Driving Public Support: Support for a Law is Higher When the Law is Named After a Victim

ABSTRACT

Despite the potential symbolic, political, and practical importance of naming a law after a victim, it is unclear whether this practice influences public opinion about the law itself. I conducted a randomized vignette survey experiment on 1,000 American adults to determine if support for a proposed distracted driving law, and the punishment it authorized, was influenced by whether it was named after a victim, as well as the victim’s race, gender, and age. I found that naming a law after a victim increased support for the law and the punishment authorized overall, but this effect was not consistent across all named laws, and instead was driven by specific types of named victims. In particular, results suggest the image of the “ideal victim” may have shifted or expanded to place greater emphasis on African American women, and less emphasis on White women.

Keywords:

• Race and crime
• named laws
• public opinion
• survey experiment
• distracted driving

Tonry (2004, p. 93) notes that “sometimes horrible events provide windows of opportunity for desirable policy changes.” A tragic crime with a sympathetic victim can serve as a “triggering event” that affords policymakers the opportunity to tap public emotions about the crime to propose a new law (McGarrell & Castellano, 1991). For example, after Casey Anthony was acquitted of murdering her daughter Caylee, policymakers around the county tapped the public outrage at the verdict by proposing Caylee’s Laws (Kallestad, 2011; Socia & Brown, 2016). These proposals may also tap broader public concerns about a perceived crime epidemic (see Cohen, 2011; Goode & Ben-Yehuda, 1994; Gottschalk, 2006; Sutherland, 1950a), as was seen by proposals for the AMBER Alert System, in response to a perceived rash of child abductions in the United States (see Fox, 2002; Zgoba, 2004a).

Policymakers can also tap this public outrage and promote policy change by naming a proposed law after a victim (Legal Information Institute, 2020). Since the early 1990s, America has seen an increase in these “apostrophe laws” (see Kulig & Cullen, 2017, p. Appendix A).¹ Naming a law after a crime victim is politically valuable, as it can help deter legislative debate by symbolically positioning supporters on the side of the victim, and painting opponents as “anti-victim,” “pro-offender,” and/or “soft on crime” (see Garland, 2001; Jones, 2012; Simon, 2007). For example, in the 1990s, legislators questioning Megan’s Law, which mandated public sex offender registries, would have positioned themselves as being “in league with predatory pedophiles and indifferent to the rape and murder of little girls and their families’ grief” (Wood, 2005, p. 13). Indeed, Fanarraga (2020) notes that over half of Federal apostrophe laws initially failed to pass as unnamed laws, and only found success after they were refiled as “named laws.”²

Beyond the political battles playing out in legislative sessions and press releases, taken for granted is the assumption that naming a law after a victim increases public support. Despite the potential symbolic, political, and practical importance of naming a law after a victim, the only evidence supporting this assumption is anecdotal and/or tangential. That is, it’s unclear whether public support for new legislation or harsher punishments is shifted by naming a law after a victim, nor is it clear how characteristics of the named victim influence this effect.

Public opinion matters because it can spur reforms in criminal justice policy (Pickett, 2019). It can influence whether a proposed law is passed, or instead dies in committee, and can help dictate the appropriate punishment attached to a law. That is, when a law’s punishment differs from what is perceived as appropriate by the public, it can undermine the law’s legitimacy and lead to backlash (Pickett, 2019; Robinson, 2013). Yet public punitivity is also flexible and can shift over time (Drakulich & Kirk, 2016; Pickett, 2019; Travis et al., 2014), and support for get-tough policies is “mushy” (Cullen et al., 2000). The recent Black Lives Matter (BLM) movement, and increased focus on Black/African American victims such as Breonna Taylor (see Drakulich et al., 2021; Duvall & Costello, 2021; Rubin, 2021), represent a potential dramatic shift in how racial minority victims are acknowledged by legislators and the public. Despite these recent developments, it remains unclear whether or how naming a law after a victim influences support for the law or the associated punishment.

In this study, I address this gap by conducting a survey experiment on a national sample of American adults. I first examine whether support for a proposed distracted driving law is influenced by, 1) whether the law is named after a victim, and/or 2) the named victim’s characteristics (age, race, gender). I then examine whether support for the level of punishment authorized by the law is influenced by victim naming and/or the named victim’s characteristics.

Prior Research

To date, no experimental research has explored how naming a law after a victim influences public support for either the law itself or the punishment it authorizes. Yet tangential research provides clues about what might be found. Below, I first consider such research in terms of how naming a law after a victim might influence support for passage and punishment. Then, I consider how specific victim characteristics may matter for this relationship.

How Naming a Law May Influence Public Opinion

Research on the “identifiable victim effect” (IVE; see Schelling, 1968), provides a theoretical explanation for how naming a law after a victim might influence public support. Specifically, IVE research suggests that individuals perceive and respond to situations differently when they are told about a specific, identifiable victim, compared to an unidentified victim. For example, providing the victim’s name when describing a crime can increase an individual’s willingness to help that victim (Jenni & Loewenstein, 1997; Lee & Feeley, 2016). Providing specific case details about a drunk driving crime (e.g. fatal injury, offender culpability) can increase respondents’ punitivity regarding drunk driving laws (Applegate et al., 1996), similar punitivity effects may come from providing identifying details about a specific victim.³

Based on the prior research, I expect that compared to an unnamed law, naming a law after a victim will increase support for passage and for punishment. Other tangential research can inform on the influence of specific victim characteristics (e.g. age, race, gender).

The Role of Victim Age

Many laws are named after child victims, particularly those targeting sex crimes (McAlinden, 2012, 2014). This makes sense, given the unique outrage and sadness that comes from the untimely death of a child (Hindmarch, 2009; Zelizer, 1985), and the support for harsh punishments for those responsible (Socia et al., 2021; Torre, 2007). Christie (1986, p. 19) notes that “very young people are particularly well suited as ideal victims.” Thus, among named laws, those named after child victims should have more support from the public, for both the law’s passage and the punishment it authorizes, compared to those named after adult victims.

The Role of Victim Race

Very little is known about how the race of the named victim influences public support for the legislation and/or associated punishment. This is perhaps a consequence of the disproportionate practice of naming laws after White victims (see Fanarraga, 2020; Kulig & Cullen, 2017; Leibman, 2009), affording few opportunities for real-life comparisons. Yet tangential research suggests what might be found. For instance, Tonry (2011) suggests that since the news and media typically portray “criminals” as Black, and “victims” as White, this influences the underlying racial preferences of the public (see also Slakoff, 2020; Slakoff & Fradella, 2019). Kulig and Cullen (2017, p. 982) further contend that naming laws after Black victims “simply does not come to mind” for many members of the public. As a result, among laws named after victims, a law named after a White victim can be expected to have more support from the public for its passage than an identical law named after a Black victim.

Regarding punishment, Hawkins (1987, p. 726) notes that “the lives and persons of Whites are more valued than those of Blacks in American society. Offenses against Whites are said to be more severely punished than those against Blacks regardless of the race of the offender.” Indeed, a long line of research shows that crimes against Black victims elicit less punishment than those against White victim (e.g. Baumer et al., 2000; LaFree, 1980; Williams et al., 2008; but see Applegate et al., 1996). Thus, a law named after a White victim is expected to have more support for punishment compared to an identical law named after a Black victim.

The Role of Victim Gender

Research suggests that victims who are men/boys are less valued than those who are women/girls, in terms of both media prioritization and public concern. Indeed, Greer (2007) notes that young men are at the bottom of the “hierarchy of victimization,” which correspond with Christie’s (1986) “ideal victim” image. Further, crimes against women/girls generally elicit greater desires for punishment than those against men/boys (Baumer et al., 2000; Williams et al., 2008), as well as longer sentences (Glaeser & Sacerdote, 2003). Therefore, I expect that among laws named after victims, those named after women/girls will have more support for both passage and punishment than those named after men/boys.

Potential Interaction Effects and the “Ideal Victim”

While the prior research suggests that differences in a named victim’s age, race, and gender might influence public opinion about the associated law, these characteristics do not operate independently. Indeed, there may be important interaction effects at play, especially as it relates to congruence with the “ideal victim” image of a young White woman or girl (Christie, 1986). Wood (2005, p. 5) contends that relying on the image of an ideal victim fosters a crime policy discourse that “authorizes punishment and silences critique.” As explained by Pickett et al. (2013, pp. 732–733), women and young children are “perceived publicly to be in the greatest need of protection from victimization, and [their] victimization generates the most visceral and retributive public reaction” (see also Dubber, 2006; Garland, 2001).

Existing research supports these contentions. Capital sentencing hearings in Delaware portrayed victims in “ideal ways” when they were White and female (Zaykowski et al., 2014), and this “ideal victim” portrayal may have driven sentencing bias (Kleinstuber et al., 2020). Other research found crimes against White women and girls elicited the most punishment (e.g. Williams et al., 2008; Williams & Holcomb, 2004). Media coverage of crime victims encourages this by focusing on missing White women/children, or otherwise reinforcing racist stereotypes (Slakoff, 2020; Slakoff & Fradella, 2019). Unsurprisingly, existing laws are disproportionately named after victims matching the “ideal victim” stereotype (Christie, 1986; Fanarraga, 2020; Kulig & Cullen, 2017; Leibman, 2009; Wood, 2005).⁴

This suggests that public opinion about a law might depend not only on whether a law is named after a victim, but also on characteristics of the named victim, either separately or in interaction. For example, among laws named after victims, I expect that a law named after a White girl will elicit the most support, and an identical law named after an African American man will elicit the least support. Kulig and Cullen (2017) note that these characteristics may be implicitly or explicitly conveyed by the victim’s name, and the resulting name of the law.

Summary of Expectations

In summary, there is little direct research examining whether naming the law after a victim influences public support for a law or the punishment it authorizes, or whether victim characteristics influence such support. Still, prior tangential research provides several expectations: First, I expect that a “named” law will have greater support than an unnamed law. Second, I expect that among named laws, support will be greater when the victim is a child (vs. an adult), is White (vs. African American), or is a woman/girl (vs. man/boy). Third, the “ideal victim” image suggests interaction effects may occur, such that support for a named law will be highest when it is named after a victim who is a White girl, and lowest when it is named after an African American man. Further, given the victim’s name can sometimes reflect a specific race and gender pairing, there may be particularly important interaction effects among race and gender, controlling for age. Finally, I expect that the same characteristics that predict support for the law will also predict support for the punishment authorized by the law.

Avoiding Penal Populism

While many of the existing laws are named after victims of heinous, violent crimes that elicit much media attention, less focus has been given to laws targeting more commonplace deviance. This is a problem, as laws targeting violent crimes elicit widespread public support, due in part to the intense emotional responses to the crimes itself and the “othering” of the perpetrators (see Goffman, 1986; Spencer, 2009; Sutherland, 1950b; Wright, 2003). This was seen in Caylee’s Law proposals (Socia & Brown, 2016), and is also a recurring theme for policies punishing of sex crimes (Harris & Socia, 2016; Socia et al., 2021). Further, certain crimes are associated with perceptions of specific kinds of victims (e.g. women as victims of domestic violence and rape), which may influence the effect of victim naming for laws targeting such crimes (see Terrance et al., 2011). Thus, in determining how naming a law after a victim influences public support, crimes that elicit knee-jerk public support (e.g. sex crimes), and/or are associated with stereotypical victim types (e.g. sex crimes, kidnapping), are not ideal choices, due to the existing intense emotions that surround such crimes.

This leads to a delicate balancing act when deciding on which law to explore in this study’s experiment. That is, the law must involve a deviant act that is: 1) common enough to be relevant to members of the public and allow for opinions to be formed, 2) not associated with a stereotypical victim type, 3) serious enough to warrant a criminal justice response, 4) yet not so serious as to have automatic widespread public support or intensely punitive reactions. As noted below, distracted driving laws meet these requirements and offer several other benefits.

Distracted Driving Laws

Distracted driving, such as when the driver of a vehicle uses a cellphone to talk, text, or surf the web, is both very common in the U.S. and comes with severe public safety consequences (Hill et al., 2015; Nelson et al., 2009; Overton et al., 2015). Accidents and deaths from distracted driving have substantially increased over the last two decades (National Center for Statistics and Analysis [NCSA], 2020, 2021; Wilson & Stimpson, 2010), and in 2018 alone, distracted driving led to 3,142 motor vehicle fatalities (NCSA, 2021). The public rightly perceives distracted driving as a serious public safety threat (Guarino, 2013; Li et al., 2014).

Yet distracted driving persists, despite a wide variety of laws targeting it throughout the country (Governors Highway Safety Association, 2021; Ibrahim et al., 2011), including some bearing the name of a victim.⁵ For instance, Michigan prohibited cellphone use by drivers under age 18 by enacting Kelsey’s Law, named after a 17-year-old girl who died in a 2010 cellphone-related driving accident (Aupperlee, 2012; C. Hall, 2013; Spelbring, 2020). New Jersey’s Kulesh, Kubert and Bolis’ Law (A-1074), passed in 2012, is named after three victims of distracted driving (Gloucester County Times, 2012). Yet naming a distracted driving law is not always successful. Over multiple legislative sessions since 2017, Oklahoma lawmakers have proposed The Bobbi White Act (Dossett, 2021) to prohibit the use of electronic devices while driving in a school zone. Named after a high school teacher killed in a distracted driving accident, it has not yet successfully passed the State Senate (Dossett, 2021; KFOR-TV & Querry, 2017). Thus, while distracted driving may not represent the “typical” violent crime targeted by named laws, it is a common form of deviance responsible for substantial and widespread public safety concerns, and laws targeting it have sometimes been named after victims.

As a result, distracted driving laws offer an important opportunity to isolate the influence of naming a law after a victim on public support for both the law itself and the punishment it authorizes. As noted earlier, this influence might otherwise be difficult to isolate when considering a law that already has widespread public support and heightened punitivity, as is seen with sex crime legislation (Klein & Cooper, 2019; Maguire & Singer, 2011; Sutherland, 1950b; Zgoba, 2004b). Thus, public support for a proposed distracted driving law would likely involve more thought and consideration, and be more susceptible to influence from naming effects, than for a law targeting heinous and/or exceptionally rare crimes.

For the purposes of this experiment, this means that respondents may be more malleable in their opinions about the need for distracted driving laws and the associated punishment, compared to more serious crimes that may see high levels of support and punitivity regardless of naming conventions. Further, unlike crimes targeted by Federal laws, there is much variation in existing state-level distracted driving laws, particularly in terms of restricting a driver’s use of a cellphone (Chase, 2014; Ibrahim et al., 2011; National Center for Statistics & Analysis, 2021). Thus, there should be similar variation in the opinions about both the need for additional distracted driving laws and what constitutes suitable punishment(s) for violations.

The Present Study

For this study, I conducted a randomized survey experiment embedded in a larger national survey of 1,000 United States adults. The experiment was designed to determine whether public support for a proposed distracted driving law, and the punishment it authorizes, were influenced by whether the law was named after a victim. It further explored how differences in the named victim’s characteristics (age, race, gender) influenced these results, separately or in tandem. Four research questions, and related hypotheses, are considered:

Table

Research question	Hypothesis
1. Does naming a proposed law after a victim increase support for the law?	1. Compared to unnamed laws, named laws will have more support.
2. Among laws named after a victim, do characteristics of the named victim influence the level of support for the law?	2a. Compared to a law named after a White victim, a law named after an African American victim will have less support.
	2b. Compared to a law named after a woman or girl, a law named after a man or boy will have less support.
	2c. Compared to a law named after a child victim, a law named after an adult victim will have less support.
3. Among laws named after a victim, are there interaction effects between the characteristics of the named victim?	3a. Support for the law will be highest when the named victim is a White girl.
	3b. Support for the law will be lowest when the victim is an African American man.
4. Do the same factors that influence support for the law also influence support for the level of punishment?	4. The same victim naming factors that influence support for the law also influence support for the level of punishment.

Research Design

The research design involved a randomized vignette within a larger national survey. Respondents considered a vignette about an unnamed state legislator proposing a new bill to implement dramatic new penalties for driving while using a cellphone without a hands-free device (i.e. distracted driving). Respondents were randomly assigned to consider one of nine possible vignette conditions: one condition involved an unnamed law, and eight other conditions involved a named law, with variations of the named victim’s age, race, and gender.

For the unnamed law condition, the vignette described the proposal as being a response to “recent high-profile accidents that were caused by distracted driving.” For the eight “named law” conditions, the vignette described the proposal as being a response to the death of a specific victim who was killed while walking down the sidewalk after being struck by a distracted driver who was checking their email on their cellphone.

After considering the randomly assigned vignette condition, respondents gauged their level of support for the law itself, and then gauged their level of support for various individual punishments for violating the law. Details about the vignette conditions are provided below.

Vignette Conditions

As noted earlier, respondents were randomly assigned to consider one of nine possible vignette conditions about the proposed distracted driving law. One vignette condition involved an unnamed law, while the other eight vignette conditions involved named laws. Each of the eight “named” conditions represented a unique combination of victim race (African American vs. White), gender (man/boy vs. woman/girl), and age (child vs. adult).⁶ The name of the victim, and the associated name of the law, were varied to reflect the four unique race-gender combinations (e.g. Latisha, Trisha, DeShawn, Shawn), and directly tap into underlying racial associations.⁷

The nine vignette conditions are provided in full in supplemental online Appendix A, and summarized below:

1. An unnamed law with no victim information

2. A law named after an African American girl (Latisha Davis/Latisha’s Law)

3. A law named after an African American woman (Latisha Davis/Latisha’s Law)

4. A law named after a White girl (Trisha Davis/Trisha’s Law)

5. A law named after a White woman (Trisha Davis/Trisha’s Law)

6. A law named after an African American boy (DeShawn Davis/DeShawn’s Law)

7. A law named after an African American man (DeShawn Davis/DeShawn’s Law)

8. A law named after a White boy (Shawn Davis/Shawn’s Law)

9. A law named after a White man (Shawn Davis/Shawn’s Law)

Support for Law and Punishment Measures

Strong Support for Law

Immediately after the presentation of the randomized vignette, respondents were asked: “How much would you support or oppose this proposed legislation?” Responses used a seven-point scale ranging from Strongly Oppose (0) to Strongly Support (6). The ordinal outcomes were almost evenly split between the “strongly support” response (49.7%) and all other responses (50.3%) (see supplemental online Appendix B), and the marginal distribution of these outcomes in ordinal logistic regression models (available upon request) indicated the “strongly support” response was dramatically different across the conditions compared to all other responses. As a result, the ordinal measure was recoded into a binary outcome of Strong support for law, which compared the “strongly support” response (1) to all other responses (0).

Support for Punishment

After responding to the overall support question, respondents gauged their support for each of eight different punishments that might be authorized by the law after an initial warning. The question was: “In considering the proposed legislation, assuming the first time someone is caught they are given a warning, how much would you support or oppose these specific penalties for a second distracted driving offense? Assume no accident or injury occurred in either offense.” The eight punishments were: 1) A fine of $100, 2) A fine of $1,000, 3) Points applied to driving record, 4) Requiring a 16-hour distracted driving course, 5) One week (40 hours) of community service, 6) Loss of a driver’s license for one month, 7) Putting the driver’s picture on a highway billboard labeled “Distracted Drivers”, and 8) Placing a large “DISTRACTED DRIVER” sticker on the rear windshield for one year.⁸ Support for each of the eight punishments was individually measured using a seven-point scale (Strongly Oppose [0] to Strongly Support [6]). These eight individual punishment measures were averaged into a single Punishment scale, where higher values meant more support for punishment. The Punishment measure had a Cronbach’s alpha of .83, indicating relatively high internal consistency.⁹

Survey Sample

The present study is part of a larger survey project that collected data from an online panel-based survey of 1,000 U.S. adults commissioned by the Center for Public Opinion at the University of Massachusetts Lowell (Dyck & Cluverius, 2020), and administered by YouGov from August 20–25, 2020. The CPO survey data are covered under the University of Massachusetts Lowell’s IRB exemption 14-069-DYC-EXM. As noted elsewhere (Socia et al., 2021), YouGov utilizes a two-stage sampling process, whereby surveys are first administered to a nonprobability “over-sample” from an online opt-in panel, which is then reduced to a representative final sample by algorithmically matching respondent characteristics to an established sampling frame (Rivers, 2006).

For the present survey, YouGov initially surveyed 1,154 respondents who were then matched down to a sample of 1,000 to produce the final dataset. The respondents were matched to a sampling frame on gender, age, race, and education. The frame was constructed by stratified sampling from the full 2017 American Community Survey (ACS) 1-year sample with selection within strata by weighted sampling with replacements (using the person weights on the public use file). See Ansolabehere and Schaffner (2014) for additional details on the sampling and survey methods of YouGov. YouGov surveys have been previously validated in election studies (e.g. Vavreck & Rivers, 2008), and used in U.S.-based polling conducted by media outlets (e.g. Cohn, 2014; YouGov, 2014), and social science opinion research (e.g. Burton et al., 2021; Norris & Mullinix, 2020; Rydberg et al., 2018; Socia & Harris, 2016).¹⁰

Missing Data

No cases were missing Strong support for law measure. However, two cases each had missing data on one of the eight individual punishment questions used to generate the overall Punishment measure. As the overall Punishment measure only considered non-missing punishment responses for a given case, an overall Punishment measure was able to be generated for all cases in the sample. As such, each model considers a full 1,000 cases.

Analytical Models

The analytical models estimated the effects of victim naming and characteristics on 1) Strong support for the law, and 2) support for the Punishment authorized by the legislation. All analyses were performed in Stata version 16.1. As noted earlier, the Strong support for law variable compared the “strongly support” response (1) to all other responses (0), and so these models were estimated using logistic regression. The Punishment measure had a more normal distribution (Skewness = −.36; Kurtosis = 2.93), and so these models were estimated using ordinary least squares (OLS) regression. All models used robust standard errors.

Testing for Interaction Effects

The various combinations of the named victim’s race, gender, and age characteristics might be thought of as potential interactions among named laws. However, the use of product terms can be misleading when exploring interactions in non-linear models (see, Mize, 2019). As such, I considered interaction effects based on the differences in marginal effects (first order differences) and the differences of differences in the marginal effects (second order differences) across the race, gender, and age combinations of the named victim.

Results

Descriptive statistics are provided in Table 1. The randomization procedure assigned an equal probability of exposure to each vignette condition, and a check of the randomization procedure indicated equal distribution of respondent characteristics (gender, age, race, education) across the vignette conditions (see supplemental online Appendix C). This suggests the randomization procedure produced substantively similar respondent groups, and any significant differences in the outcome measures are likely due to the variations in the vignette conditions. Strong support for law was binary, with 49.7% of respondents (N = 497) indicating strong support. Prior to standardizing, the Punishment measure was essentially continuous, ranging from 0 to 6, with a mean of 3.88 (SD 1.27). After standardizing, the measure had a mean of 0 and a standard deviation of .68.

Table 1. Descriptive statistics.

N = 1000 Variable	Mean	SD	%
Strong support for law			49.7%
Support for punishment	0.00	0.68
Vignette Conditions
Named law			88.8%
Child victim			43.5%
White victim			44.0%
Woman/Girl victim			43.9%
Name of law, victim age
Unnamed law			11.2%
Latisha’s law, child			10.1%
Latisha’s law, adult			12.1%
Trisha’s law, child			10.7%
Trisha’s law, adult			11.0%
DeShawn’s law, child			11.6%
DeShawn’s law, adult			11.0%
Shawn’s law, child			11.1%
Shawn’s law, adult			11.2%

Strong Support for Law and Support for Punishment by Condition

Table 2 explores strong support for the law and support for punishment based on the eight unique named law conditions, as compared to the unnamed law condition. Further, comparisons between the named law conditions are also provided for each outcome.

Table 2. Predicting strong support for law and support for punishment.

	Strong support for law				Support for punishment
	b	SE	OR	Contrasts	β	SE	Contrasts
Name of law, victim age^1
a. Latisha’s law, child	0.82**	0.28	2.26		0.15	0.10
b. Latisha’s law, adult	0.85**	0.27	2.35	d	0.25**	0.09
c. Trisha’s law, child	0.49	0.27	1.63		0.17	0.10
d. Trisha’s law, adult	−0.09	0.28	.92	b	−0.01	0.09
e. DeShawn’s law, child	0.61*	0.27	1.84		0.14	0.10
f. DeShawn’s law, adult	0.36	0.27	1.44		0.07	0.10
g. Shawn’s law, child	0.24	0.27	1.27		0.18*	0.09
h. Shawn’s law, adult	0.51	0.27	1.66		0.20*	0.09
Constant	−0.44*	0.19	.65		−0.13	0.07
N	1,000				1,000
	Pseudo R^2 = .02				R^2 = .02
	Wald χ^2(8) = 22.92**				F(8,991) = 1.91

Note: Models present robust standard errors. Strong support for law model presents logistic regression estimates, and support for punishment model presents ordinary least squares estimates. The contrasts column indicates that marginal estimates are significantly different at an alpha level of .05 after applying a Bonferroni correction for multiple comparisons.

¹Ref. Unnamed law.

* p < .05; ** p < .01; *** p < .001.

Strong Support for Law

As shown in Table 2, compared to an unnamed law, the odds of strong support were significantly higher for Latisha’s law (child or adult), and for child DeShawn’s law. None of the other named law conditions had strong support levels that were significantly different from the unnamed law condition. The Contrasts column indicates which named laws were statistically different from each other based on pairwise comparisons with a Bonferroni correction. Among the named laws, the only significant difference was between adult Latisha’s law and adult Trisha’s law.

The left side of Figure 1 graphically presents the marginal effects of the probability of strongly supporting the law for each named law condition, along with 95% confidence intervals, in comparison to an unnamed law (0 on the Y axis). As was indicated in Table 2, compared to an unnamed law, the probability of strongly supporting the law was significantly higher when the law was named after an African American woman, girl, or boy.

Figure 1. Conditional marginal effects with 95% CI for each outcome.

Note: Conditional marginal effects for each law are in comparison to an unnamed law (0 on the Y axis).

A follow up model predicting strong support based only on a dichotomous “Named Law” measure (with unnamed law as the comparison) is presented in supplemental online Appendix D. This model indicates that respondents presented with a named law vignette are 1.61 times more likely to strongly support the law compared to respondents presented with an unnamed law vignette (p < .05). Thus, across all iterations of the named victim, there is a significant “naming” effect on the likelihood of strong support.

Support for Punishment

As shown in the second model in Table 2, compared to an unnamed law, support for punishment was higher for adult Latisha’s law, as well as for both Shawn’s laws (child or adult). None of the other named law conditions had punishment levels that significantly differed from the unnamed law condition. Note that among the named laws, there were no significant pairwise differences after applying a Bonferroni correction.

The right side of Figure 1 graphically presents the marginal effects on support for punishment for each named law condition, with 95% confidence intervals, as compared to an unnamed law (Y axis = 0). As was indicated in Table 2, compared to an unnamed law, support for punishment was significantly higher for an African American woman, and a White man or boy.

A follow up model predicting support for punishment based only on a dichotomous “Named Law” measure (with unnamed law as the comparison) is presented in supplemental online Appendix D. This model indicates that respondents presented with a named law vignette led to a standardized punishment score that was .14 points higher than respondents presented with an unnamed law vignette (p < .05). Thus, across all iterations of the named victim, there is a significant “naming” effect on the level of punishment authorized.

Testing for Interaction Effects

Interaction effects for combinations of victim gender, race, and age are explored using predictive margins in Table 3 (Strong support for law) and Table 4 (Support for punishment). For each effect considered (gender, race, and age), the marginal estimates are presented in the first two columns for each unique pairing of the other two victim characteristics (i.e. the gender effects of man/boy vs. woman/girl for each combination of victim race and age). The third column presents first-order differences (e.g. the differences between the marginal estimates for man/boy and woman/girl within each race-age combination). The last column, Contrasts, indicates which of the 1^st order differences were significantly different from each other (i.e. second-order differences). The marginal estimates for different combinations of victim characteristics are visually presented in supplemental online Appendix E.

Table 3. Probability of strong support for law: Predictive margins and differences in effects by victim characteristics.

Effect of victim gender
	Man/Boy		Woman/Girl		Man/Boy vs. Woman/Girl
	Margin	SE	Margin	SE	1st differences	SE	Contrasts
a. AA Adult	0.48*	0.05	0.60*	0.04	0.12	0.07	b
b. White Adult	0.52*	0.05	0.37*	0.05	−0.15*	0.07	a
c. AA Child	0.54*	0.05	0.59*	0.05	0.05	0.07
d. White Child	0.45*	0.05	0.51*	0.05	0.06	0.07
Effect of victim race
	African American		White		AA vs. White
	Margin	SE	Margin	SE	1st differences	SE	Contrasts
a. Man	0.48*	0.05	0.52*	0.05	0.04	0.07	b
b. Woman	0.60*	0.04	0.37*	0.05	−0.23*	0.06	a
c. Boy	0.54*	0.05	0.45*	0.05	−0.09	0.07
d. Girl	0.59*	0.05	0.51*	0.05	−0.08	0.07
Effect of victim age
	Adult		Child		Adult vs. Child
	Margin	SE	Margin	SE	1st differences	SE	Contrasts
a. Deshawn	0.48*	0.05	0.54*	0.05	0.06	0.07
b. Shawn	0.52*	0.05	0.45*	0.05	−0.07	0.07
c. Latisha	0.60*	0.04	0.59*	0.05	−0.01	0.07
d. Trisha	0.37*	0.05	0.51*	0.05	0.14	0.07

Note: Standard errors calculated using delta method. Contrasts indicate significant 2^nd order differences. Differences include Bonferroni correction for multiple comparisons. African American abbreviated AA.

* p < .05.

Table 4. Support for punishment: Predictive margins and differences in effects by victim characteristics.

Effect of victim gender
	Man/Boy		Woman/Girl		Man/Boy vs. Woman/Girl
	Margin	SE	Margin	SE	1st differences	SE	Contrasts
a. AA Adult	−0.06	0.07	0.13*	0.06	0.18	0.09	b
b. White Adult	0.07	0.06	−0.13*	0.06	−0.20	0.09	a
c. AA Child	0.01	0.07	0.02	0.07	0.01	0.09
d. White Child	0.05	0.05	0.04	0.07	−0.01	0.09
Effect of victim race
	African American		White		AA vs. White
	Margin	SE	Margin	SE	1st differences	SE	Contrasts
a. Man	−0.06	0.07	0.07	0.06	0.13	0.09	b
b. Woman	0.13*	0.06	−0.13*	0.06	−0.26*	0.09	a
c. Boy	0.01	0.07	0.05	0.05	0.04	0.09
d. Girl	0.02	0.07	0.04	0.07	0.02	0.10
Effect of victim age
	Adult		Child		Adult vs. Child
	Margin	SE	Margin	SE	1st differences	SE	Contrasts
a. Deshawn	−0.06	0.07	0.01	0.07	0.07	0.09
b. Shawn	0.07	0.06	0.05	0.05	−0.02	0.08
c. Latisha	0.13*	0.06	0.02	0.07	−0.11	0.09
d. Trisha	−0.13*	0.06	0.04	0.07	0.18	0.09

Note: Standard errors calculated using delta method. Contrasts indicate significant 2^nd order differences. Differences include Bonferroni correction for multiple comparisons. African American abbreviated AA.

* p < .05.

Effect of Victim Gender

As shown in Table 3, the only significant (p < .05) difference in marginal estimates (1^st order differences) for the effect of victim gender was for White adults. This suggests a decreased probability of strongly supporting the law when the adult White victim changed from a man (Shawn) to a woman (Trisha). There was no significant “gender gap” in strongly supporting the law for victims who were either African American adults, African American children, or White children. However, as indicated in the Contrasts column, there was a significant 2^nd order difference (p < .05) between the gender gaps of African American adults (.12) and White adults (-.15). This suggests that the increased probability of strongly supporting the law when the adult African American victim changed from a man (Deshawn) to a woman (Latisha) was significantly different from the decreased probability when the adult White victim changed from a man (Shawn) to a woman (Trisha). This difference in the gender gap is also seen in support for punishment in Table 4. Specifically, the increased support for punishment resulting from changing adult Deshawn to adult Latisha (.18) is significantly different from the decreased support for punishment resulting from changing adult Shawn to adult Trisha (-.20).

Effect of Victim Race

There was a significant racial gap for both the probability of strongly supporting the law and support for punishment among victims who were women. Specifically, the probability of strongly supporting the law decreased by .23 (Table 3), and support for punishment decreased by .26 (Table 4), when the adult victim changed from Latisha to Trisha. None of the other racial gaps were significant among the remaining gender-age combinations. As indicated in the Contrast column of both tables, the racial gap is significantly larger for adult women than for adult men.

Effect of Victim Age

Unlike the results for gender and race gaps, there were no significant age gaps for any of the race-gender combinations for either the probability of strongly supporting the law (Table 3) or support for punishment (Table 4). Nor were any of the racial gaps significantly different from each other. In other words, among different race-gender combinations, there was no significant difference in either outcome measure between an adult victim and a child victim.

Overall, these results suggest that among the named laws, there were only a couple of significant interaction effects, specifically indicating a gender gap in strongly supporting the law among White adult victims, a racial gap for support for punishment, and a difference in both the gender and racial gaps for both outcomes. The significant gender gap means that the likelihood of strongly supporting the law was significantly higher when the victim was adult Shawn compared to adult Trisha. The significant racial gap means that support for punishment was significantly lower when the victim was adult Trisha compared to adult Latisha. The significant second-order differences suggest that the gender gaps were significantly different between adult African American victims and adult White victims, while the racial gaps were significantly different between victims who were men and victims who were women.

Discussion

In this study, I conducted an online survey experiment to consider how support for a proposed distracted driving law, and the punishment it authorized, varied based on whether the law was named after a victim, and the characteristics of the named victim. Compared to an unnamed law, a law named after a victim was overall more likely to have strong support, and when considering the individual victim types, this effect was driven by laws named after African American women, girls, and boys. Compared to an unnamed law, a law named after a victim was overall also more likely to elicit a higher level of authorized punishment, and this was driven by laws named after African American women, and White men and boys.

Results also suggested an important interaction effect of victim race, such that such that a law named after a White woman (Trisha) had a lower probability of strong support, and lower support for punishment, compared to a law named after an African American woman (Latisha). This racial gap was not present among laws named after men, nor among laws named after children of either gender. Further, when comparison the sizes of the gender gaps and the racial gaps, results indicate that for both outcomes, the gender gap is larger for victims who are African American adults compared to White adults, and the racial gap is larger for victims who are men compared to women.

Overall, while it does seem that naming a law after a victim can, at times, increase strong support for the law, and for the punishment it authorizes, this presents a conundrum: how does one reconcile the expectations of the “ideal victim” research with the findings of the present study? That is, the ideal victim stereotype suggests that the public perceives White women/girls as the most “valued” victims, and thus laws named after these victims should have the most support from the public for both passage and for authorized punishment. Yet in this study, support for the law was generally highest for laws named after African American women/children, and lowest for White women. Support for punishment was highest among laws named after African American women and White men/boys, and (again) lowest for White women. Laws named after other victim types fell between these extremes.

These results may mean that the image of the “ideal victim” does not consistently influence decisions about supporting laws named after such victims, nor the associated punishment authorized. Yet this seems unlikely, given how policymakers publicly discuss proposed laws in ways that highlight such victims (e.g. Socia & Brown, 2016). Instead, these results may suggest that the image of the ideal victim has shifted, even if temporarily, and may also be influenced by the type of outcome considered (e.g. support for a law vs. level of punishment).

The most obvious explanation of this shift would be from the current events that occurred while this survey was fielded in August 2020. Specifically, this period saw widespread protests across America that focused on the overall devaluation of Black/African American lives in society (see Chenoweth & Pressman, 2020; Putnam et al., 2020), and a related push for “Breonna’s Law,” both of which may have influenced views regarding African American victims (see Cecil, 2022; Tesler, 2020). That is, it may be that the ideal victim image may have expanded or shifted from that of a White woman, to place more emphasis on African American women. This would explain the significant racial gap in both outcomes for victims who were women. If so, then could represent one of Tonry’s (2004) “policy windows” for future legislative proposals, at least in terms of naming laws after African American victims.

Further research measuring a respondent’s specific perceptions of an ideal victim would be required to confirm these speculations, as this cannot be confirmed with this study’s cross-sectional data. As Pickett (2019) astutely notes, reliance on cross-sectional studies is a problem as they obviously do not speak to trends over time. Even assuming the image of the ideal victim has changed, it is unclear whether this will persist, or is just a dramatic but fickle and temporary shift in public opinion (Drakulich & Kirk, 2016; Pickett, 2019; Travis et al., 2014).

There were some interesting differences between which conditions influenced support for the law and support for its punishment. That is, despite similarities between the two models in terms of the direction of effects for the conditions (see Figure 1), laws named after African American children were more influential for support for the law, while laws named after White men/boys were more influential for support for the punishment level. Although the exact reason for these differences cannot be confirmed with these data, it may be that support for the law taps a utilitarian perspective of reducing future crimes, and thus more influenced by perceptions of an innocent and vulnerable victim, while support for the level of punishment represents a retributive perspective of retaliating against a wrong, and thus more influenced by the social status of the victim (Gerber, 2021). However, research on vehicular homicides suggests that desires for more punishment are also linked to perceptions of victim innocence (Glaeser & Sacerdote, 2003). Yet neither explanation clarifies why adult Latisha’s law had the highest support for both outcomes, nor why adult Trisha’s law had the lowest levels of support for both outcomes.

Still, these findings suggest at least partial support for an identifiable victim effect influencing support for the law and its punishment (Lee & Feeley, 2016; Schelling, 1968; Wiss et al., 2015). That is, named laws were generally more likely to have strong support and higher punishment levels than an unnamed law. This can explain how unnamed legislation that initially fails to pass can find success after being renamed after a victim (Fanarraga, 2020; see also Brown, 2012). Overall, these results suggest that support for a law and its punishment level can be increased by simply naming it after a particular victim, at least for certain types of victims.

Along with increased support, there are other arguments for using this practice. First, the symbolic gesture of naming a law after a victim may bring some small measure of peace to the victim’s family by memorializing the victim and/or providing “social good” from the tragic loss (Kulig & Cullen, 2019). For example, the mother of Breonna Taylor reportedly called Louisville’s “Breonna’s Law” a testament to her daughter’s agenda to save lives as an ER technician (Duvall & Costello, 2021). Second, naming a law after a victim may increase the public’s awareness of that legislation. That is, the image and name of a memorable victim may help capture the attention of the public, and thus increase potential compliance. For example, under New Hampshire’s “Jessica’s Law,” motorists are required to clear snow and ice off of their vehicles before they drive (City of Lebanon, 2021). It seems reasonable to assume that New Hampshire residents would be more familiar and aware of “Jessica’s Law” than they would of an unnamed law referenced only by the associated legal code, NH RSA 265:79-b.

However, there are also potential drawbacks to naming a law after a victim. That is, many ineffective, wasteful, and/or harmful policies have been proposed and passed in the name of a sympathetic victim (e.g. Caylee’s Law). The act of naming a law after a victim has been used as a political weapon to “beat the hell out of the opposition” and get a proposal passed (see Brown, 2011, p. 442). The political benefits of naming a law after a victim does not mean the law itself is beneficial to society. This is particularly relevant given the trend towards penal populism (Bottoms, 1995; Pratt, 2007), whereby policymakers propose crime legislation in response to public opinion and outrage, rather than on policy effectiveness and justifiable need. It may be that passing laws in the name of sympathetic victims increases the overall punitivity of the justice system, perhaps with disproportionate legal consequences for minority communities.

These results may help inform on other efforts to regulate commonplace behaviors that threaten the public. For example, COVID-19 mask ordinances have been imposed at both the local and the state levels since 2020. Despite research showing that such ordinances reduce the transmission of COVID-19 (Hartley & Perencevich, 2020; Shacham et al., 2020; Van Dyke et al., 2020), such legislation has not seen uniform adoption or public support across the U.S. (Jacobs & Ohinmaa, 2020; Shacham et al., 2020; The New York Times, 2021). Naming such ordinances after a victim who died of COVID-19 may increase public support, although given the extreme politicization of COVID-19 mitigation strategies (e.g. Kahane, 2021), naming techniques alone may not be enough to cut through the sizable influence of partisan politics.

Limitations

Perhaps the most important limitation is that this study was fielded during the nationwide “Black Lives Matter” protests, which likely influenced findings. Further, vignettes represent an artificial reality, and thus respondents may have reacted differently to “real world” proposals.

The vignette design itself is another limitation. Specifically, while victim naming was measured with dichotomous indicators for name use and the victim’s race, gender, and age, the vignette incorporated these details into a brief story about a specific victim killed by an unnamed distracted driver. Thus, results may have been influenced by the story about the victim, as suggested by psychological reactance theory (see Rosenberg & Siegel, 2018, pp. 286–287). Results may have been different if only the name of the law/victim were to change across the vignettes, without the accompanying story. As such, future research should compare names that tap stereotypes about the victim’s age, race, and/or gender, versus names that are more neutral.

This study did also not consider the influence of naming a law after multiple victims. Future research may want to look into this effect by building on the experimental work of Kogut and Ritov (2005). Such knowledge has clear relevance, as some laws incorporate multiple victims’ names (e.g. The Matthew Shepard and James Byrd Jr. Hate Crimes Prevention Act).

This study also considers a relatively commonplace form of deviance, rather than the rare heinous and violent crimes more commonly targeted by victim naming practices. A proposal targeting a more serious crime would likely have had stronger overall support for the law and punishment, due to punitive knee-jerk reactions to such crimes. Thus, conclusions may not generalize to proposals targeting crimes that are much more serious, but also much rarer. Relatedly, distracted driving legislation does not tap into the partisan political influence present in other types of legislation (e.g. gun control). Given the lack of prior research on the influence of naming effects on public support, this study should be considered an exploratory first step.

Due to space limitations, this study did not consider the unique influence of respondent demographics. While the successful randomization process meant respondent demographics should not have influenced the conclusions about the experimental manipulations, future research should consider how these demographics might influence support for such laws.

Finally, this experiment used the term African American rather than Black. As noted earlier, the Black racial term has very specific negative connotations (E. V. Hall et al., 2015; Pinsker, 2014), and support may have been lower for laws named after Latisha or DeShawn had Black been used.¹¹ The term Black may have more directly tapped into feelings towards the BLM movement and/or racial animus (see Drakulich et al., 2021). Future research should consider how variations in describing a victim’s race might influence public opinion.

Conclusion

This study used a national public opinion survey experiment to explore how naming a distracted driving law after a victim influenced support for either the law, or the punishment it authorized. Naming a law after a victim was found to increase support for the law and its punishment, but only for certain victims. Results suggested that the image of the “ideal victim” may have shifted or expanded to place greater emphasis on African American women, and less emphasis on White women. In highlighting the lack of laws named after African American/Black victims, Kulig and Cullen (2017) asked “Where is Latisha’s Law?” This study suggests that there may be an open policy window for proposing such laws, but whether that is a good thing or not depends very much on the details of the proposed law and its consequences for society.

Acknowledgments

I would like to thank Drs. Kareem Jordan, John Cluverius, Rebecca Stone, Craig Harper, Taylor John Wright, Elizabeth Brown, Justin Pickett, Robert Norris, Teresa Kulig, doctoral candidate Stacy Pearson, my wife Kathleen Hawkes, and many twitter followers who responded to random requests for information and references. Per usual, I am also indebted to Dr. Jason Rydberg for his helpful advice regarding various methodological issues, even if his advice about triangulating nonlinear interaction effects was a three way interaction of doodoo, caca, and poopoo as far as reviewer 3 was concerned. Relatedly, I found the resources on Dr. Trenton Mize’s website (https://www.trentonmize.com/) incredibly helpful in tackling the thorny interaction effects in this project, and I thank him for generosity sharing this information with the public. I am also grateful to the anonymous reviewers and Dr. Bryanna Fox for their helpful comments and suggestions on earlier versions of this manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This research was supported by the Center for Public Opinion and the Office of the Dean of Fine Arts, Social Sciences, and Humanities at the University of Massachusetts Lowell.

Notes on contributors

Kelly M. Socia

Kelly M. Socia, Ph.D., is an associate professor in the School of Criminology and Justice Studies at the University of Massachusetts Lowell, where he is also a fellow for the Center for Public Opinion. His research interests include punitive views, sex offense policies, public opinion, and policymaking.

Notes

1 This practice can be controversial, given the ineffective, symbolic policies and empty political promises commonly found in such laws (see Bravin, 2011; Duvall & Costello, 2021; Frank, 2016; Pason et al., 2017).

2 At the state level, after a 2019 Massachusetts impaired driving bill failed to pass with the uninspiring title ‘An Act Implementing the Recommendations of the Special Commission on Operating under the Influence and Impaired Driving, 2019,’ it was refiled in late 2021 as the ‘Trooper Thomas Clardy Law,’ (Murphy, 2021; WWLP, 2021). Massachusetts legislators also proposed ‘Nero’s Bill’, named after an injured police dog (Associated Press, 2021).

3 A news conference about New York’s Leandra’s Law, named after a girl who was killed by a drunk driver, provides a clear example of how legislators mention specific victims (or their families) to justify increased punishments. In the press release, New York Assembly Speaker Sheldon Silver (2009) mentioned a promise to the victim’s father that “…in memory of Leandra, [the Assembly] would advance DWI legislation that is tougher and more comprehensive than any that currently exists.” In other words, Speaker Silver is suggesting that one legitimate reason for increasing the punishment for DWI offenses is that it would symbolically honor the victim.

4 The terms ‘male’ and ‘female’ are the historic terms used in the literature. However, for the rest of this manuscript, I will use woman/man and girl/boy depending on whether I am referring to adults or children, respectively. There is also little agreement on whether White should be capitalized, and the reasons both for and against capitalization are compelling (see Ewing, 2020; Halley et al., 2011; Watson, 2013). In this manuscript, I adhere to APA (2021) guidelines, and capitalize both White and Black, acknowledging there are valid arguments against capitalization.

5 In a bit of irony, at the same time a new distracted driving bill was making its way through the Ohio legislature, a state senator was caught attending a virtual legislative meeting while driving (DeNatale, 2021; Ohio Office of Budget & Management, 2021). Despite the ‘office’ Zoom background, the senator’s seatbelt was a clear giveaway.

6 The term African American was used, given research finding that Black has more specific negative connotations (E. V. Hall et al., 2015; Pinsker, 2014), and can describe a much broader category of individuals (e.g., African Americans, Black Africans, Haitians, Jamaicans), compared to the term African American (Kareem Jordan, Personal Communication via Facebook Messenger, 2020).

7 Kulig and Cullen (2017, p. 993) noted that few laws carry names such as Latisha, and as that name is associated with African American culture, it would “evoke, in most observers, the assumption that the victim was Black.” I posit that DeShawn will yield a similar assumption of an African American male victim (Behind the Name, 2021).

8 These punishments were chosen to provide a wide variation in both the severity of the punishment (e.g., a $1,000 fine vs. a $100 fine, a 16-hour driving class vs. 40 hours of community service vs. 30 days of no license), and the type of punishment (e.g., financial loss, time commitment, loss of driving privilege, and public shaming.) These also represent either existing or otherwise realistic punishments for distracted driving violations.

9 Analysis of the unrotated factor loadings and scree plots suggested the eight punishments loaded well onto a single Punishment factor. However, rotated loadings suggested an alternative two-factor solution may be possible, with the first factor loading highly on the ‘tangible’ punishments (fines, license points, driving class, community service, and loss of license punishments), and the second factor loading highly on the two ‘embarrassment’ punishments (e.g. billboard and sticker). Given both the high Cronbach’s alpha and unrotated factor outcome for the single factor solution, I present only the single punishment measure in the final models. Alternative analyses (not shown) suggested relatively similar conclusions between models independently predicting the two separate factors.

10 Miratrix et al. (2018, p. 276) note that “for scholars examining population treatment effects using the high-quality, broadly representative samples recruited and delivered by top online survey firms, sample quantities, which do not rely on weights, are often sufficient. Sample average treatment effect (SATE) estimates tend not to differ substantially from their weighted counterparts, and they avoid the statistical power loss that accompanies weighting.” For these reasons, the models used in this study present unweighted results. Alternative models that applied weights largely did not change overall conclusions. As expected, weighting did increase standard errors, resulting in some coefficient estimates becoming non-significant, but with the same general size and direction of the effect as the unweighted estimates. Weighted results are available upon request.

11 Although a survey experiment by Wozniak (2020) found that White respondents allocated less public funding towards programs and services when explicitly told such funds would be used in ‘African American communities.’