Accountability in Government Contracting Arrangements: Experimental Analysis of Blame Attribution Across Levels of Government

Abstract

While governments increasingly turn to third-party providers to deliver public services and government responsibilities are increasingly shifted from the federal to the state and local levels, both contracting and the division of powers under federalism blur lines of accountability. Because recent experiments on blame shifting find mixed results and citizens have different expectations of federal, state, and local government, we ask the following: How does blame attribution in third-party governance compare across levels of government? To address this question, we employ a timely survey experiment to examine who is responsible for a prisoner’s death in the case of interstate prisoner transport, which is one of the few services that is provided across all levels of government and by government contractors. The results show that contracting reduces the level of blame attributed to the government and that blame for contract failures varies by the level of government. Across levels of government, we find the local government sees the largest reduction in blame by contracting out. Findings have implications for accountability in contracting arrangements in public safety contexts.

Keywords

Contracting levels of government blame attribution experiment prison transport

The increase in contracting out is well known in service delivery, but the blurring of the lines between sectors gives citizens little clarity in who to hold accountable (e.g., Johnston & Romzek, 1999; Posner, 2002) and threatens democratic norms (Rosenbloom & Piotrowski, 2005). While politicians may use contracting out to deflect blame (Hood, 2002, 2007, 2011), recent experiments have found mixed results for such blame shifting (James et al., 2016; Marvel & Girth, 2016; Piatak et al., 2017). At the same time, the trend toward decentralization of the responsibility for policy implementation has moved from primarily the federal to state and local governments (e.g., T. Conlan, 2006; Dinan, 2008; Krane, 2002). Federal legislators have political incentives to devolve some public policies to state and local governments to claim credit for responding to constituent needs but avoid the blame of imposing unpopular taxes (T. J. Conlan, 2010; Cutler, 2004; Peterson, 1995). Like in contracting, the lines of accountability may be blurred if citizens hold a certain level of government—federal, state, or local—responsible for a given policy area (Arceneaux, 2005; Konisky, 2011; Leland et al., 2020; Schneider & Jacoby, 2003), even if they have little control.

Both contracting out and devolution in modern federal networks create distance between elected officials and citizens because accountability models are more complex. Politicians and public managers alike are more remembered for their failures than success. If delivering a service is unpopular and prone to failure, we may see shifts across levels of government and the use of contracting to avoid blame. Prison privatization, for example, tends to be driven by political pressures rather than efficiency goals (Nicholson-Crotty, 2004; Price & Riccucci, 2005). Unfortunately, prisons also experience a great deal of service delivery failures, making this the appropriate context to examine blame attribution in contracting arrangements in the federal system of the United States.

One aspect of accountability is the public’s expectations of government. As such, we build upon prior blame experiments in the context of service delivery in trash collection (Marvel & Girth, 2016; Piatak et al., 2017) and street maintenance (James et al., 2016) to examine who citizens blame for service delivery failure in the context of a federal system for a service prone to catastrophic delivery failure, prisoner transport. Federal, state, and local governments play unique roles, and the allocation of those functions may yield different contracting experiences (Romzek, 2000). Likewise, we view federalism to vary public support for different types of institutions. Because both contracting and our federal system blur lines of accountability, we ask the following: How does blame attribution in third-party governance vary across the federal, state, and local levels of government?

Using a timely survey experiment in the context of prisoner transport, we examine how contracting influences attribution of responsibility to the government or private provider across levels of government. When faced with a prisoner death in interstate transport, respondents attribute lower levels of blame in the contracting scenarios than in the in-house government provision scenarios. Looking across levels of government, the county government has significantly lower levels of blame than the state government. In addition, the local government has a significant interaction effect with contracting out in reducing levels of blame. These findings have implications for public management in decisions to contract out and public perceptions of government agencies across levels of government that likely reflect a variety of factors, including trust, brand favorability, and proximity.

Next, we examine the context of prisoner transport and theory behind blame attribution, a critical part of how individuals view the level of responsibility when there is service delivery failure. This is followed by our hypotheses, a description of the data and methods, a presentation of the results, and closes with a discussion of the key findings and implications.

Context: Prisoner Transport

Prisoner transport is frequently contracted out (see Kim & Price, 2014; Price & Riccucci, 2005) and prone to service delivery failure. As such, there have been several court cases to determine who is accountable, government or the private contractor, when things go wrong. For example, in West v. Atkins (1988), a prisoner injured himself during a volleyball game and was treated at a hospital but sued arguing he was treated with deliberate indifference. The question before the court was

whether a physician under contract with a state agency . . . could be sued under a federal statute permitting suits against those who act “under the color of the law” to deprive a person rights he or she might enjoy under the Constitution or laws of the United States. (Rohr, 2002, p. 3)

In a unanimous decision, the court ruled that the physician’s actions are under the color of state law. However, in Richardson v. McKnight (1997), McKnight, a prisoner, sued a privately run prison as he ended up in the hospital with injuries after being handcuffed too tightly during transport. The prison claimed qualified immunity, but in a 5-4 decision, the court ruled in favor of McKnight. Justice Breyer wrote the majority opinion stating, “our examination of history and purpose . . . reveals nothing special enough about the job or about its organizational structure that would warrant providing these private prison guards with a governmental immunity” (Richardson v. McKnight 1997). Justice Scalia dissented arguing that prior decisions have given immunity to private contractors serving a public function. Granted the disparate views of the court at the time, public opinion on who is to blame in certain situations may play a role in how future cases are determined.

There has been a long-standing pattern of deaths and abuses throughout privatized facilities. Private transport companies have been linked to 24 deaths, more than 50 crashes, 60 prisoner escapes, and 14 sexual assaults from 2000 to 2016 (Hager & Santo, 2016). In February 2019, members of the U.S. Congress demanded information about the abuse and deaths of prisoners transported through Prisoner Transportation Services LLC, a different company than the one the U.S. Department of Justice penalized in 2011 (Santo, 2019; Hager & Santo 2016), because of the number of criminal charges facing employees. While some private contractors have been held accountable for their actions or neglect, the public’s views about prisoners shape media attention as well.

Recently, there have been several newspaper accounts of instances where prisoners have been victims of cruel behavior during transport, including assault and traveling for an excessive number of days at a time in a cage, resulting in prisoner escapes or even death during such time (Hager & Santo, 2016; LaFlamme, 2019; Perez, 2017; Rehwald, 2018; Rhea, 2017; Santo, 2019; Hager & Santo, 2016; Satter, 2019; Webb, 2019). Prisoners have also been taken to the incorrect state, traveling more than 900 miles in the wrong direction (Alexander, 2019). Understanding the dynamics of citizen’s blame attribution in prisoner transport is a timely topic. Unfortunately, multiple news articles over the past few decades to draw from to ensure mundane realism.

Blame Attribution Theory

Accountability focuses on who the public expects to be responsible and for what, but who citizens hold responsible or blame when something goes wrong plays a key role in shaping those expectations. Therefore, we examine how contracting out influences blame attribution, as a part of the greater accountability to citizens across levels of government. First, we focus on the influence of the decision to contract out whether blame is attributed to the government or private provider. Do citizens always blame the government or does contracting out deflect blame to the private contractor? Next, we examine how blame varies across levels of government. Do citizens place more or less blame on federal, state, or local government agencies? Finally, we examine differences in blame shifting when contracting out across levels of government. Do certain levels of government shift more blame by contracting out than others?

Assignment of blame or responsibility is an issue of interest across disciplines. Social psychologists focus on blame attribution theory, which examines how a rational actor assigns blame in a specific context. According to blame attribution theory, people assess blame based on the level of personal control they perceive the actor to have as well as their perceptions of the outcome (Alicke, 2000; Alicke et al., 2008). In this sense, citizens will attribute blame if they find the outcome to be predictable and if they think there was intent. Alicke (2000) expands upon the role of intent, arguing that personal control can have both a positive and negative influence on blame attribution. Testing this theory with an experiment, Alicke et al. (2008) find physicians are blamed for negligence, but that blame is mitigated when precautions are taken. Intent or personal control plays a large role in blame attribution.

Similarly, political scientists assess blame or responsibility according to Heider’s (1958) triangle of responsibility model, where rational actors will attribute blame based on variance and personal causality. Like the degree of intent, the strength of the linkages between prescription-event (controllability), prescription-identity (the actor’s ties to the prescription), and identity-event (variance in the environment) influences who people hold responsible (Schlenker et al., 1994). Information on the strength of these linkages and the degree of personal control or intent can shape who people hold responsible.

However, not all blame is attributed neutrally based on the information provided. Politics and negativity bias infiltrate blame attribution decisions. Blame attribution is often a political game, where attribution is based on partisan rationalizations (Rudolph, 2003a, 2003b, 2006), competition (Mortensen, 2016), and blame avoidance (Weaver, 1986). In addition, political actors try to manipulate public opinion, such as partisan elites through issue framing (McGraw, 1991) and Congress by altering accountability chains (Arnold, 1990). Indeed, Jilke and Baekgaard (2020) find partisan bias plays a greater role when there are clear lines of responsibility. Negativity bias is found in both politics and administration. Political failures overshadow success (Lau, 1982, 1985; Weaver, 1987), and citizen dissatisfaction or poor performance outweighs citizen satisfaction or good performance (James, 2011; James & Moseley, 2014; James & Van Ryzin, 2017; Olsen, 2015). In addition, Sievert et al. (2020) find past blame increases future blame attributions. Therefore, negative outcomes have greater blame attribution than responsibility is given for positive outcomes. As such, we examine who the public holds responsible when something goes wrong in the case of contracting across levels of government.

While citizens may perceive a certain actor to be responsible for a given policy or public service, citizens often form opinions without complete information. Accountability can be difficult to determine, especially with the added complexity of lengthened accountability chains. One way citizens make decisions is based on cues, which are a shortcut for citizens to obtain information as a substitute for full or complex information (Downs, 1957; James, 2011; Lupia & McCubbins, 1998), such as understanding complicated lines of accountability. Therefore, we examine who citizens blame for service delivery failure when given cues about who is providing the service (the identity of Heider’s responsibility model) and in each context (the variance in the environment or link between prescription and event in Heilder’s responsibility model).

Hypotheses

Contracting and Blame

Current research now examines how contracting influences blame attribution. The government faces a public sector bias. This negative bias influences perceptions of performance, despite objective performance data (Marvel, 2015, 2016). In the case of contracting, those dissatisfied with services are more likely to think the government is the provider rather than a nonprofit provider (Van Slyke & Roch, 2004). Examining the politics side, James et al. (2016) find only delegation to a public unit inside the government reduces blame to local politicians, and Marvel and Girth (2016) find sector influences perceptions of mayoral control. Contrary to contracting helping to deflect blame (Hood, 2002, 2007, 2011), these studies find support for public sector bias where politicians are blamed even when services are contracted out.

Building upon the political experiments, Piatak et al. (2017) examine blame attribution to the local government in the case of contracting and service delivery failures. Using an experiment of illegal dumping by the trash services, they find citizens attribute blame appropriately to whichever sector is providing the service, except for in the case of budget shortfalls where blame is shifted from the contractor to the government (Piatak et al., 2017). Because blame attribution is based on intent and controllability of the event (Alicke, 2000; Schlenker et al., 1994), citizens should blame those who have direct control of providing the service. For example, research finds that voters are willing to punish government (in this case, an incumbent mayor) for a flood if they believe the city is responsible for flood preparations (Arceneaux & Stein, 2006). In the case of military contracting, Johnson et al. (2019) found the federal government is held accountable, regardless of contracting, perhaps because citizens see the government as responsible for the military rather than contractors. Although the political logic behind the decision to contract out is to deflect blame, perhaps citizens blame those responsible for providing the service when given cues about the provider. Therefore, we hypothesize the following:

Hypothesis 1: Governments that contract services will receive less blame than governments that provide the service directly.

Federalism and Blame Attribution

The United States federalist system allows citizens to shift responsibility in a policy area from one level of government to another. The legislative theory behind devolving public services to lower levels of government is to respond to the needs of constituents while passing the costs of providing the service to a lower level of government (e.g., Peterson, 1995). Policy studies show that the public has preferences for whether the federal, state, or local government level should be responsible for a given policy (Arceneaux, 2005; Schneider et al., 2010; Schneider & Jacoby, 2003; Shaw & Reinhart, 2001) based on political ideology (Schneider et al., 2010), evaluations of government performance (Arceneaux, 2005), support for devolution to the local government (Wolak, 2016), and trust (Leland et al., 2020; Maestas et al., 2020). Even in policy areas with shared responsibility, greater trust in the state government compared with the federal government leads to public preferences for state responsibility (Leland et al., 2020). Citizen trust in government has been on the decline since the 1970s, where citizens now consistently trust the local government level more than the federal and state government levels (Cole & Kincaid, 2000; Moore, 2013; Wolak & Palus, 2010). This may be due to proximity as citizens are more likely to interact with their local governments from visiting the library to going to school. Because citizens prefer different levels of government to carry out public policies, does citizen blame attribution vary across levels of government?

However, political actors also reframe public issues to shift blame across levels of government. In the case of Hurricane Katrina, Maestas et al. (2008) found the media primarily attributed blame to the state and that those closely following the national media mirrored this blame attribution, especially among Republicans. Relatedly, Gomez and Wilson (2008) found individuals with lower levels of political sophistication more likely to blame the federal government by default following Hurricane Katrina. In an experiment, Marvel (2014) found the influence of politics on blame shifting across levels of government in the case of the Boston Marathon bombing where blame statements about the Federal Bureau of Investigation (FBI) were particularly salient among Republications, whereas exculpatory statements were particularly salient among Democrats. Following the default blame to the federal government in the case of crises (Gomez & Wilson, 2008; Marvel, 2014) and borrowing the logic behind citizen trust shaping policy preferences (Leland et al., 2020), we hypothesize the following:

Hypothesis 2: Local government will receive the least blame and the federal government will receive the most blame for service delivery failure.

Methods and Data

To test these hypotheses, we employ a survey experiment with a fictional news article based upon real news stories that involve a prisoner being loaded into a van and due to neglect, dies along the way. The scenario varies both the contracting and in-house service provision, and involves a prisoner escaping and killing someone. While this may sound extreme, we found several recent news stories (see Hager & Santo, 2016; LaFlamme, 2019; Perez, 2017; Rhea, 2017; Santo, 2019; Satter, 2019; Webb, 2019), where these types of service delivery failures occur, which adds to the mundane realism. Thus, respondents are likely to view this type of scenario as something that they may realistically see and form an opinion about who is to blame. The vignette also varies whether the jurisdiction that was responsible for the transport was a County Sheriff’s Office, State Department of Corrections, or the Federal U.S. Marshal Service, which are the typical agencies responsible for handling prisoner transport at each level of government. Each respondent received one of these six randomly assigned conditions and then was asked about who was to blame in this scenario and at what level they blamed the government (Appendix A). The level of blame to the government was used as the outcome to test the hypotheses and is a five-point Likert-type scale (1 = no blame, 2 = a little blame, 3 = some blame, 4 = a lot of the blame, 5 = all the blame).

The survey experiment was part of a national study that ran January 28th and 29th of 2018. The experiment came first, followed by demographic and attitudinal questions and the final questions were memory check questions to see if the respondent was paying attention to the question. The respondents to this survey are members of the Amazon Mechanical Turk (MTURK) community. MTURK workers are paid for their time to complete tasks and experiments. Research shows that MTURK samples can be used in public administration experiments, but attention should be paid to whether the respondent spent enough time on the survey to answer meaningfully (Stritch et al., 2017). For completing the survey, the participants¹ earned two dollars. This amount is often considered good compensation by MTURK workers, which should cause them to pay attention to the scenarios. In addition, we included an attention check question at the end of the omnibus that asked about where the prisoner was being transported to address the concern about respondents speeding through the experiment and not paying attention. In our analysis, we only used those that correctly responded to the attention check question. Finally, the question that allows us to compute age asked them to input the four-digit year in which they were born and four entered two digit birthdays. Presumably, someone that had a birthday likely put 60, but to remove any doubt, we took out the four people whose computed age was greater than 150 years as these are likely errors from the way people responded to the question. The final sample size was 667 respondents from an original sample of 855 respondents that completed the survey.

The computer program randomly assigns respondents into one of the six conditions. The random assignment was checked by comparing the relevant demographic characteristics. Table 1 shows that the groups are approximately equivalent in terms of age distribution. While neither age nor gender were statistically significant at standard social science levels (α < .05), the distribution of female respondents is low in the federal/in-house condition and the test approaches statistical significance (p = .076). This suggests that a multivariate model that controls for the sex of the respondent may be appropriate in this situation.

Table 1.

Group Characteristics.

Condition	Average age (years)	Percent female (%)
County/in-house	38.24	51.4
State/in-house	37.42	51.6
Federal/in-house	36.83	35.0
County/contract	38.04	45.6
State/contract	36.43	45.7
Federal/contract	38.77	53.6
Mean	37.61	47.3
N	667	670
Test	F = 0.964 (significance = .441)	χ² = 9.962 (significance = .076)

Bivariate tests of the hypotheses are sufficient for most experiments; however, we take the additional step of testing the hypotheses in a multivariate model which should be the most conservative test of the data and is used in similar studies (James et al., 2016; Marvel & Girth, 2016; Piatak et al., 2017). We use multivariate regression to model the level of blame attribution. The variables in the model include age, race, and sex of the respondent along with attitudinal questions that may influence blame attribution. The first variable is political affiliation, and it is measured on a Likert-type scale with 1 being extremely liberal and 7 being extremely conservative. The sample leans slightly liberal with a value of 3.31. The descriptive statistics for the variables in the model can be found in Table 2.

Table 2.

Descriptive Statistics.

	Mean	SD	Minimum	Maximum
Political ideology	3.31	1.72	1	7
White	0.76	0.43	0	1
Black	0.09	0.28	0	1
Hispanic	0.06	0.23	0	1
Other	0.10	0.30	0	1
Age	37.61	10.85	19	74
Female	0.47	0.50	0	1
Valid N (listwise)	667

Results

The results of the analysis show that there is support for the hypotheses in the literature with some important differences. Particularly regarding the second hypothesis, the results offer a nuanced view of how citizens likely view blame for service delivery failure across the different levels of government. We first present the descriptive and bivariate statistics, and then we present the results of the multivariate models.

Descriptive Results

In the contract or in-house service conditions (Figure 1), the in-house service providing governments received a mean blame of 3.91 as compared with the governments that contracted the service that receive a mean blame of 3.41. This half-point difference is highly significant (F = 47.21, p < .001) and supports previous research that shows that governments generally avoid blame when contracting out.²

Figure 1.

Level of blame to government for service failure by contract condition.

When we examine the level of blame that respondents attribute to the different levels of government in the scenarios (Figure 2), we see significant differences in the level of blame attribution (F = 6.44, p < .01). As we expect by theory, the lowest level of blame goes to the county, the local government, with a mean blame of 3.54 and the highest level of blame is the state government with a mean blame of 3.84. Interestingly, the mean level of blame for the federal government is only 3.58, which is close to the level of blame attributed to the local government. This was not as expected by theory, but we discuss why this might be in the discussion section and turn to multivariate regression models to better interpret the effects.

Figure 2.

Level of blame to government for service failure by the level of government.

Regression Results

To test our hypotheses, we use ordinary least squares (OLS) regression³ to show the nuanced ways that citizens likely view service failures in the multiple levels of government (Table 3). Model 1 tests Hypotheses 1 and 2. First, the contract scenario reduces the blame to the government (p < .01). The model predicts that the contract scenario will have almost a half a point lower level of blame attribution (b = −0.481). This finding supports the first hypothesis that government agencies that contract out services will receive less blame than governments that provide the service directly.

Table 3.

OLS Regression Models of Level of Blame Attribution to Service Providing Organization.

	Model 1
Variables	B (SE)
Contract to private provider	−0.481** (0.073)
County	−0.295** (0.088)
Federal	−0.255** (0.089)
Age	0.002 (0.003)
Female	−0.064 (0.073)
African American	0.030 (0.132)
Hispanic or Latino/a	−0.166 (0.159)
Other race	0.244* (0.122)
Political ideology	−0.109** (0.021)
(Constant)	4.371** (0.163)
Adjusted r²	0.116

p < .05. **p < .01.

The second hypothesis that the level of government influences the level of blame attribution receives mixed support in Model 1. Specifically, the county receives nearly a .3 lower level of blame than the state (p < .01). Interestingly, the federal government also receives a lower level of blame relative to the state (p < .01), which is contrary to the literature on trust in government, that suggests lower levels of government will have lower levels of blame. In partial support of our second hypothesis, local governments have lower levels of culpability compared with the state government.

Interestingly, the African American and Hispanic/Latino(a) respondents did not attribute blame differently from the baseline White category. However, the Other race attributes higher levels of blame (p < .05). In terms of political ideology, the more conservative the person’s political affiliation, the lower the perceived blame attributed to the service providing organization (p < .01) in this scenario.

Discussion

In examining shifts in accountability for service delivery failure in the case of contracting a complex human service, we find respondents who received a vignette where prisoner transport was contracted out were more likely to shift blame from the government to the private company. Across levels of government, we find both the local government and the federal government levels receive lower levels of blame compared with the state government level. We build on prior work on blame attribution in two ways: first by examining blame in the context of contracting a complex, failure-prone service, and second by examining how the amount of blame attributed varies across levels of government.

Although the political logic behind contracting out may be to deflect blame from the government to the contractor (Hood, 2002, 2007, 2011), little support has been found for blame shifting in contracting scenarios. In experiments on blame attribution in contracting street maintenance and trash collection, scholars find citizens view politicians more responsible when services are delivered by the government (James et al., 2016; Marvel & Girth, 2016). Looking at bureaucratic accountability, Piatak et al. (2017) find citizens hold contractors or government responsible depending on who is delivering the service (trash collection), except in the case of budget shortfalls are present. In looking at the amount of blame given to either party, we find citizens place higher levels of blame on government than contractors, which is in line with political accountability research. This may also be due to public views about accountability, like military contracting (Johnson et al., 2019); citizens may have greater expectations for the government compared with private contractors.

In addition to blame attribution shifting in third-party governance, citizens also attribute blame across levels of government (Gomez & Wilson, 2008; Maestas et al., 2008; Marvel, 2014). Over the last 40 years, local government has enjoyed the highest level of trust when compared with state and federal governments with the federal government receiving the lowest level of citizen confidence (Cole & Kincaid, 2000). Trust shapes public preferences for government responsibility (Leland et al., 2020), that can in turn influence assessments of blame. In this experiment, the lowest level of blame goes to the county as predicted (which enjoys a higher level of citizen trust—that is, Cole & Kincaid, 2000), but this is also similar to the level of blame attributed to the federal government (which bears the lowest level of trust). The highest level of blame arises from the state government contract failure scenario. Perhaps the public holds the state level most responsible for prisoners. These results demonstrate the complexity of accountability of the U.S. federal system and offer a more nuanced view of how citizens likely view blame for the different levels of government when a service is contracted out.

Although the federal government may be the default level of government to blame for national crises, such as hurricanes and terrorist attacks (Gomez & Wilson, 2008; Marvel, 2014), the media can shape public perceptions of responsibility. Maestas et al. (2008) found the news blamed the state for Hurricane Katrina where individuals who followed the news mimicked blame attribution to the state. In addition to levels of trust, perhaps brand favorability influences may influence our results. To ensure mundane realism, we use the actual names of the government agencies responsible for prisoner transport across levels of government, including using the U.S. Federal Marshal Service. However, this may have been a trade-off. Teodoro and An (2018) find evidence of brand favorability for certain federal government agencies when using the specific agency name compared with the generic “federal government.” While the U.S. Federal Marshal Service was not one of the agencies experimentally tested, one could see how it may have brand favorability as it is featured in the TV series Justified and Gunsmoke as well as the novel turned movie True Grit, where the U.S. Marshal is played by beloved John Wayne in 1969 and Jeff Bridges in 2010. Tommy Lee Jones also starred in the 1990 film U.S. Marshals, which indicates that the U.S. Federal Marshals has received significant levels of positive publicity for decades. We would also like to note that a previous experiment, Marvel (2014), still found citizens put the most blame on the FBI compared with the Boston police department, and the FBI, like the U.S. Marshal Service, has also been portrayed favorably by TV. So, in the case of blaming one level of government over another, brand favorability may have taken a back seat to the circumstances of terrorism. The context may be more of a significant driver of brand favorability and could be related to dread risk when it comes to catastrophic events. Future research should examine the relationship between brand favorability of specific law enforcement agencies and citizens, especially because in law enforcement, government relies to an extent on citizen involvement for coproduction of service delivery, such as providing valuable information about crime.

Finally, we discover an interesting ideological link to blame attribution. In this case, the more conservative a respondent, the less blame attributed to the organization for service delivery failure. While partisan identity influences who people blame (Arceneaux & Stein, 2006; Bisgaard, 2015; Malhotra & Kuo, 2008; Marvel, 2014; Tilley & Hobolt, 2011), less is known about the degree of blame. More experiments are needed to explore this link, but we suspect it may have to do with who was harmed in the scenario, in this case, the prisoner. The experimental cue of a prisoner death could be viewed as a citizen in need of protection to some and an unworthy criminal to others. For example, Jilke and Tummers (2018) find deservingness cues influence which students teachers decide to help. If the scenario were switched to one where the prisoner escaped and caused harm to innocent people would conservatives be more likely to blame the organization? Future research should examine the link between perceptions of deservingness, political ideology, and contracting arrangements.

While this experiment contributes to the study of contracting across levels of government, there are some notable limitations. First, is that the two conditions presented in this experiment give the appearance that both may be manipulated for policy purposes. In reality, only the decision to contract out may be an intentional policy design choice. We view the second scenario of the level of government as simply an exogenous factor that may play into the choice of whether to contract the service or not. The fact that different levels of government experience different levels of blame should theoretically influence the decision on the policy decision of whether to contract the service, but this has not ever been examined to the best of our knowledge. In addition, the scenario of interstate transport may lead respondents to assume states are primarily responsible for prisoner transport, when this is intended to demonstrate a shared area of governance.

The second limitation of the study is that the sample of experiment respondents comes from MTURK. While MTURK studies have been used extensively for experiments in public administration (i.e., Jilke et al., 2016; Kaufmann & Tummers, 2017; Pedersen & Stritch, 2018), it is worth considering whether there are elements of the sample that may influence the response. The sample is diverse with 24% of the sample being non-White, age ranging from 19 to 74 years, and a 50/50 split of male to female. One consideration that we did not think about when we designed the survey, though, is that we do not have information on geography. This information was not considered important when the survey was designed, but it could be considered important in this type of analysis if the majority of the sample came from the DC metro or Virginia area, then this may influence some of the findings with regard to local, state, and federal government blame. While we do not believe this to be the case, unobserved factors can limit the generalizability beyond the sample. Additional concerns about the sample include whether the person is truly living in the United States. Using methods such as virtual private networks to circumvent detection of whether they truly live in the United States is also a concern, and future studies that study U.S. population should undertake methods to block these respondents with their attendant reduction in data quality (Kennedy et al. 2020). However, as a test of causal mechanisms, these respondents should be randomly distributed in the samples and are likely only a source of noise in the data. This has led researchers to conclude that MTURK samples are sufficient for experiments (Stritch et al., 2017); however, future iterations of this experiment and others like it may need to ask about the geography of the respondent to increase generalizability and attempt to block respondents that may not be from the United States.

Conclusion

Our findings have implications for public managers and elected officials across levels of government—federal, state, and local—in clarifying the lines of accountability and understanding citizen blame attribution in times of service delivery failure and contracting out. Service delivery failure of any type, especially in the extreme case of a contract failure in public safety, erodes both trust and the legitimacy of government which are critical in a representative democracy.

Because research has found public opinion mimics blame attribution in the media (Maestas et al. 2008), past blame influences future blame (Sievert et al., 2020), and citizens accurately blame those carrying out the service except when given cues to shift blame to the government (Piatak et al., 2017), public managers should ensure transparency in contracting and promote an informed citizenry. We find citizens place lower levels of blame for governments that have contracted the service, particularly at the local level. Public trust and risk perceptions play a role in public preferences for government responsibility (Leland et al., 2020; Maestas et al., 2020). Policymakers should consider both our federalist system and contracting arrangements in distributing responsibility for public programs.

Footnotes

Appendix A

Table A1.

Last Week, the Following Story Appeared in the Local Newspaper.

	Service provision
	In-house	Contracted
Level of government
County	An inmate died yesterday while being transported to California from a county in a neighboring state for extradition. Mr. Johnson was loaded in a van by two guards run by the County Sheriff’s Office to transport prisoners. He was crammed in a cage which held nine other men and women. There was no air conditioning in the vehicle and temperatures were around 105 degrees. After a twenty-hour drive, they arrived and found Johnson dead. The cause of death is unknown.	An inmate died yesterday while being transported to California from a county in a neighboring state for extradition. Mr. Johnson was loaded in a van by two guards run by Interstate Transport Inc., frequently hired by the County Sheriff’s Office to transport prisoners. He was crammed in a cage which held nine other men and women. There was no air conditioning in the vehicle and temperatures were around 105 degrees. After a twenty-hour drive, they arrived and found Johnson dead. The cause of death is unknown.
State	An inmate died yesterday while being transported to California from a neighboring state for extradition. Mr. Johnson was loaded in a van by two guards run by the State Department of Corrections to transport prisoners. He was crammed in a cage which held nine other men and women. There was no air conditioning in the vehicle and temperatures were around 105 degrees. After a twenty-hour drive, they arrived and found Johnson dead. The cause of death is unknown.	An inmate died yesterday while being transported to California from a neighboring state for extradition. Mr. Johnson was loaded in a van by two guards run by Interstate Transport Inc., frequently hired by the State of Department of Corrections to transport prisoners. He was crammed in a cage which held nine other men and women. There was no air conditioning in the vehicle and temperatures were around 105 degrees. After a twenty-hour drive, they arrived and found Johnson dead. The cause of death is unknown.
Federal	An inmate died yesterday while being transported to California from a neighboring state for extradition. Mr. Johnson was loaded in a van by two guards run by the Federal U.S. Marshal Service to transport prisoners. He was crammed in a cage which held nine other men and women. There was no air conditioning in the vehicle and temperatures were around 105 degrees. After a twenty-hour drive, they arrived and found Johnson dead. The cause of death is unknown.	An inmate died yesterday while being transported to California from a neighboring state for extradition. Mr. Johnson was loaded in a van by two guards run by Interstate Transport Inc., frequently hired by the Federal U.S. Marshal Service to transport prisoners. He was crammed in a cage which held nine other men and women. There was no air conditioning in the vehicle and temperatures were around 105 degrees. After a twenty-hour drive, they arrived and found Johnson dead. The cause of death is unknown.

Note. Please read the story, and then, we will ask you some questions about it.

Appendix B

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Suzanne Leland

Notes

Author Biographies

Suzanne Leland is a professor in the Department of Political Science and Public Administration at the University of North Carolina at Charlotte. She studies intergovernmental relations, urban policy and state and local service delivery.

Zachary Mohr is an associate professor in the Department of Political Science and Public Administration at the University of North Carolina at Charlotte. He studies cost accounting, budgeting, and organizational control in government. He is also the director of the POLS Lab at UNC Charlotte that facilitates survey, field, and lab experiments for public administration, political science, and public policy research.

Jaclyn Piatak is an associate professor in the Department of Political Science and Public Administration at the University of North Carolina at Charlotte. Her research focuses on how public services are delivered, such as through contracting, and who delivers them, examining sector differences and issues in human resource management.

References

Alexander

(2019). Mistake takes N.C. inmate transport vehicle on a detour—To the wrong state. Charlotte Observer. https://www.charlotteobserver.com/news/local/article234148247.html

Alicke

M. D.

(2000). Culpable control and the psychology of blame. Psychological Bulletin, 126(4), 556–574.

Alicke

M. D.

Buckingham

Zell

Davis

(2008). Culpable control and counterfactual reasoning in the psychology of blame. Personality and Social Psychology Bulletin, 34, 1371–1381.

Angrist

J. D.

Pischke

J. S.

(2008). Mostly harmless econometrics: An empiricist’s companion. Princeton University Press.

Arceneaux

(2005). Does federalism weaken democratic representation in the United States? Publius: The Journal of Federalism, 35(2), 297–311.

Arceneaux

Stein

R. M.

(2006). Who is held responsible when disaster strikes? The attribution of responsibility for a natural disaster in an urban election. Journal of Urban Affairs, 28(1), 43–53.

Arnold

R. D.

(1990). The logic of congressional action. Yale University Press.

Bisgaard

(2015). Bias will find a way: Economic perceptions, attributions of blame, and partisan-motivated reasoning during crisis. The Journal of Politics, 77(3), 849–860.

Cole

R. L.

Kincaid

(2000). Public opinion and American federalism: Perspectives on taxes, spending, and trust—An ACIR update. Publius: The Journal of Federalism, 30(1), 189–201.

10.

Conlan

(2006). From cooperative to opportunistic federalism: Reflections on the half-century anniversary of the commission on intergovernmental relations. Public Administration Review, 66(5), 663–676.

11.

Conlan

T. J.

(2010). From new federalism to devolution: Twenty-five years of intergovernmental reform. Brookings Institution Press.

12.

Cutler

(2004). Government responsibility and electoral accountability in federations. Publius: The Journal of Federalism, 34(2), 19–38.

13.

Dinan

(2008). The state of American federalism 2007–2008: Resurgent state influence in the national policy process and continued state policy innovation. Publius: The Journal of Federalism, 38(3), 381–415.

14.

Downs

(1957). An economic theory of democracy. Harper and Row.

15.

Gomez

B. T.

Wilson

J. M.

(2008). Political sophistication and attributions of blame in the wake of Hurricane Katrina. Publius: The Journal of Federalism, 38(4), 633–650.

16.

Hager

Santo

(2016). Inside the deadly world of private prisoner transport. The Marshall Project. https://www.themarshallproject.org/2016/07/06/inside-the-deadly-world-of-private-prisoner-transport

17.

Heider

(1958). The psychology of interpersonal relations. John Wiley.

18.

Hood

(2002). The risk game and the blame game. Government and Opposition, 37(1), 15–37.

19.

Hood

(2007). What happens when transparency meets blame-avoidance? Public Management Review, 9(2), 191–210.

20.

Hood

(2011). The blame game: Spin, bureaucracy, and self-preservation in government. Princeton University Press.

21.

James

(2011). “Managing citizens” expectations of public service performance: Evidence from observation and experimentation in local government. Public Administration, 89(4), 1419–1435.

22.

James

Jilke

Petersen

Van de Walle

(2016). Citizens’ blame of politicians for public service failure: Experimental evidence about blame reduction through delegation and contracting. Public Administration Review, 76(1), 83–93.

23.

James

Moseley

(2014). Does performance information about public services affect citizens’ perceptions, satisfaction, and voice behaviour? Field experiments with absolute and relative performance information. Public Administration, 92(2), 493–511.

24.

James

Van Ryzin

G. G.

(2017). Incredibly good performance: An experimental study of source and level effects on the credibility of government. The American Review of Public Administration, 47(1), 23–35.

25.

Jilke

Baekgaard

(2020). The political psychology of citizen satisfaction: Does functional responsibility matter? Journal of Public Administration Research and Theory, 30(1), 130–143.

26.

Jilke

Tummers

(2018). Which clients are deserving of help? A theoretical model and experimental test. Journal of Public Administration Research and Theory, 28(2), 226–238.

27.

Jilke

Van Ryzin

G. G.

Van de Walle

(2016). Responses to decline in marketized public services: An experimental evaluation of choice overload. Journal of Public Administration Research and Theory, 26(3), 421–432.

28.

Johnson

A. P.

Geva

Meier

K. J.

(2019). Can hierarchy dodge bullets? Examining blame attribution in military contracting. Journal of Conflict Resolution, 63(8), 1965–1985.

29.

Johnston

J. M.

Romzek

B. S.

(1999). Contracting and accountability in state Medicaid reform: Rhetoric, theories, and reality. Public Administration Review, 59(5), 383–399.

30.

Kaufmann

Tummers

(2017). The negative effect of red tape on procedural satisfaction. Public Management Review, 19(9), 1311–1327.

31.

Kennedy

Clifford

Burleigh

Jewell

Waggoner

(2020). The shape of and solutions to the MTurk quality crisis. Political Science Research and Methods, 8(4), 614–629.

32.

Kim

Price

B. E.

(2014). Revisiting prison privatization: An examination of the magnitude of prison privatization. Administration & Society, 46(3), 255–275.

33.

Konisky

(2011). Public preferences for environmental policy responsibility. Publius: The Journal of Federalism, 41(1), 76–100.

34.

Krane

D. A.

(2002). The state of American federalism, 2001–2002: Resilience in response to crisis. Publius: The Journal of Federalism, 32(4), 1–28.

35.

LaFlamme

(2019, August 8). Former Sabattus woman settles lawsuit in prisoner transport case. Sun-Journal. https://www.sunjournal.com/2019/08/08/former-sabattus-woman-settles-lawsuit-in-prisoner-transport-case/

36.

Lau

R. R.

(1982). Negativity in political perception. Political Behavior, 4(4), 353–377.

37.

Lau

R. R.

(1985). Two explanations for negativity effects in political behavior. American Journal of Political Science, 29, 119–138.

38.

Leland

Chattopadhyay

Maestas

Piatak

(2020). Policy venue preference and relative trust in government in federal systems. Governance. https://doi.org/10.1111/gove.12501

39.

Lupia

McCubbins

M. D.

(1998). The democratic dilemma: Can citizens learn what they need to know? Cambridge University Press.

40.

Maestas

C. D.

Atkeson

L. R.

Croom

Bryant

L. A.

(2008). Shifting the blame: Federalism, media, and public assignment of blame following Hurricane Katrina. Publius: The Journal of Federalism, 38(4), 609–632.

41.

Maestas

Chattopadhyay

Leland

Piatak

(2020). Fearing food: The influence of risk perceptions on public preferences for uniform and centralized risk regulation. Policy Studies Journal, 48(2), 447–468.

42.

Malhotra

Kuo

A. G.

(2008). Attributing blame: The public’s response to Hurricane Katrina. The Journal of Politics, 70(1), 120–135.

43.

Marvel

J. D.

(2014). The Boston marathon bombings: Who’s to blame and why it matters for public administration. Public Administration Review, 74(6), 713–725.

44.

Marvel

J. D.

(2015). Public opinion and public sector performance: Are individuals’ beliefs about performance evidence-based or the product of anti–public sector bias? International Public Management Journal, 18(2), 209–227.

45.

Marvel

J. D.

(2016). Unconscious bias in citizens’ evaluations of public sector performance. Journal of Public Administration Research and Theory, 26(1), 143–158.

46.

Marvel

J. D.

Girth

A. M.

(2016). Citizen attributions of blame in third-party governance. Public Administration Review, 76(1), 96–108.

47.

McGraw

K. M.

(1991). Managing blame: An experimental test of the effects of political accounts. American Political Science Review, 85(4), 1133–1157.

48.

Moore

(2013). State and local governments trusted more than the feds. YouGov.com. https://today.yougov.com/topics/politics/articles-reports/2013/12/02/state-local-governments-trusted

49.

Mortensen

P. B.

(2016). Agencification and blame shifting: Evaluating a neglected side of public sector reforms. Public Administration, 94(3), 630–646.

50.

Nicholson-Crotty

(2004). The politics and administration of privatization: Contracting out for corrections management in the United States. Policy Studies Journal, 32(1), 41–57.

51.

Olsen

A. L.

(2015). Citizen (dis) satisfaction: An experimental equivalence framing study. Public Administration Review, 75(3), 469–478.

52.

Pedersen

M. J.

Stritch

J. M.

(2018). Internal management and perceived managerial trustworthiness: Evidence from a survey experiment. The American Review of Public Administration, 48(1), 67–81.

53.

Perez

(2017, July 20). Prisoner transport officer arrested in alleged rape. Albuquerque Journal, C2.

54.

Peterson

P. E.

(1995). The price of federalism. Brookings Institution Press.

55.

Piatak

Mohr

Leland

(2017). Bureaucratic accountability in third-party governance: Experimental evidence of blame attribution during times of budgetary crisis. Public Administration, 95(4), 976–989.

56.

Posner

P. L.

(2002). Accountability challenges of third party governance. In Salamon

L. M.

(Ed.), The tools of government: A Guide to the new governance (pp. 523–551). Oxford University Press.

57.

Price

B. E.

Riccucci

N. M.

(2005). Exploring the determinants of decisions to privatize state prisons. The American Review of Public Administration, 35(3), 223–235.

58.

Rehwald

(2018, December 12). Nonprofit organization shines light on “deadly world” of inmate transport Industry. Springfield News-Leader. https://www.news-leader.com/story/news/local/ozarks/2018/12/12/inmate-transport-industry-deaths-escapes-inmate-abuse-marshall-project/2275173002/

59.

Rhea

J. P.

(2017). Highway to hell: The privatized prison transportation industry and the long road to reform. West Virginia Law Review, 120(1), 203–252.

60.

Richardson v. McKnight. (1997) 117 S. Ct. 2100.

61.

Rohr

J. A.

(2002). The ethical aftermath of privatization and contracting out: A constitutional analysis. Public Integrity, 4(1), 1–12.

62.

Romzek

(2000). Symposium on accountability in public administration: Reconciling democracy, efficiency and ethics-dynamics of public sector accountability in an era of reform. International Review of Administrative Sciences, 66(1), 21–44.

63.

Rosenbloom

D. H.

Piotrowski

S. J.

(2005). Outsourcing the constitution and administrative law norms. The American Review of Public Administration, 35(2), 103–121.

64.

Rudolph

T. J.

(2003a). Institutional context and the assignment of political responsibility. Journal of Politics, 65(1), 190–215.

65.

Rudolph

T. J.

(2003b). Who’s responsible for the economy? The formation and consequences of responsibility attributions. American Journal of Political Science, 47(4), 698–713.

66.

Rudolph

T. J.

(2006). Triangulating political responsibility: The motivated formation of responsibility judgments. Political Psychology, 27(1), 99–122.

67.

Santo

(2019). Members of congress seek answers from prisoner transport company. https://www.themarshallproject.org/2019/02/28/congress-seeks-answers-from-prisoner-transport-company

68.

Satter

(2019, January 13). Prisoner transport officer faces new sex-assault charge. Northwest Arkansas Democrat, 19.

69.

Schlenker

B. R.

Britt

T. W.

Pennington

Murphy

Doherty

(1994). The triangle model of responsibility. Psychological Review, 101(4), 632.

70.

Schneider

S. K.

Jacoby

W. G.

(2003). Public attitudes toward the policy responsibilities of the national and state governments: Evidence from South Carolina. State Politics & Policy Quarterly, 3(3), 246–269.

71.

Schneider

S. K.

Jacoby

W. G.

Lewis

D. C.

(2010). Public opinion toward intergovernmental policy responsibilities. Publius: The Journal of Federalism, 41(1), 1–30.

72.

Shaw

G. M.

Reinhart

S. L.

(2001). Trends: Devolution and confidence in government. The Public Opinion Quarterly, 65(3), 369–388.

73.

Sievert

Vogel

Reinders

Ahmed

(2020). The power of conformity in citizens’ blame: Evidence from a survey experiment. Public Performance & Management Review, 43(1), 53–80.

74.

Stritch

J. M.

Pedersen

M. J.

Taggart

(2017). The opportunities and limitations of using Mechanical Turk (Mturk) in public administration and management scholarship. International Public Management Journal, 20(3), 489–511.

75.

Teodoro

M. P.

S. H.

(2018). Citizen-based brand equity: A model and experimental evaluation. Journal of Public Administration Research and Theory, 28(3), 321–338.

76.

Tilley

Hobolt

S. B.

(2011). Is the government to blame? An experimental test of how partisanship shapes perceptions of performance and responsibility. The Journal of Politics, 73(2), 316–330.

77.

Van Slyke

D. M.

Roch

C. H

. (2004). What do they know, and whom do they hold accountable? Citizens in the government–nonprofit contracting relationship. Journal of Public Administration Research and Theory, 14(2), 191–209.

78.

Weaver

R. K.

(1986). The politics of blame avoidance. Journal of Public Policy, 6(4), 371–398.

79.

Weaver

R. K.

(1987). The politics of blame. The Brookings Review, 5(2), 243–247.

80.

Webb

(2019, November 8). Prisoner transport van crashes in southern Medina County, injuring 9 inmates and 3 others. Ashland Times-Gazette. https://www.times-gazette.com/article/20191108/NEWS/311089891

81.

West v. Atkins. (1988) 108 S. Ct. 2250.

82.

Winship

Mare

R. D.

(1984). Regression models with ordinal variables. American Sociological Review, 49(4), 512–525.

83.

Wolak

(2016). Core values and partisan thinking about devolution. Publius: The Journal of Federalism, 46(4), 463–485.

84.

Wolak

Palus

C. K.

(2010). The dynamics of public confidence in US state and local government. State Politics & Policy Quarterly, 10(4), 421–445.