Respondent: Political system approval $_{t}$				0.093***	0.092***	0.093***
				(0.010)	(0.010)	(0.010)
Constant	1.141*	0.809	1.008*	−0.514	1.887***	−0.454
	(0.454)	(0.440)	(0.477)	(0.372)	(0.574)	(0.372)
N	14,584	14,584	14,584	22,280	22,280	22,280

Maximum likelihood estimates using MLwiN 2.00. Standard errors are shown in parentheses. ***p $<$ 0.001, **p $<$ 0.01, *p $<$ 0.05.

dependent variable makes the use of experimental methods a particularly valuable tool with which to supplement the results of the observational analysis.

I design an experiment that constitutes a more direct test of Hypotheses 1 and 2 (particularly Hypothesis 2, since the experimental treatment is modeled after criticism from Amnesty International), and that examines whether exposure to external information affects individual human rights perceptions. I should note that the observational analysis has rather strong external validity (due to the large number of countries included within the GIMS and WVS datasets and the inclusion of multiple naming and shaming variables), but weaker internal validity, since it does not directly test whether individual perceptions are affected by naming and shaming, only whether the citizens of shamed countries have more negative perceptions of their country’s human rights conditions than the citizens of non-shamed countries. Conversely, the experimental analysis has strong internal validity (given that it directly tests the causal mechanism), but weaker external validity, since it has a smaller sample size drawn from only two countries. Together, the observational and experimental analyses are each strong where the other is weak, and they provide stronger support for Hypotheses 1 and 2 together than either does alone.

In October 2011, I ran a web-based experiment on human rights perceptions in the United States and India. The experiment consisted of an online survey available through Amazon’s Mechanical Turk (MTurk), where users can complete a variety of tasks from their home or work computer in exchange for financial compensation. In this case, users selected into and then completed a survey on the human rights conditions in their country, during which they were randomly assigned to read criticism of their country’s human rights conditions from an international organization. MTurk users tend to be more diverse (or, at the very least are no more specialized) than college students (Berinsky, Huber & Lenz, 2012). Moreover, MTurk panels have been used to successfully replicate several classic experiments that were originally conducted using student samples (Arceneaux, 2012).

The online survey was available only in those countries with enough MTurk users to ensure adequate participation (i.e. the United States and India). I collected approximately 414 responses: 209 from the USA, and 205 from India. Responses were collected between 18 October 2011 and 21 October 2011. Survey respondents provided basic information about themselves, including their age, gender, education level, annual income, and their perception of the overall economic conditions within their country. I also asked whether respondents believed that their government reflects the will of the people.¹²

After the initial questions, respondents were randomly assigned to one of two conditions, a treatment group and an untreated control group. Respondents in the US treatment group were asked to read the following paragraph:

Amnesty International accused the United States today of repeated violations of human rights. The group said it would publish a report about the violations, which included the mistreatment of racial minorities, execution of juvenile offenders, instances of police brutality, torture, and ill-treatment in jails. In all these areas, Amnesty said, it found violations of human-rights standards, including those under the International Covenant on Civil and Political Rights, which the United States ratified in 1992.

Respondents in the Indian treatment group were asked to read the following paragraph:

The human rights group Amnesty International has documented a pattern of repression by Indian authorities including systematic disappearances, widespread torture, rape and extrajudicial executions. Amnesty says the pattern of disappearances is just one of several forms of gross human rights violations committed by Indian forces. Amnesty’s membership in Asia and the Pacific is planning a campaign to focus on human rights violations in India. Indian authorities will be urged to ensure that perpetrators of human rights violations are brought to justice.

After exposure (or no exposure) to the treatment, survey respondents were asked to rate the quality of human rights in their country.¹³ Respondents were also asked to rate the quality of specific human rights included within the Universal Declaration of Human Rights (UDHR), including freedom from torture, freedom of speech, and related topics.

To maximize the statistical power of the experimental results, I included an instructional manipulation check (IMC) at the end of the questionnaire in order to identify those respondents who were more likely to have actually read the treatment (Oppenheimer, Meyvis & Davidenko, 2009). Survey respondents who did not give the correct answer to the IMC were dropped from the dataset (as recommend by Oppenheimer, Meyvis & Davidenko, 2009).¹⁴

Because the experimental analysis incorporates data at the individual level only, there are no macro-level variables and therefore no need for a multilevel analysis. I estimate a fixed-effects OLS regression using the survey responses from both countries. The results are shown in Table IV. The dependent variable in the first model is survey respondents’ perceptions of the general human rights conditions in their country on a scale from 1 to 7. The dependent variable in the second model is survey respondents’ perceptions of UDHR rights in their country on a scale from 1 to 7.¹⁵ The dependent variable in the third model is survey respondents’ perceptions of torture in their country on a scale from 1 to 7 (where 1 = the respondent strongly disagrees that their government does not torture people and 7 = the respondent strongly agrees that the government does not torture people). Because the USA and India have different economic conditions, I include an interaction term for perceptions of the local economy and whether or not the survey respondent lives in India.

The regression coefficient for treatment exposure is negatively signed for all three dependent variables,

Table IV.

Experimental analysis of human rights perceptions

Dependent variable	General rights	UDHR rights	Torture
Treatment	−0.1540	−0.0316	−0.2749
	(0.1840)	(0.1249)	(0.2062)
Respondent age	0.0064	0.0078	0.0043
	(0.0082)	(0.0051)	(0.0089)
Respondent gender	−0.2723	−0.2647*	−0.3938
	(0.1949)	(0.1308)	(0.2114)
Respondent education	0.4148	0.2965*	0.3977
	(0.2186)	(0.1405)	(0.2300)
Respondent: Perception of economy	−0.0725	0.0261	−0.0083
	(0.1150)	(0.0805)	(0.1396)
Respondent: Will of the people	0.4067***	0.3024***	0.3240***
	(0.0607)	(0.0406)	(0.0698)
India	−1.784***	−1.208**	−.1202
	(0.5686)	(0.4433)	(0.7263)
Perception of economy * India	0.2995*	0.1972	0.2131
	(0.1578)	(0.1179)	(0.1923)
Constant	2.402*	2.790	1.801
	(0.9783)	(0.6467)	(1.010)

Observations: 291. Robust standard errors are shown in parentheses. Two-tailed test: ***p $<$ 0.001, **p $<$ 0.01, *p $<$ 0.05.

including respondents’ perceptions of the general human rights conditions within their country (Table IV, Model 1), the conditions of specific rights mentioned within the UDHR (Model 2), and the level of torture committed by their government (Model 3), suggesting that survey respondents do react negatively to the naming and shaming of their country. The regression coefficients, however, are all statistically insignificant. Nevertheless, these results provide suggestive evidence for the theoretical argument that the observational analyses cannot because they were produced using data on individuals who were randomly selected to read about the hypothetical naming and shaming of their country.¹⁶ I should note that the treatment was modeled after actual press releases from Amnesty International, which is the type of international nongovernmental organization captured by the variable HRO Shaming in the observational analysis. Recall that I found strong results for the variables UNCHR Resolution and HRO Presence, and weaker results for the variable HRO Shaming. Future research could improve upon the experiment in this article and perhaps generate stronger results by modeling the experimental treatment upon criticism from the United Nations Commission on Human Rights (UNCHR), as well as by increasing the length of the treatment, the number of human rights violations referred to in the treatment, and the number of questions that survey respondents are asked about the conditions in their country. The formation of public opinion regarding human rights can be a complex, drawn-out process. Individuals must first observe human rights violations (or, alternatively, the naming and shaming of their government for abuses) and then care enough about human rights to form or change their opinions about the conditions in their country.

Regarding the external validity of the experimental results, although MTurk users tend to be more diverse than college students (Berinsky, Huber & Lenz, 2012), they also tend to be more highly educated than the general population. This does not necessarily mean that, on average, they will be more receptive to outside information. However, recent research by Hayes & Guardino (2011) on the influence of foreign voices appearing on US mass media suggests that highly educated individuals are more strongly influenced by foreign voices when they are also ideologically predisposed to the opinion being expressed by those voices. In other words, there is a triple interaction between an individual’s education, political ideology, and exposure to foreign voices on mass media. I did not design the experiment to examine those sorts of interactions. However, it is worth noting that, given the above average education level among MTurk users, it is possible that some survey respondents are more prone to accepting external sources of information than others because of their political ideology.

Because the United States and India were the only countries with enough MTurk users to ensure adequate participation in the online survey, I cannot replicate the experiment in other, less democratic countries, to see whether the results are generalizable. But the fact that the United States and India do not experience as many human rights violations as other less developed countries does not extirpate the results. Whether an individual lives in a country with a historically low or high number of human rights violations, they should still update their beliefs whenever they receive new information regarding the human rights conditions within their country (whether they experience violations themselves, or hear about the naming and shaming of their country, etc.).

Conclusion

Although naming and shaming is meant to coerce governments into stopping human rights violations, its track record has been mixed (Hafner-Burton, 2008; Krain, 2012). Governments often decrease the number of violations that they commit afterwards in order to decrease the pressure upon their government, for a myriad of reasons both economic and political. In those cases, naming and shaming has accomplished its stated goal. But shaming can also be followed by an overall increase in violations (Hafner-Burton, 2008). This is not surprising if the shamed government is worried that that shame has somehow hurt its ability to remain in power. Naming and shaming must have had some effect upon the domestic population that made the government feel as though it must further repress its own people. The empirical analyses in this article suggest that naming and shaming might indirectly affect government behavior by influencing public opinion.

Hafner-Burton (2008: 693) claims that naming and shaming ‘could have the accidental side effect of providing incentives for groups to orchestrate acts of violence large enough to attract the spotlight’ and that ‘governments sometimes react to these security challenges by repressing human rights even further’. In other words, citizens become aware that naming and shaming exists, and that they are able to cause their government to receive additional shame by upping their attacks on the government, causing it to commit additional violations. But there are multiple mechanisms through which naming and shaming could increase rebellion. Gurr (1970: 13) defined relative deprivation as ‘a perceived discrepancy between men’s value expectations and their value capabilities’ – between ‘the goods and conditions of life to which people believe they are rightfully entitled’ and ‘the goods and conditions they think they are capable of attaining or maintaining’. Certainly ‘conditions of life’ include basic human rights, and naming and shaming sends a signal to the shamed country’s citizens that those rights exist and that they are not being respected. It could incite further government repression by increasing discontent with the government. Discontent, after all, ‘is the basic, instigating condition for participants in collective violence’ (Gurr, 1970: 13), and when a country is shamed, its citizens perceive their conditions more negatively, thus increasing the likelihood of rebellion and potentially increasing the government’s need to repress. Citizens will rebel because their perceptions have changed, not only of their ability to bring shame upon their government, but of the severity of their government’s crimes. Strategic models of rebellion and repression already assume that citizens are at least somewhat aware of the human rights violations committed by their government, and more generally that people are capable of updating their beliefs based on new information (Gurr, 1970; Lichbach, 1996; Moore, 2000).

In sum, human rights perceptions have been linked with political dissent (Anderson, Regan & Ostergard, 2002), and political dissent and naming and shaming have both been linked with government repression (Carey, 2010; Hafner-Burton, 2008; Moore, 1998; Pierskalla, 2010). But it is possible that all four variables belong to the same causal chain: that naming and shaming has a negative effect upon human rights perceptions, thus increasing the level of political dissent within the shamed country and increasing the government’s need to repress the general population. Scholars cannot know whether government repression results from a link between naming and shaming and human rights perceptions until we have direct evidence that such a link even exists. I believe that the empirical analyses in this article constitute strong enough evidence that future research should now investigate whether naming and shaming is indirectly linked with government behavior via shifts in public opinion and, perhaps, whether naming and shaming could be more effectively designed in order to influence public opinion.

Hafner-Burton (2008: 713) notes that shining a spotlight upon governments for human rights violations ‘is followed by complex politics of human rights abuse and enforcement’. Indeed, naming and shaming does not simply place international pressure upon the shamed government. It modifies the domestic pressure that that government faces by affecting human rights perceptions among the general population, further complicating the ‘complex politics’ of human rights enforcement.

Footnotes

Replication data

Statistical analyses were conducted using Stata 10.0 and MLwiN 2.0. The datasets, codebook, MLwiN worksheets, and do-files for the empirical analyses in this article can be found at http://www.prio.no/jpr/datasets and at the author’s website, .

Acknowledgements

I would like to thank three anonymous reviewers and Sabine Carey for their thoughtful comments on previous versions of this manuscript, as well as Will Moore, Megan Shannon, Sean Ehrlich, Jennifer Jerit, Mark Souva, Scott Clifford, and participants at the SPSA Annual Meeting in 2012. Any remaining errors are my own.

Notes

References

Amnesty International (2012) Iran must quash human rights lawyer’s conviction ahead of prison term (http://www.amnesty.org/en/news/iran-must-quash-human-rights-lawyer-s-conviction-ahead-prison-term-2012-05-04, accessed 5 May 2012).

Anderson

Christopher

Regan

Patrick M

Ostergard

Robert L

(2002) Political repression and public perceptions of human rights. Political Research Quarterly 55(2): 439–456.

Arceneaux

Kevin

(2012) Cognitive biases and the strength of political arguments. American Journal of Political Science 56(2): 271–285.

Berinsky

Adam J

Huber

Gregory A

Lenz

Gabriel S

(2012) Evaluating online labor markets for experimental research: Amazon.com’s Mechanical Turk. Political Analysis 20(3): 351–368.

Bob

Clifford

(2005) The Marketing of Rebellion: Insurgents, Media, and International Activism. Cambridge: Cambridge University Press.

Bueno de Mesquita

Bruce

Smith

Alastair

Siverson

Randolph M

Morrow

James D

(2003) The Logic of Political Survival. Cambridge, MA: MIT Press.

Carey

Sabine C

(2010) The use of repression as a response to domestic dissent. Political Studies 58(1): 167–186.

Carlson

Matthew

Listhaug

Ola

(2007) Citizens’ perceptions of human rights practices: An analysis of 55 countries. Journal of Peace Research 44(4): 465–483.

Cavallaro

James L

Brewer

Stephanie Erin

(2008) Reevaluating regional human rights litigation in the twenty-first century: The case of the inter-American court. American Journal of International Law 102(4): 768–827.

10.

Chapman

Terrence L

(2009) Audience beliefs and international organization legitimacy. International Organization 63(4): 733–764.

11.

Cingranelli

David L

Richards

David L

(2010) The Cingranelli-Richards (CIRI) human rights dataset (http://www.humanrightsdata.org).

12.

Davis

David R

Murdie

Amanda

Steinmetz

Coty Garnett

(2012) ‘Makers and shapers’: Human rights INGOs and public opinion. Human Rights Quarterly 34(1): 199–224.

13.

Franklin

James C

(2008) Shame on you: The impact of human rights criticism on political repression in Latin America. International Studies Quarterly 52: 187–211.

14.

Freedom House (2000) Freedom of the Press Historical Data (http://www.freedomhouse.org/report-types/freedom-press).

15.

Gallup International (2000) Gallup International Millennium Survey (http://www.gallup-international.com).

16.

Gibney

Mark

Cornett

Linda

Wood

Reed

(2007) The political terror scale 1976 to 2006 (http://www.politicalterrorscale.org).

17.

Grieco

Joseph M

Gelpi

Christopher

Reifler

Jason

Feaver

Peter D

(2011) Let’s get a second opinion: International institutions and American public support for war. International Studies Quarterly 55(2): 563–583.

18.

Gurr

Robert

(1970) Why Men Rebel. Princeton, NJ: Princeton University Press.

19.

Hafner-Burton

Emilie M

(2008) Sticks and stones: Naming and shaming the human rights enforcement problem. International Organization 62(4): 689–716.

20.

Hayes

Danny

Guardino

Matt

(2011) The influence of foreign voices on U.S. public opinion. American Journal of Political Science 55(4): 830–850.

21.

Holsti

Ole R

(1996) Public Opinion and American Foreign Policy. Ann Arbor, MI: University of Michigan Press.

22.

Krain

Matthew

(2012) J’accuse! Does naming and shaming perpetrators reduce the severity of genocides or politicides? International Studies Quarterly 56(3): 574–589.

23.

Lebovic

James H

Voeten

Erik

(2006) The politics of shame: The condemnation of country human rights practices in the UNHCR. International Studies Quarterly 50(4): 861–888.

24.

Lebovic

James H

Voeten

Erik

(2009) The cost of shame: International organizations and foreign aid in the punishing of human rights violators. Journal of Peace Research 46(1): 79–97.

25.

Lichbach

Mark I

(1996) The Cooperator’s Dilemma. Ann Arbor, MI: University of Michigan Press.

26.

Livingston

Steven

Eachus

Todd

(1995) Humanitarian crises and U.S. foreign policy: Somalia and the CNN effect reconsidered. Political Communication 12(4): 413–429.

27.

Moore

Will H

(1998) Repression and dissent: Substitution, context, and timing. American Journal of Political Science 42(3): 851–873.

28.

Moore

Will H

(2000) The repression of dissent: A substitution model of government coercion. Journal of Conflict Resolution 44(1): 107–127.

29.

Murdie

Amanda

Davis

David

(2012) Shaming and blaming: Using events data to assess the impact of human rights INGOs. International Studies Quarterly 56(1): 1–16.

30.

Nyhan

Brendan

Reifler

Jason

(2010) When corrections fail: The persistence of political misperception. Political Behavior 32(2): 303–330.

31.

Oppenheimer

Daniel M

Meyvis

Tom

Davidenko

Nicholas

(2009) Instructional manipulation checks: Detecting satisficing to increase statistical power. Journal of Experimental Social Psychology 45(4): 867–872.

32.

Pierskalla

Jan Henryk

(2010) Protest, deterrence, and escalation: The strategic calculus of government repression. Journal of Conflict Resolution 54(1): 117–145.

33.

Poe

Steven C

Neal Tate

Keith

Linda Camp

(1999) Repression of the human right to personal integrity revisted: A global cross-national study covering the years 1976–1993. International Studies Quarterly 43(2): 291–313.

34.

Rasbash

Jon

Browne

William

Goldstein

Harvey

Yang

Min

Plewis

Ian

Healy

Michael

Woodhouse

Geoff

Draper

David

Langford

Ian

Lewis

Toby

(2000) A User’s Guide to MLwiN. London: Multilevel Models Project at the Institute of Education, University of London.

35.

Smith

Jackie

Wiest

Dawn

(2005) The uneven geography of global civil society: National and global influences on transnational association. Social Forces 84(2): 621–652.

36.

Steenbergen

Marco R

Jones

Bradford S

(2002) Modeling multilevel data structures. American Journal of Political Science 46(1): 218–237.

37.

World Values Survey Association (2009) World Values Survey 1981–2008 official aggregate v.20090902 (www.worldvaluessurvey.org).

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB