Reshaping Court Systems: Issue Environments and the Establishment of Drug Courts

Abstract

State court systems are being reshaped by the widespread adoption of drug courts. However, there has been limited attention to what drives the decision to create drug courts in the states. I link the establishment of drug courts to local issue environments found in each state that support proactive judicial and legal elites. I propose hypotheses that link the density of drug courts to judicial professionalization, state funding of courts, drug arrest rates, and levels of court consolidation. I test these hypotheses with panel data from 2009 to 2014. I find that states with more professional judiciaries and higher rates of drug arrests are more likely to adopt drug courts, that higher levels of court consolidation show a small negative effect and, finally, that higher levels of state funding for court systems do not have consistent effects. I conclude that there is support for (a) using the issue environments approach to analyze the establishment of drug courts and (b) considering issue environments when analyzing other community oriented changes in state court systems.

Keywords:

• State courts
• court reform
• court consolidation
• drug courts
• problem solving courts
• judicial independence
• judicial professionalization

Acknowledgements

I acknowledge the useful and constructive comments of three anonymous reviewers of this paper. I also received useful feedback on many points on an earlier version of this paper from participants at a panel of the 2019 Midwest Political Science Association Convention.

Disclosure statement

I have no conflicts of interest to disclose. No organization or institution provided funding for this study.

Notes

1 It is important to realize that drug courts do not contravene usual legal procedures. Drug court defendants have already been made liable to legal sanctions for their offenses, usually by a guilty plea. If they qualify for entrance to a drug court, they voluntarily enter the programs set up by the court to avoid application of those sanctions (Nolan 2001). See Chapter 3 of Nolan’s book for more particular examples of differences in procedure.

2 This was done in part because of a lack of complete data on most of these courts and in part because this study is not a general examination of all community oriented courts.

3 For examples of comparative research on court unification, see Flango (1975), Berkson (1978), Glick (1980, 1981), Scheb and Methany (1988), and Lightcap (2003). Most of this work was firmly in earlier traditions of research on policy innovations in the states pioneered by Walker (1969), tying court unification innovations to the political effects of economic development. Lightcap’s work uses a different model, as does the research presented here.

4 See also Douglas and Hartley’s (2011) study of drug court funding. While not concentrating on the establishment of drug courts, their work is comparative.

5 This has changed as drug courts have become more prominent. As of 2009, there were statutes in 31 states providing for drug courts in some fashion (Huddleston and Marlowe 2011, Table 7, p. 38). Most of this legislative activity, however, has taken place after the fact.

6 The creation of drug courts bears some resemblance to “street-level policy entrepreneurship” as described in recent literature (Arnold 2015). However, the judges involved are almost always elected officials and have wide discretion over their work. They are not subordinate bureaucrats and are initiating new procedures directly through rule-making and administrative changes, not by manipulation of existing regulatory structures. The process also differs from research concerning entrepreneurial judges. This line of research has concentrated more on judges leading changes in jurisprudence and jurisdiction of already established courts, not on using rulemaking and administrative authority to establish new courts using new judicial processes (see, for example, McIntosh and Cates 1997; Crowe 2007).

7 This elite confidence is reflected in the consensus Douglas, Raudla, and Hartley (2015) found; judges and court personnel in localities that had established drug courts accepted the treatment model they used, were willing to create new courts through court rules or state statutes, and had the requisite expertise to write grant proposals for seed money (see also Farole et al. 2005).

8 There is a voluminous literature on the reasons for this reform program. The original and still unsurpassed statement for it is Roscoe Pound’s famous 1906 address to the American Bar Association (Pound 1937).

9 For instance, Los Angeles County in California had 18 different drug court programs in 2014 (Bureau of Justice Assistance n.d.). All of these were part of the unified circuit court of the county and operated in courthouses in separate cities and neighborhoods. This is an extreme example, but the most populous counties in each state have drug courts, often multiple ones.

10 There are three difficulties here. First, there are counties in some states that are covered by drug courts established by judicial circuits rather than by individual counties. This is usually an expedient brought on by the small populations of the counties; but it does reflect the spread of drug courts and their jurisdiction as effectively as for more populated areas. Second, Virginia allows any established city to separate itself from the county containing it and set up a separate government recognized by the state as equivalent to a county. I treated both counties and independent cities as “counties” for this state. The third problem is that over half of Alaska has not been incorporated for local government. I counted the existing counties and their drug courts only and left the unorganized “census areas” out of the dataset. The data were drawn from Bureau of Justice Assistance Drug Court Technical Assistance/Clearinghouse Project Summary of Drug Court Activity by State and County Reports (n.d.).

11 Many of the judges who end up running drug courts are drawn from limited jurisdiction courts. However, it is the general jurisdiction judges who usually make the decisions that allow the courts to be established in the first place. Indeed, general jurisdiction judges are usually the main instigators of community oriented reforms of all kinds in local court systems (Tobin 1999; Berman and Feinblatt 2002).

12 Determining the number of state general jurisdiction judgeships was more complicated than for federal district judges. It may be difficult to believe, but there is no single comprehensive source of time series data for the number of state general jurisdiction judges. I proceeded in three steps. First, I took the number of general jurisdiction judgeships for each state from Examining the Work of State Courts for 2009 and 2010 (LaFountain et al. 2011, 2012). I then compared these numbers to the State Court Organization judgeship figures for 2016 (Strickland et al. 2017). If there was no change from 2010—this was often the case—then I extrapolated figures from 2010 to 2011–2014. If there was a change or there was missing data for 2016, I went to the state’s Administrative Office of the Courts internet site, examined their annual reports for changes, then included those readings as part of the final index figures for 2011 to 2014.

13 I also calculated all the models in this study using the actual percentages of direct state judicial and legal expenditures in each state and found no substantial differences from the estimates using Berkson’s index.

14 Again, gaps in the data required a two-step procedure. First, I compared the National Center for State Courts (2015, 2018) state court structure charts for 2009 and 2010 to their 2014 structure. If there was no change—again, this was usually the case—I extrapolated the index readings for 2010 to 2011–2013. If there was a change, I went to the state’s Administrative Office of the Courts internet site, examined their annual reports to determine changes in court structure between 2011 and 2013, and calculated the index for each year accordingly.

15 Berry and his colleagues have also developed ideology scores for state legislatures. I did not use these indicators since the usual decision to establish drug courts is taken at the local level, not at state capitals. Further, as Berry, Fording, Hanson, and Crofoot (2021), point out, the citizen ideology scores are a reliable indicator of the “operational ideology” of state populations; i.e. how willing the population is to see more expansive government policies. These data have good face validity and are well suited for the time-series cross-sectional models used in this study (Berry, Fording, Hanson, and Crofoot 2021).

16 Often income figures should be transformed to accommodate differences between the states. Examination of plots of the distribution of per capita income for 2009 to 2014 failed to reveal any gross distortions. I used income per capita in thousands of dollars to aid interpretation of the estimates. I did not use inflation adjusted income figures for the calculations because the average inflation rate for 2009–2014 was 1.66% (Bureau of Labor Statistics n.d.).

17 The main difficulty with these data is that they are “sluggish,” i.e. readings on the independent variables change slowly over time relative to the dependent variable. This can make estimates of the effect of independent variables unreliable (Wilson and Butler 2007; Clark and Linzer 2015). This is a problem with institutions like courts that change slowly and makes analysis of panel data involving them more difficult. Some scholars (Beck 2001; Gelman and Hill 2007) recommend against using fixed effects models at all with datasets like this one. Based on their simulations using sluggish data, Clarke and Linzer advise using fixed effects models for datasets like the one used here as the best course of action. Although their limits for using fixed effects models are very close to those in these data, I have followed their guidance.

18 Table A.2 contains the levels of significance for the estimates using a two-sided t-test, adjusted R-squared coefficients to show the amount of variance explained by the models, average VIF or GVIF (generalized variance inflation factor scores required for models with factors) scores for the indicators in each model to indicate multicollinearity, and F statistic readings for significance of the models overall. The VIF and GVIF readings for the models are of particular interest. The general rule for VIF and GVIF scores is that multicollinearity is a problem if the score for any indicator is over 5 (Studenmund 1997). This is not the case for indicators in any of the models used in Figure 2; the highest score for any of them is 1.64.

19 The graph in Figure 2 was generated using Solt and Hu’s (2015) dotwhisker package for the R computer language.

20 This result may arise from the complex relationship between economic development and both drug use and its consequences in the United States. The relationship may be confounded by the influence of racial discrimination and its role in depressing economic growth, thus laying the ground for social disorder and consequent drug use (Donnelly et al 2021). Further research is obviously needed to elucidate the link between economic developments to the proliferation of drug courts, but that is a project for another day.

21 The graph in Figure 3 was generated using Wickham’s (2016) ggplot2 package for the R computer language.