Combining Frequent Alcohol Testing with Swift-Certain-Fair Sanctions: Summary of the Peer-Reviewed Literature on 24/7 Sobriety and Ideas for Future Research

1. Introduction

Most people who consume alcohol to do not encounter serious problems related to their use, but for those who do, the harms can be severe. Heavy alcohol use is associated with increased risk of acute injury, weakened immune system response, chronic illnesses affecting the heart, liver, and pancreas, and several forms of cancer (Zakhari 2006). The harms of alcohol use also extend to others. Heavy alcohol use is associated with intimate partner violence (Klostermann and Fals-Stewart, 2006); firearm violence, including suicide (Branas et al. 2016); youth homicide (Parker et al. 2011); and other forms of violence (Duke et al. 2018). Twenty-five percent of the 13,384 persons killed in alcohol-involved traffic crashes in 2021 were someone other than the intoxicated driver (Stewart 2023).

In the United States, alcohol-related harms are disproportionately concentrated among persons involved in the criminal legal system (Greenfeld 1998; Rand et al. 2010). Estimates suggest the costs of alcohol-involved crime exceed $80 billion per year (Miller et al. 2006). Among current drinkers, persons arrested for driving under the influence of alcohol (DUI) and those who had alcohol-involved traffic crashes are more likely to have an alcohol use disorder (AUD) (Yao et al. 2018). Approximately 30% of persons incarcerated in U.S. state and federal prisons reported drinking alcohol at the time of the offense (Maruschak et al. 2021). Persons under community supervision are also substantially more likely to have an AUD, and evidence indicates that probationers with an AUD are more likely to experience violations than other clients (Galvin et al. 2022). Thus, interventions that deter heavy alcohol use may help reduce criminal legal system involvement and its costs.

Individuals in the United States aged twenty-one and over have an effective “license” to purchase and consume as much alcohol as they want (Kilmer and Humphreys 2013). This raises an important question: when is it appropriate to suspend someone’s “license to drink” because their use is leading them to repeatedly threaten public safety? Conditions of bail, probation, and parole may require individuals to abstain from alcohol, but this can be difficult to enforce. Compared to many other drugs, alcohol leaves the body quickly, making it harder to detect via traditional alcohol tests. Further, there may be resistance to sanctioning someone under supervision for their first substance use violation, weakening any deterrent effect the condition was intended to have.

This is not the case with most 24/7 Sobriety programs (hereinafter, 24/7), which take the concept of suspending someone’s “license to drink” very seriously. These programs focus on alcohol consumption itself, rather than the DUI event that may result from excessive consumption (DuPont and Humphreys 2011). First implemented in South Dakota roughly twenty years ago, 24/7 requires those ordered to abstain from alcohol to be monitored via very frequent testing (e.g., twice-daily breathalyzers, remote alcohol monitoring). Participants face swift, certain, and moderate sanctions for testing positive or missing a test—typically a night or two in jail. The goal is to hold people accountable and create a credible deterrent threat. The program was created by practitioners trying to reduce alcohol-involved crime and is consistent with what we know about criminal deterrence and behavioral economics (Bickel et al. 2014; Nagin 2013; Paternoster 2010; Sloan et al. 2014). It has now spread from South Dakota to jurisdictions in other states, and a modified version of the program is operational in England and Wales.

This article is intended to provide more information about 24/7, discuss the peer-reviewed evidence concerning its effectiveness, and offer ideas for future research opportunities. Section 2 provides background about how the program started and spread beyond South Dakota, paying special attention to how the program has been adapted over time and place. Sections 3 and 4 summarize the peer-reviewed evidence on 24/7 and how it has (and has not) been incorporated into various research reviews. Section 5 discusses ideas for future research, especially in the context of conducting randomized controlled trials. Section 6 concludes.

2. How 24/7 Spread within and beyond South Dakota

The concept for 24/7 was introduced in 2003 in response to South Dakota Governor Mike Rounds’s call for ideas to reduce the state’s reliance on incarceration (Rapid City Journal 2003). The state’s new attorney general, Larry Long, saw promise in enforcing abstinence orders for persons arrested for crimes where alcohol was a contributing factor. When he was a prosecutor, he observed that people who frequently went before the bench were often there because of an alcohol-related crime (Long 2009). Therefore, limiting alcohol consumption could reduce criminal involvement. He convinced the governor to implement a pilot program that combined abstinence orders with twice-a-day breathalyzer tests of arrestees with multiple DUI offenses.

In 2005, South Dakota began its 24/7 pilot in its two most populous counties (Pennington and Minnehaha) and three rural counties (Bennett, McCook, and Tripp). Through word of mouth—among judges, prosecutors, and sheriff department officials—other counties began implementing the program. With this expansion came adaptations. Some jurisdictions added transdermal alcohol monitoring devices as an alternative to breathalyzer testing. Some also added urinalysis and sweat patch screening for illegal drug use. Some judges began assigning participants to 24/7 for offenses other than DUI (e.g., domestic violence cases where alcohol use was believed to be a contributing factor).

In February 2007, the state legislature enacted House Bill 1072 to establish 24/7 Program coordinator, formalize program rules, and provide resources to counties who wished to run the program. The state also established a schedule of fees paid by participants—$2 per day for breathalyzers and $6 per day for remote alcohol monitoring, plus fixed initiation fees—and they defined the apportionment of these funds to the state and counties. These steps mitigated some of the uncertainty surrounding implementation and the risk of financial loss to jurisdictions. The program spread quickly throughout the state.

The 24/7 program quickly drew national attention, despite the lack of peer-reviewed evidence of its efficacy. It won awards from the National Highway Traffic Safety Administration, the McGovern Foundation, and the Council of State Governments. Neighboring states also began replicating the program. In 2008, North Dakota started a large-scale pilot in fourteen of the state’s fifty-three counties, making approximately 75% of the state population eligible. Montana followed suit, beginning a pilot in 2010 and authorizing the program in state law in 2011. Laws authorizing 24/7 programs were enacted in states such as Idaho, Iowa, and Wyoming,¹ and some jurisdictions in states such as Alaska, Nebraska, Utah, Washington, and Wisconsin have also run similar programs; however, program design and implementation choices can vary by jurisdiction.

There have been two notable federal efforts to encourage expansion of 24/7. Under the 2015 Fixing America’s Surface Transportation (FAST) Act, states could apply for funding to support implementation of 24/7 programs over fiscal years 2016 to 2020. Subsequently, a bipartisan group of sponsors introduced the Supporting Opportunities to Build Everyday Responsibility Act of 2022 (the SOBER Act) in the U.S. House of Representatives to incentivize the implementation of 24/7 programs nationwide.² If passed, the SOBER Act would have allocated $250 million over five years to states to introduce and expand existing 24/7 programs, as well as facilitate program evaluations.

Finally, a “cousin” of 24/7 has made its way across the pond. After piloting “mandatory sobriety” in Humberside, Lincolnshire, North Yorkshire, and South London, the program is now being rolled out across England and Wales (Alcohol Abstinence & Monitoring Requirements; Hodges 2022). The program only uses alcohol monitoring bracelets, and insights from one of the pilot programs suggest it may differ from 24/7 in other important ways:

Whereas those in South Dakota were subject to an immediate night or two in jail for non-compliance, [Dr. Laura] Bainbridge reports that in the South London pilot there was no possibility of “flash incarceration” and the first violation was only met with a warning. However, even with the very low expected sanction for non-compliance, a preliminary analysis suggested that more than 90% of the 111 individuals in the pilot were fully compliant.

(Kilmer 2019, 1706)

3. Review of the Peer-Reviewed Literature

There now exists a strong peer-reviewed literature suggesting that 24/7 participants largely abstain from heavy drinking during the program and that 24/7 may be an effective means of addressing the harms associated with DUI and other alcohol-involved crimes.

Heavy Drinking

South Dakota’s 24/7 required frequent monitoring with twice-a-day breathalyzer tests (or, to a lesser extent, continuous alcohol monitoring bracelets) and imposed swift and certain sanctions for testing positive or skipping a test—often a brief (1–2 night) jail stay. North Dakota imposed similar requirements, though the exact specifications were altered to allow for some drinking (BAC<0.02) and greater use of continuous alcohol monitoring bracelets. The non-zero BAC level may have implications for compliance given that it reduces the certainty of the sanction for drinking (Midgette et al. 2021).

As reported by Kilmer and Midgette (2020), “From 2005 through February 2017, more than 30,000 unique South Dakotans participated in 24/7, accumulating more than five million days without a detected alcohol violation, missed test, or tampering event” (2). Citing data from the South Dakota Attorney General (2019), Midgette and colleagues (2021) note that: “More than 99% of the breathalyzers tests were taken and passed (including no-shows in the denominator) and more than 99% of the days on the bracelet have neither confirmed alcohol use nor an attempt to tamper with the device” (651).³ Midgette et al. (2021) calculated similar figures for 24/7 participants in North Dakota submitting to breathalyzers tests, observing rates for testing negative at 95.8% with an additional 2% excused. The results suggest that 24/7 participants largely abstained from heavy drinking during their participation in the program.

Public Health and Safety

If heavy drinking is an underlying cause of participants’ risky behaviors, then the high compliance rates should be accompanied by lower levels of future alcohol-involved offending and health consequences. The peer-reviewed literature provides insight into 24/7’s effects on these outcomes with two types of studies: ecological analyses at the county level and individual-level studies.

County-level analyses.

The county-level studies exploit the variation in the implementation of 24/7 across counties within a state to assess whether there has been a reduction in the prevalence of adverse outcomes among counties with the program relative to those without the program. Given that all counties eventually adopted the program, the approach essentially compares changes in outcomes among counties that had implemented the program versus those that had not yet implemented the program. The identification strategy relies on the assumption that the roll-out of 24/7 across counties was not related to the outcome of interest—for example, the assumption that those counties with worse outcomes were not the first to implement the program. Therefore, the timing of those changes could be used to identify the effect of the program. An assessment of the roll-out supports this plausible exogeneity (see Kilmer and Midgette 2020).

The first county-level evaluations of South Dakota’s 24/7 program showed significant reductions in county-level criminal legal outcomes among counties with the program versus those that had not yet implemented the program. The program was associated with a 12% reduction in repeat DUI arrests and a 9% reduction in domestic violence arrests (Kilmer et al. 2013), as well as a 4% reduction in mortality (Nicosia et al. 2016). Using the same analytic approach, a county-level study of North Dakota’s 24/7 found a 9% reduction in DUI arrests—a similar reduction to that documented for South Dakota (Midgette et al. 2020).

Individual-level analyses.

The individual-level studies also rely on quasi-experimental methods. As 24/7 was not a randomized controlled trial, there is concern that selection of individuals into the program would bias the findings. For example, if those most (or least) likely to recidivate were more likely to enter 24/7, then regressions based on observational data would underestimate (or overestimate) the effect of the program. Therefore, these studies use instrumental variable style approaches (e.g., bivariate probit, two-stage residual inclusion with an individual frailty). Specifically, the individual-level studies exploit that same variation in county-level program availability to address the potential selection effects and assess whether 24/7 participation is associated with a lower likelihood of adverse events among those in the program vs. not in the program. This identification strategy requires that 24/7 program availability in the arrest county was strongly predictive of individual-level participation, as measured by the F-statistic or equivalent. In addition, the county-level availability cannot influence individuals’ outcomes (e.g., arrest, mortality) through any mechanism except individual-level participation, an assumption that is intuitive.

Similar to the county-level studies, the individual-level studies also point to a reduction in both re-arrest and deaths. The first individual-level study among repeat DUI arrestees demonstrated that 24/7 participants in South Dakota were substantially less likely to be re-arrested than non-participants: 49% lower risk during 1-year, 35% during 2-year; 26% during 3-year follow-ups (Kilmer and Midgette 2020). Although based on only a nascent program, findings from another instrumental variables analysis focused on Montana also point to beneficial effects on recidivism (Midgette and Kilmer 2021). Another study of North Dakota using different methods also indicated improvements in citations and crashes among participants (Vachal and Kubas 2018).

Finally, an instrumental variables analysis focused on those arrested for DUI in South Dakota demonstrated that 24/7 participants were substantially less likely to die than non-participants during five-year follow-up, particularly in models that address selection into the program (Nicosia et al. 2023). These approaches suggest that participants were about 50 percent less likely to die during follow-up than non-participants. The results are robust to sensitivity based on selection of the index arrest and other factors.

There are other studies of 24/7 programs that have not yet been published in the peer-reviewed literature. One example of this “grey” literature is an evaluation of Utah’s 24/7 program. Utah’s 24/7 program was originally designed as a randomized controlled trial, which would have been important contribution to our understanding of 24/7. Unfortunately, the authors of the evaluation noted concerns with the randomization process (Mueller et al. 2021), meaning our ability to draw strong conclusions are constrained by concerns about randomization, small sample size, and limited follow-up.

4. Summaries of Research on 24/7

The evidence on 24/7 has been summarized in both policy and academic literature. The 2018 report from the National Academies of Sciences, Engineering, and Medicine, Getting to Zero Alcohol-Impaired Driving Fatalities, describes the evidence on 24/7 as promising; although the authors noted that all peer-reviewed evaluations relied on quasi-experimental methods and aggregated data (this was prior to the publication of the aforementioned individual-level studies [see Kilmer and Midgette 2020; Midgette and Kilmer 2021; Nicosia et al. 2023]). The National Institute of Justice’s evidence-based practices summary resource, CrimeSolutions.gov, recognizes both the North Dakota and South Dakota programs as “promising” evidence-based practices based on single county-level studies of each program by Midgette et al. (2020) and Kilmer et al. (2013), respectively. Other studies we discuss in the prior section are not included in this assessment, and CrimeSolutions.gov policy requires the programs to be assessed separately despite their similarities.

Other reviews have grouped 24/7 with Hawaii’s Opportunity Probation with Enforcement (HOPE) and derivative programs as a category of interventions described by Kleiman (2014) as “Swift, Certain, and Fair” (SCF). Yet there are some important differences between 24/7 and HOPE. The 24/7 program focuses predominantly on alcohol-related criminal offenses and is more “certain” than HOPE, the latter relying on randomized testing rather than mandatory high-frequency testing. Larkin (2015) describes these SCF programs as, “sensible, humane, and effective mechanisms for dealing with substance abuse and crime.” More recently, in a review of criminal legal practices that promote desistance from crime, Doleac (2022) frames 24/7 as evidence that programs which prioritize increasing the perceived probability of punishment can be effective deterrents. Cullen and colleagues (2018) and Pattavina and colleagues (2023) find the evidence of SCF program effects to be to mixed at best, but both reviews include a single peer-reviewed 24/7 study in their evidence base, misclassify the research designs, and do not include in their reviews other extant peer-reviewed evaluations of 24/7.⁴ Midgette and colleagues (2023) note that while the body of evidence on 24/7 specifically is promising, there is still much to learn about implementation of SCF programs, noting we do not yet know which components and design choices are most important.

5. Future Directions for Research on 24/7

Based on a decade’s worth of peer-reviewed evidence on 24/7 and the continued problems that heavy drinking poses for many under community supervision (and sometimes others in their households and communities), it seems reasonable for other jurisdictions to conduct their own pilot programs and see if it works for them—they won’t know unless they try.

For community corrections agencies committed to the idea of evidence-based practices, we hope they will consider helping to build the evidence on 24/7 by rigorously evaluating the pilots, possibly with a randomized controlled trial (RCT).

The remainder of this section is divided into two parts. The first highlights the many questions that remain about 24/7 and the causal mechanisms that may be driving the results. The second offers some considerations for jurisdictions and funding organizations interested in implementing or supporting an RCT.

Outstanding Questions about 24/7

This section highlights eight questions, fleshing out six previously listed by Kilmer (2019) and adding two others.

How long should individuals stay in the program to produce a lasting effect?

When 24/7 spread throughout South Dakota, the state did not mandate how long participants should stay in the program; it was left up to judges who could also change duration depending on how the participant was doing. Predictably, this generated a fair amount of variation. Because both judicial and participant behavior influence program duration, it would likely be misleading to simply correlate time in program with various outcomes while controlling for other factors.

So, how can we figure out the optimal duration of 24/7 for various types of individuals? One option is to analyze what has happened in North Dakota, which initially allowed judges to decide as well, but then passed a law mandating days in the program by offense type and criminal history. Fortunately, there’s some ongoing research focused on this. Another option is to build this into an RCT.

Do certain types of alcohol testing lead to better 24/7 outcomes?

While most 24/7 participants blow into a breathalyzer twice a day, a large number wear alcohol monitoring bracelets. Some jurisdictions have also used remote breath devices that allow participants to test from anywhere, and there are other approaches to alcohol testing as well (e.g., kiosks). We know that different types of testing detect different patterns of alcohol use with varying levels of confidence (see Brobbin et al. [2023] for an overview of the literature), so this is an important question to answer. There may also be something important about the process associated with different types of testing. Does the routine of going to a sheriff s department twice a day and having positive interactions with staff and negative tests also influence the effectiveness of 24/7? Because participants often choose whether they want to pay more for remote monitoring, researchers cannot simply compare outcomes by testing type and generate a credible answer to this question. This is something that could be easily answered in an RCT, but considering the role that third-party vendors play in this process is critical (further addressed in the next section).

What is the minimum level of sanction needed to produce a deterrent effect?

Most jurisdictions with 24/7 require those testing positive or who have missed a test to spend a night or two in jail. What if the sanction was only four hours in holding cell or some type of house arrest? Would this generate the same deterrent effect? What about other types of sanctions that do not involve incarceration?

Would combining sanctions with positive incentives for compliance improve outcomes?

There is a tremendous amount of research supporting the idea that small rewards (e.g., movie passes or gift certificates) for extended periods of abstinence can reduce substance use among those with alcohol and other substance use disorders; this is typically referred to as contingency management (Petry et al. 2000). Because 24/7, as it has been implemented in most places, is focused on the “stick” of potential punishment, could better outcomes be achieved by incorporating some carrots into the model (Humphreys and Kilmer 2020)?

Would combining 24/7 with treatment lead to better outcomes?

While 24/7 doesn’t require individuals to participate in substance use disorder treatment, it does not forbid them from doing so. One could imagine an approach that combines 24/7 with outpatient or medication treatment. Another approach could initially mandate individuals whose alcohol use has led them to repeatedly threaten public health and safety to 24/7, and mandate treatment after repeated positive or missed tests; something Dr. Angela Hawken (2010) refers to as “behavioral triage.” Whether combining these programs would produce better outcomes is very much an empirical question.

How does 24/7 participation affect the family members and intimates of participants?

The peer-reviewed research on the program has focused on secondary analyses of administrative data, largely focusing on criminal recidivism or mortality at the county and individual levels. To assess the full costs and benefits of a program with the potential to drastically reduce heavy drinking, researchers should consider outcomes for those who live with participants and possibly other family members. How did 24/7 affect participants’ intimates; did it lead to more or less stress in their relationships? Did it influence the amount of fighting and violence in the household? With one county-level study finding that 24/7 implementation was associated with a decrease in arrests for domestic violence (Kilmer et al. 2013), it is critical that primary data collection with study participants and their families be included in future studies.

Which intervention is most cost-effective at reducing future arrests, traffic crashes, mortality and other alcohol-related outcomes, including domestic violence?

If one’s goal is to reduce the negative consequences of heavy alcohol consumption among those subject to community corrections, there are many options that could make a difference (e.g., 24/7, substance use disorder treatment, interlock devices). However, these approaches come with different costs, so what many decision makers want to know is which intervention, or combination of interventions, produces the most “bang for the buck.” Ideally, researchers would look beyond one outcome (e.g., DUI arrests) and produce cost-benefit ratios for the various interventions and the various sought-after outcomes. This type of analysis is ripe for an RCT.

Can 24/7 generate the same results in more urban areas outside of the Great Plains?

We cannot know the answer to this until some urban jurisdictions outside of the Great Plains experiment with the program and rigorously evaluate their efforts. After piloting a modified version of the program in a handful of places in England, including South London, the Alcohol Abstinence and Monitoring Requirements were rolled out in England and Wales. While there were some positive results from descriptive analyses of the South London pilot (Bainbridge 2019; Finlay and Humphreys 2017), we are unaware of any rigorous analysis of the pilots or of the more recent rollout.

Logistical challenges related to transportation and testing volume mean that urban jurisdictions will probably have to rely more on remote monitoring or kiosks than in-person breathalyzers that have been most common in the Great Plains. There is also the issue of people missing their scheduled tests. In many urban areas, it is not uncommon for those with multiple bench warrants—sometimes for violent offenses—to be free in the community with little risk of apprehension unless they get picked up for something else. Would law enforcement make it a priority to target those who miss an alcohol test even if they do not give similar attention to those with bench warrants who are alleged to have been involved in more serious crimes?

This is not meant to discourage more urban jurisdictions from considering 24/7, but it highlights how the program may have to be adapted to succeed. One idea would be to start with more intensive supervision for a small number of participants. Once a credible deterrent threat has been created, jurisdictions could slowly increase the number of participants without dedicating more officers or resources to the program.

Considerations for a 24/7 RCT

We are aware of only one RCT that has been conducted thus far (Weber County, UT), but its small sample size and other implementation issues raise questions about the findings from that effort. We offer a few ideas for jurisdictions considering their own RCT as well as for government agencies or philanthropic organizations that may be interested in funding this type of research.

Allow for a “burn in” period before starting the evaluation.

Implementing frequent testing with swift-certain-fair sanctions may require a very different approach than what some jurisdictions are used to. Thus, it may take time to iron out the details and work through logistical challenges that may influence the programs’ ability to produce a deterrent effect. This suggests that researchers should allow some time before starting a formal evaluation. Indeed, the available evidence suggests that the most positive results from the multi-site trial of HOPE were for the jurisdiction that already had experience implementing SCF (Humphreys and Kilmer, 2020).

Ensure a large enough sample size to account for non-compliance by community corrections officers or judges.

Even waiting through a “burn-in” period will not guarantee that judges or community corrections officers will always follow the study protocols. They may add people to the 24/7 condition when they are supposed to receive a different intervention. They may also fail to sanction someone for a violation. The extent of this “non-compliance” will determine how much can be learned. There are statistical methods that allow researchers to account for deviations from the RCT protocols (e.g., see Kilmer’s [2008] analysis of a drug testing experiment in California), but these approaches may require a larger sample size. The necessary sample size will depend on the outcome(s) being evaluated as well as predictions about the amount of non-compliance. We strongly urge jurisdictions to consult with an analyst who can conduct these sample size calculations.

Ensure all outcomes are pre-registered and published simultaneously.

While pre-registration of outcomes and methods is becoming more common in RCTs involving criminal legal interventions, that does not mean that researchers always publish all the outcomes at the same time; this can affect the inferences and narratives about the program. Indeed, as noted by Humphreys and Kilmer (2020), this happened with the multi-site demonstration field experiment (DFE) of HOPE and the main outcome evaluation published in Criminology and Public Policy (CPP):

When the DFE was pre-registered at ClinicalTrials.gov (2012), five primary and six secondary outcomes were listed. The first measure listed under the primary outcomes was “Illicit Drug Use.” Although the discouraging pre-registered outcomes were reported in the CPP article, the pre-registered—and positive—drug outcomes were not. As Hawken notes (2018), neither were illegal drug use and related outcomes such as health care and labor market outcomes incorporated into the subsequent cost-effectiveness analyses of the DFE.

It is imperative that those funding and publishing these studies make sure researchers are presenting all the pre-registered outcomes, especially those listed as primary outcomes.

Consider the role of third-party vendors who may own the alcohol-testing technologies.

Some courts and community corrections departments do not own the supervision technologies they require individuals to use or wear (e.g., alcohol monitoring or house arrest devices). Instead, they work with third-party vendors who charge individuals directly for using their technologies (and sometimes collect processing fees at the beginning and end of supervision periods). Depending on the specifics of the RCT, these vendors’ bottom lines could be affected, and instances where the impact is negative could create challenges. Of course, study funders could simply bypass these vendors and work directly with the technology companies. If jurisdictions and researchers do end up working with local third-party vendors, that adds another element of the program design that could vary by location and affect outcomes if a multi-site trial is conducted.

6. Concluding Thoughts

The 24/7 programs largely focus on reducing alcohol consumption among justice-involved individuals whose alcohol use has led them to repeatedly threaten public safety. Participants are ordered to abstain from alcohol use and subject to frequent alcohol testing (e.g., twice daily breathalyzers, remote alcohol monitoring); those testing positive face an immediate sanction—typically a night or two in jail. Unlike drug courts and other forms of coerced substance use disorder treatment, 24/7 does not require participants to enter a treatment program or attend self-help group meetings; however, these approaches are not necessarily mutually exclusive. The empirical evidence suggests important gains with respect to heavy drinking, criminal legal involvement, and mortality. Several questions remain about the design of these programs (e.g., optimal program length, possible benefits of incorporating positive incentives, efficacy of different alcohol testing technologies), many of which can best be answered in an experimental setting. There is also much to be learned from experimental and quasi-experimental studies of 24/7 programs outside of the Great Plains.