Appointed or Elected: How Justices on Elected State Supreme Courts Are Actually Selected

Abstract

During at least part of the post–World War II period, the constitutions of thirty-six states called for the popular election of the judges of the states’ highest courts. In practice, only slightly more than half of those judges (excluding strictly interim appointees) initially obtained their positions by election. This article examines the likelihood of initial election in actual practice, how it has varied over time, and various factors that might be related to election versus appointment (e.g., type of election, mandatory retirement). It concludes that state norms play a substantial role in determining patterns of actual selection.

Full text

Appointed or Elected: How Justices on Elected
State Supreme Courts Are Actually
Selected
Herbert M. Kritzer
During at least part of the post–World War II period, the constitutions of thirty-six
states called for the popular election of the judges of the states’highest courts. In practice,
only slightly more than half of those judges (excluding strictly interim appointees) initially
obtained their positions by election. This article examines the likelihood of initial election in
actual practice, how it has varied over time, and various factors that might be related to
election versus appointment (e.g., type of election, mandatory retirement). It concludes
that state norms play a substantial role in determining patterns of actual selection.
INTRODUCTION
One of the most visible ways that the American legal system stands apart from
other systems around the world is the use of elections to select and/or retain judges at
the state level.1This article focuses on the initial selection of state supreme court justices
in those states where the state constitutions or state statutes call for justices to be elected,
“elected by the people”in the words of several state constitutions.2Specifically, it exam-
ines the actual initial selection of justices in those states, considering two broad questions:
what percentage of justices are in fact initially elected and what factors can account
for whether a justice was initially elected or initially appointed?
The phenomenon of judicial elections was largely a product of the mid-nineteenth
century, when many states shifted from executive appointment (or legislative election)
of judges to popular elections and new states chose to elect their judges. Shugerman
Herbert M. Kritzer (kritzer@umn.edu) is Marvin J. Sonosky Chair of Law and Public Policy emeritus,
University of Minnesota Law School, United States, and Professor of Political Science and Law emeritus,
University of Wisconsin–Madison, United States. Damon Cann, Melinda Gann Hall, Charles Geyh, James
Gibson, and Gregory Goelhauser provided helpful input as this article was being developed. The reviewers
for Law and Social Inquiry gave me excellent feedback on the initial submission to the journal.
1. The three other countries that use popular elections in some way are Japan for retention of supreme
court justices (see O’Brien 2006, 359), Switzerland for local judges in some cantons (see Wipf 2002, 1569),
and Bolivia for national-level courts after 2009 (see Driscoll and Nelson 2015).
2. In New York and Maryland, the highest court is called the Court of Appeals, and to confuse matters,
in New York, the upper trial court and intermediate court of appeals are called the Supreme Court, trial
division and appellate division. Additionally, in Texas and Oklahoma, there are separate top courts for crim-
inal and noncriminal matters, with the latter called the “Supreme Court”and the former the “Court of
Criminal Appeals.”Throughout this article, I will use the term “supreme court”to refer to a state’s highest
court or, in the case of Texas and Oklahoma, highest courts.
Law & Social Inquiry
Volume 00, Issue 00, 1–36, 2022
© The Author(s), 2022. Published by Cambridge University Press on behalf of the American Bar Foundation. 1
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike
licence (https://creativecommons.org/licenses/by-nc-sa/4.0/), which permits non-commercial re-use, distribution, and reproduction
in any medium, provided the same Creative Commons licence is included and the original work is properly cited. The written
permission of Cambridge University Press must be obtained for commercial re-use.
Law&
Social
Inquir
y
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

(2012, 6) argues that the core reason for this shift was to increase the independence of
judges from other political elites. Whether there was an increase in independence is
unclear because leaders of political parties, including the very same governors who
previously had made the appointments, largely continued to control the nomination
process, at least until states adopted direct primaries or nonpartisan election of judges
starting in the early twentieth century (Hall 1984, 354).3
A key issue that states using elections must deal with is how to fill midterm vacan-
cies arising due to death, resignation, or retirement occurring during a justice’s term.
The three possibilities are (1) leave the position open until the next regularly scheduled
election, (2) hold a special election to fill the position, or (3) empower someone, usually
the governor, to appoint someone to fill the position until either the next regularly
scheduled election or the end of the departing justice’s term. Given the problems of
leaving a vacancy unfilled for an extended period of time (e.g., the increased possibility
of tie votes) and the cost of running a statewide special election, all states but one use
appointments to fill midterm supreme court vacancies, some until the next regularly
scheduled election and some for the remainder of the departing justice’s term; the
one state currently using special elections is Louisiana, where state supreme court jus-
tices are elected by district rather than statewide. With some exceptions, discussed
below, the appointee can then run to continue in the position when the subsequent
election is held.
In 2017 the Brennan Center for Justice (BCJ) published a report highlighting the
initial appointment of state supreme court justices serving as of August 2016 in the
twenty-two states where the state constitution still specifies that justices are to be pop-
ularly elected (Berry and Lisk 2017).4According to the report, only 54.9 percent of the
then-sitting justices in those states had been initially elected to their positions.
Moreover, the authors reported that in three states—Georgia, Minnesota, and North
Dakota—none of the sitting justices were initially elected (Berry and Lisk 2017,
1–2); in three other states—Pennsylvania,5Louisiana, and West Virginia—all of the
3. The use of popular elections for the highest courts in each state, mostly labeled the state “supreme
court,”probably peaked in the first third of the twentieth century when thirty-eight states used such elec-
tions. The role of elections started to decline in the 1930s when California shifted to an appointment system
for appellate judges combined with referendum-style “retention elections”in which voters are asked to vote
yes or no regarding whether a justice should be continued in office (Salyer 2016, 192–94); the referendum-
style election also serves as a method of confirmation for new judges joining the California Court of Appeals
or the California Supreme Court when judges leave at the end of their terms, but for the California Supreme
Court this is rare, having happened only three times since the 1930s (Marvin Baxter in 1990, Tani G.
Cantil-Sakauye in 2010, and Mariano-Florentino Cuéllar in 2014). In 1940 Missouri adopted its “nonparti-
san court plan”(see Peltason 1945), which is commonly called the Missouri Plan, or by its supporters “merit
selection.”In an additional twelve states elections play a role in the selection of some or all lower court
judges; in four of those states, popular elections have either never or not during the period of my study been
used for appellate judges. The decline in the use of elections accelerated starting around 1960 as more states
adopted versions of the Missouri Plan. Three states—Illinois, New Mexico, and Pennsylvania—combine
partisan elections for initial terms with retention elections for subsequent terms. Today, in only twenty-
two states do popular elections continue to play a role in the initial selection of appellate judges, although
another sixteen states use retention elections after initial gubernatorial appointment.
4. As noted below, the BCJ report did not include the Texas Court of Criminal Appeals. My analysis
includes both the Texas Supreme Court and the Texas Court of Criminal Appeals.
5. The authors were incorrect about Pennsylvania because toward the end of the period they consid-
ered, an appointee in Pennsylvania successfully ran in the subsequent election.
2 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

sitting justices were elected. The authors also noted that “over one-third of justices who
were initially appointed were unopposed in their first election, and 29 percent had
never been opposed”(2). They argue that this constitutes a “disconnect in many states
between the formal system of judicial selection and practice. If a state chooses to elect
judges in order to forward certain goals, such as public accountability or democratic
input, the regular use of interim appointments raises questions about whether its system
is, in fact, serving these purposes”(2). Thus, the BCJ suggests important normative
issues concerning how judges are initially selected.
Today, the purpose of electing judges is most frequently described as a method of
enhancing accountability to the electorate (DeBow et al. 2002; Bonneau and Hall
2009,2;O’Malley 2010). However, the theory of political ambition (Schlesinger
1966) tells us that the key to accountability is retention in office, not initial selection
to office. Following from this, there is a body of research providing evidence that state
supreme court justices who must stand in some type of election to retain their seats may
be more responsive to the political preferences of the voters on hot-button issues—such
as criminal justice and abortion—than are justices who do not face the electorate (Brace
and Boyea 2008; Shepherd 2009; Canes-Wrone, Clark, and Park 2012; Canes-Wrone,
Clark, and Kelly 2014). There appear to be little or no such effects on decisions
regarding lower-salience issues (Canes-Wrone, Clark, and Semet 2018). There is also
evidence that state supreme court justices who are retained through partisan or nonpar-
tisan elections are more willing to overturn legislation (and to reverse precedents) than
are judges who face reappointment rather than reelection (Lindquist 2017). One can
debate whether this type of accountability is what one wants in the courts, or what
degree of accountability there should be, but accountability turns on retention rather
than initial selection.
This does not mean that initial selection is unimportant because who is initially
selected can affect court decisions, as is clearly demonstrated by research on decision
making by federal judges who do not face a retention. However, this is not “account-
ability.”There is some evidence of a loyalty effect, with federal appointees more defer-
ential to their appointers than even those appointed by other presidents of the same
party (Howell 2003, 136–74). Loyalty does have its limits, as demonstrated by
Trump appointees’decisions rejecting challenges to the 2020 presidential election.
There is no similar research or evidence concerning loyalty effects among appointed
state judges.
The initial selection process does have the potential to impact the legitimacy of
the courts and the decisions they make. The evidence concerning the direct impact of
the formal selection system on measures of legitimacy is mixed, with some research
showing a difference between election and appointment and some showing no differ-
ence (Kritzer 2015,53–56; Nelson 2017). However, the method of initial selection
could indirectly affect legitimacy through possible effects on demographic representa-
tiveness, what Pitkin (1967) labeled “descriptive representation.”Importantly, what
is relevant here is not the formal system of selection but how the way judges are actually
selected impacts descriptive representation and how that in turn affects the perceived
legitimacy of the court and court decisions. There is evidence that descriptive represen-
tation has effects on legitimacy (Achury et al. forthcoming, but see Overby et al. 2005;
Scherer and Curry 2010; Scherer forthcoming). In my conclusion, I will consider how
Appointed or Elected 3
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

the actual mode of selection of supreme court judges in a state impacts descriptive
representation.
Although my motivation for the analysis presented in the following pages is pri-
marily empirical, I will return to the normative issues related to methods of judicial
selection in my conclusion. My central focus is on what accounts for the extreme vari-
ation across the states with elected supreme courts regarding whether justices are ini-
tially elected or initially appointed to fill midterm vacancies. To that end, I consider a
seventy-five-year period, 1946–2020. Unlike the BCJ study, I omit from my analysis two
states, Louisiana and Arkansas, where appointees are barred by law from running in the
subsequent election, as well as appointees in three other states for certain periods of time
where, for various reasons, appointees either chose not to run (Pennsylvania, 1975–
2014) or were appointed in a system where the formal selection system made initial
appointment the norm (Utah, 1968–1984; New Mexico, post-1988). I discuss these
omissions in more detail in the data section of this article.
Because there have been important political and structural changes over the sev-
enty-five-year period that potentially affect how state supreme court justices initially
come to their positions on elected courts, the first part of my empirical analysis examines
whether there has been a significant shift between initial appointment and initial elec-
tion over the seventy-five-year period. There have been two major political changes.
The first is the demise of the one-party Democratic South with the Republican
Party becoming dominant in statewide politics in most of those formerly Democratic
states. The second change has been the general polarization of politics in the
United States, which is also evident in state supreme court elections. Since at least
1980, state supreme court elections of all types—partisan, semi-partisan (nomination
by party processes but nonpartisan general elections), nonpartisan, and retention—have
become more partisan with the increase most striking in states using ostensibly nonpar-
tisan elections (see Kritzer 2015, 179–200; Kritzer 2018, 412–14; Weinschenk et al.
2020; Kritzer 2021–22).
There have also been two noteworthy structural changes, one major and one
minor, that have implications for whether new judges will be initially appointed or
elected. The major change has been to the formal selection systems and has been largely
a shift away from partisan elections to nonpartisan elections, Missouri Plan systems,6or
in one state—New York—gubernatorial appointment.7Of the thirty-six states that used
some type of popular election for initial selection of their supreme court justices as of
1946, twenty-two states had by 2020 changed their systems, several more than once so
that by 2020, only twenty-two states continued to use popular elections in some way for
the initial selection of justices; the decline was particularly strong regarding partisan
6. The generic Missouri Plan involves all judges being initially appointed by the governor from a list of
nominees prepared by a nominating commission; after an initial term, often shorter than a full term, the
appointee stands in a retention election (a referendum on whether the judge should continue in office) for a
full term. There are many variations around this basic plan, such as including a requirement of some form of
legislative confirmation of appointees. Proponents refer to this system as “merit selection,”but there are
other systems using a nominating commission without retention elections that are also labeled “merit select.”
7. As this is written, two states have shifted back to partisan elections: North Carolina in 2018 from
nonpartisan elections adopted in 2004 (Kritzer 2020,46–52) and Ohio in 2021 from semi-partisan elections
involving nomination in party primaries and nonpartisan general election, a system adopted in the second
decade of the twentieth century (see Auman 1931, 411–15).
4 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

elections, dropping from twenty-one states to only six as of 2020.8These changes are
important because prior research that considered the frequency of initial election in
states with elected supreme courts found that the likelihood of initial election was
greater in partisan election states than in nonpartisan states (Herndon 1962,63–66;
Atkins and Glick 1974, 447; Flango and Ducat 1979,27–28; Dubois 1980,
105–06).9Thus, one might expect to observe an overall decrease in the initial election
of justices over the period I examine.
The second type of structural change has been the adoption of age-based manda-
tory retirement of judges in many states since 1946. At the start of the period I examine,
only four states had mandatory retirement rules for judges. An additional nineteen states
imposed mandatory retirement over the next sixty-five years. As of 2020, twenty-one
states still had mandatory retirement because it ended in two states, Wisconsin by laps-
ing and Illinois where a statute creating mandatory judicial retirement was struck down
as not authorized by the state constitution (Maddux v. Blagojevich, 911 N.E.2d 508 (Ill.
2009)). Mandatory retirement might be expected to increase the incidence of midterm
vacancies and hence reduce the likelihood of initial election of justices.
In addition to the question of what, if anything, has changed over the seventy-five-
year period, there is the question of whether factors other than election type or the
presence of a mandatory retirement rule might help explain the variation in the likeli-
hood of state supreme court justices actually being initially elected in states where the
constitution calls for election. Other factors I consider are region, party competitive-
ness, and state norms. Thus, the analysis that follows examines several questions/
hypotheses:
•How does the likelihood of initial election vary by state?
•Has the percentage of justices initially elected changed over time?
•Does any temporal relationship vary by type of election system used by a state?
•Do the major political changes that have occurred in the South account for any of the tem-
poral changes?
•What, if any, effect is there of a state changing from partisan to nonpartisan (and in North
Carolina, back to partisan) elections?
•What role do rules concerning mandatory retirement or other maximum age limitations play
in whether justices tend to be initially elected or appointed?
•Is the likelihood of initial election related to party competitiveness in a state and is any such
relationship conditioned by election system?
•What is the role, if any, of state norms regarding judicial selection, and how can that be
assessed?
8. From 2004 through 2016, only five states used partisan elections; this was the period during which
North Carolina used nonpartisan elections. The number increased to six in 2018 when North Carolina
restored partisan elections, and then to seven starting in 2022 as a result of Ohio switching from a
semi-partisan system to a fully partisan system for appellate courts. The number of states using nonpartisan
elections was essentially stable, twelve in 1946 and thirteen in 2020, although which states used nonpartisan
elections did change over the period.
9. Several other studies (Winters 1962, 198; Cook 1972, 244; Atkins 1976, 154; Adamany and Dubois
1976, 738; Bowers 1990, 6; Sheldon and Maule 1997, 21n13; Holmes and Emrey 2006, 6), most dealing
with a single state, noted the high incidence of appointment rather than election; some of these studies
included lower court judges as well as state supreme court judges.
Appointed or Elected 5
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

The remainder of this article consists of six sections. Section II uses a critical discussion
of the BCJ study to identify some complexities in looking at the initial election versus
initial appointment issue. Section III discusses the data, explaining what is and is not
included. Section IV opens with an overview of the variation across the states and then
examines temporal patterns and whether temporal patterns vary by election type or
region (South versus non-South); included in this section is an analysis of whether pat-
terns change when a state shifts from partisan to nonpartisan election. Section V con-
siders two other explanations for variation across states: mandatory retirement and party
competitiveness. Section VI considers the idea of state norms as a partial explanation
and presents several analyses assessing the role of such norms. Section VII concludes
with a brief summary and revisits the normative issues inherent in the initial selection
of judges.
THE BRENNAN CENTER STUDY
The BCJ study provided a snapshot of how state supreme court justices sitting as of
August 2016 obtained their positions in the twenty-two states where the state consti-
tution calls for the election of the members of the state’s highest court. As noted pre-
viously, the study found that only 54.9 percent of the 153 justices were elected (Berry
and Lisk 2017, 4). As noted in the Introduction, the BCJ reported that the percentage
elected varies from 0 in three states to 100 in three other states; the mean and median
percentage elected were 54.9 and 57.1 respectively.
With one minor exception, the BCJ study did not look at any of the factors that
might explain the likelihood of election rather than appointment. The one exception
was to point out that in Louisiana, where all seven justices sitting in August 2016 were
elected, an interim appointee may not run in the subsequent election to fill the seat to
which the interim justice had been appointed (Louisiana Constitution of 1974, Article
V, §22(B)). The authors of the study failed to note that the Louisiana Constitution also
calls for the governor to call a special election to fill interim vacancies to be held within
twelve months unless there is less than twelve months remaining in the departing jus-
tice’s term (Louisiana Constitution of 1974, Article V, §22(B)); in practice, the special
elections occur within just a few months. Thus, at any point in time, it is unlikely that
there would be an appointed interim justice sitting on the Louisiana Supreme Court.
The BCJ authors also failed to note that the Arkansas Constitution also prohibits
an interim appointee from running in the subsequent election for the seat to which the
justice had been appointed (Arkansas Constitution of 1874, Amendment 29, §2).
However, unlike in Louisiana, in Arkansas the governor does not call a special election,
and the interim appointee serves out the term of the departing justice. In August 2016,
there was one interim appointee sitting on the Arkansas Supreme Court, Chief Justice
Howard Brill, who had been appointed in August 2015 to complete the term of retiring
Chief Justice Jim Hannah;10 in November 2016 John “Dan”Kemp was elected to that
position, taking office on January 1, 2017.
10. See “Howard Brill.”Ballotpedia. Accessed July 5, 2021. https://ballotpedia.org/Howard_Brill.
6 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

There were two other quirks in judicial selection processes in operation in 2016
that the BCJ authors failed to consider. Regarding Pennsylvania, as previously noted, for
a period of about forty years ending in 2015, there was a norm that interim appointees
refrain from running in the subsequent election; the appointee who broke the norm by
running in 2015 lost in the subsequent Democratic primary. The first appointee since
1969 to run successfully was appointed in June 2016 (see Mendicino 2016). The second
quirk concerns New Mexico, where the report shows only one of the five justices having
been elected. The system in New Mexico calls for the governor to make appointments
from a list prepared by a nominating commission for all vacancies on the state supreme
court, including those occurring when justices choose not to run for reelection and
leave office at the end of their terms; appointees can then run in a partisan election
held at the time of the next general election. The only elected justice on the New
Mexico Supreme Court in 2016 was Democrat Barbara Vigil; in 2012, she had defeated
the interim justice, who had been appointed by the Republican governor less than two
months before the election. Thus in New Mexico, the system is designed with a pre-
sumption that justices will be appointed while allowing for the possibility of election.11
If one omits Louisiana, Arkansas, Pennsylvania, and New Mexico from the calcu-
lation in the BCJ report, the percentage of justices initially elected drops from 54.9 to
50.8 percent. Thus, just barely over half of justices sitting in August 2016 on state
supreme courts where there was no effective presumption for or against appointment
were initially elected to their seats. One might further adjust these figures by including
the Texas Court of Criminal Appeals, which is the final court for criminal cases in
Texas but was not included in the BCJ report. As of August 2016, eight of the nine
judges on that court had been initially elected; with those nine judges included, the
percentage of justices and judges of state high courts initially elected goes back up
to 53.3 percent, very close to the figure reported in the BCJ report.
This discussion of the BCJ report may seem like nitpicking, particularly given that
my final adjustment produces a figure virtually identical to that in the report. However,
the point here is that one must take care in assessing the frequency of appointment of
high court judges in states with elected judiciary because the rules governing selection
and how those rules affect which states are included in the calculation play an impor-
tant role in the percentages one obtains.
DATA
To examine how justices of state supreme courts initially come to their positions, I
relied on data I have assembled on state supreme court elections between 1946 and
2020 to identify the justices taking the bench during this period and how they obtained
their positions, initial election or initial appointment.12 For each of those elected, the
data indicated whether the seat was open, was being contested by an incumbent who
11. The system in New Mexico was adopted after voters rejected a full Missouri Plan system (see
Kritzer 2020, 187–96); the current system is something of a hybrid between a Missouri Plan system and
an election system.
12. These data are archived and publicly available on Dataverse; they can be accessed and downloaded
at https://dataverse.harvard.edu/dataset.xhtml?persistentId =doi:10.7910/DVN/1P1JFG.
Appointed or Elected 7
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

had been appointed since the last election, or was being contested by an incumbent
who had been previously elected to the seat. New justices include the winners of all
open seats, all justices who had been appointed since the last elections and sought, suc-
cessfully or unsuccessfully, to retain their positions, and all justices who defeated an
incumbent regardless of whether that incumbent had been previously elected to the
seat or had been recently appointed and was standing for election for the first time.
Thus, I excluded from the analysis justices appointed on a strictly interim basis, by
law, custom, or personal choice; this means I did not include persons appointed to mid-
term vacancies who then did not run in the subsequent election for the position to
which they had been appointed.
Consequently, I omitted new justices in five states for some or all of the period
examined:
•Arkansas and Louisiana, entire period: as noted previously, the constitutions of those states
prohibit an appointed justice from running to succeed him- or herself.13
•Pennsylvania, 1975–2014: after the state adopted a system using partisan elections for initial
terms and retention elections for subsequent terms, a norm, noted previously, came to exist
that interim appointees refrain from running in the subsequent election; this norm ended in
2015 when appointee Correale Stevens ran unsuccessfully to retain the seat to which he had
been appointed in 2013.14
•New Mexico, 1989–2020: in 1988 voters approved a unique system for filling vacancies on
the state supreme court that is a compromise between a pure Missouri Plan appointment sys-
tem and a partisan election system (see Kritzer 2020, 191–96). As discussed previously, under
that system, regardless of whether a vacancy occurs during a term or at the end of the depart-
ing justice’s term, the governor appoints a new justice from a list of nominees prepared by a
nominating commission; the appointee can then run in a partisan election at the time of the
next even-year general election. Once a justice has won a partisan election, the justice can
seek additional terms by standing in retention elections.15
•Utah, 1969–1984: during that period, legislation provided that every vacancy, whether
interim or at the end of the departing justice’s term, was to be filled by appointment; all
incumbents including recent appointees could be challenged in nonpartisan elections, but
13. Arkansas Constitution of 1874 Amendment 29, § 2; Louisiana Constitution of 1974, Article V,
§22(B). The prohibition in Louisiana was added in explicit terms to the Louisiana Constitution only in 1973
(effective 1974). However, in Louisiana justices of the state supreme court are elected by district and for fifty
years or more prior to 1973 interim vacancies were temporarily filled by the Supreme Court appointing a
Court of Appeals judge who did not reside in the district where the vacancy occurred to temporarily fill the
vacancy, with a special election to be held within four months if more than two years remained of the unex-
pired term (Louisiana Constitution of 1921, Article III, § 7); the requirement that the temporary appointee
not reside in the district of the vacancy served to preclude the appointee from running for the position.
14. Stevens came in third in his party primary. Subsequently, Sallie Mundy ran successfully in 2017 to
retain the seat to which she had been appointed in 2016.
15. Between 1990 and 2020, fourteen vacancies occurred on the New Mexico Supreme Court.
Because of a quirk in the system, two of the partisan elections that followed those vacancies did not include
an appointee. Both involved vacancies occurring after the party primaries in which case it falls to the party
executive committee to make the nomination; in one case the appointee’s political party chose not to nomi-
nate the person the governor had appointed (see Kritzer 2020, 198). Of the remaining twelve vacancies,
eight appointees won the subsequent partisan election and four lost, one in the party primary. The net result
was that sixteen new justices were appointed to the New Mexico Supreme Court between 1988 and 2020,
and only six new justices were initially elected, four by defeating appointees and two running for seats not
contested by an appointee.
8 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

no challengers defeated an appointee between 1969 and 1984,16 after which Utah adopted a
full Missouri Plan system (see Kritzer 2020, 169–78).
Omitting the five states as specified above and those interim appointees who did not run
in the subsequent election,17 1,120 new justices joined state high courts over the sev-
enty-five-year period, close to half (568, 50.7 percent) of whom were initially elected.
There are three ways that justices can be initially elected: winning open-seat elections,
defeating recently appointed incumbents (i.e., incumbents standing in their first elec-
tion since being appointed), or defeating previously elected incumbents. About two-
thirds of those initially elected (67.4 percent) won open-seat elections, with the remain-
der evenly split between defeating a recently appointed incumbent (16.6 percent) and
defeating a previously elected incumbent (16.0 percent). Preliminary analyses showed
that this split varied little over time with the possible exception of the 1971–1980
decade when 75.9 percent of new justices won open-seat elections and only 4.6 percent
defeated previously elected incumbents.
CHANGE OVER TIME, REGION, AND ELECTION TYPE
Overview
Figure 1a shows the percentage of justices elected in each state that used elections
for some or all of the period 1946–2020. Similar to the BCJ report, that percentage
ranged from 0 to 100, although the one hundred in Nebraska is based on only four
new justices who joined the Nebraska Supreme Court between 1946 and 1962, after
which the state switched to the Missouri Plan.18 Figure 1b shows that the range is simi-
lar for partisan, semi-partisan, and nonpartisan election systems; the outlying values for
Utah reflect that there were only two new justices—both elected—between 1946 and
1950 when partisan elections were used and only five—all appointed—between 1952
and 1968 when nonpartisan elections were used. Although the range is similar for par-
tisan and nonpartisan election states, partisan states tend to be in the upper part of the
figure and nonpartisan states in the lower part. Overall, in partisan states, 58.0 percent
of new justices were initially elected, in semi-partisan states 52.9 percent, and in non-
partisan states 40.4 percent.19 These differences are consistent with the prior research
noted in the Introduction.
16. In fact, only one of the six appointees during this period faced a challenge; the other appointees,
and unchallenged elected incumbents, stood in retention elections. One elected incumbent was defeated for
reelection.
17. I counted 105 interim appointees who chose not to run in the subsequent election, and hence that
election became an open-seat contest.
18. Three other states, all of which changed to a Missouri Plan during the period studied, had fewer
than ten new justices join their state supreme courts when popular elections were the formal method of
selection: Iowa (7), Kansas (7), Wyoming (6). The highest number of new justices was Texas at 115,
but this includes both the Texas Supreme Court and the Texas Court of Criminal Appeals.
19. Chi-square =30.87, p<.001.
Appointed or Elected 9
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

Variation over Time
Figure 2 shows by decade the percentage of new justices initially elected. As the
figure shows, that percentage has varied between 42.5 percent to 59.8 percent, although
that latter figure covers only the period 1946–1950. There is no overall trend. Leaving
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
MD (16)
MN (46)
GA (39)
SD (15)
UT (7)
ID (32)
FL (18)
ND (26)
OR (42)
TN (21)
NC (66)
MS (55)
WA (55)
WI (41)
NM (30)
WY (6)
MI (49)
KS (7)
PA (22)
NV (32)
MT (42)
WV (35)
TX (115)
AL (54)
OH (57)
OK (20)
AZ (16)
CO (17)
IA (7)
IN (15)
NY (19)
KY (55)
IL (39)
NE (4)
All Election Types
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
MD (16)
MN (46)
GA (23,16)
SD (15)
UT (2,5)
ID (32)
FL (14,4)
ND (26)
OR (42)
TN (21)
NC (56,10)
MS (35,20)
WA (55)
WI (41)
NM (30)
WY (6)
MI (49)
KS (7)
PA (22)
NV (32)
MT (42)
WV (30,5)
TX (115)
AL (54)
OH (57)
OK (20)
AZ (16)
CO (17)
IA (7)
IN (15)
NY (19)
KY (22,33)
IL (39)
NE (4)
By Election Type
Partisan Semi-Partisan Nonpartisan
Numbers next to state abbreviations are numbers of observations; if two numbers,
first is for partisan elections, second for nonpartisan elections.
(a)
(b)
FIGURE 1.
Percentage Elected for Each State Using Elections.
10 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

aside that first half decade, the percentage initially elected increases from 42.9 percent
for 1951–1960 to 57.5 percent for 1991–2000 before dropping back to 44.8 percent.
The increase from 1951 through 2000 might relate to some of the changes noted in
the Introduction, but those would not explain the drop-off over the last two periods
shown in the figure.
Does any of the variation in Figure 2 reflect changes in state selection systems
given the differences in likelihood of initial election by type of election noted pre-
viously? Figure 3 shows how these figures vary over time, but again there are no con-
sistent trends. In some decades, the percentage initially elected is highest in the
semi-partisan election states. In every decade, the percentage initially elected is least
for nonpartisan states, particularly in the 1951–1960 decade; moreover, without that
decade, percentage elected would vary the least in nonpartisan election states.
There is no clear difference over time between partisan and semi-partisan states,
with a greater percentage elected in partisan states in some decades and a greater
percentage in semi-partisan states in others. Given the similarity in the patterns
for partisan and semi-partisan states, I combine those two groups of states in many
oftheanalysesthatfollow.
Region
As noted in the Introduction, one of the major political changes in the United
States over the period under consideration was the shift from one-party Democratic
control in southern states to the dominance of the Republican Party in those states.
20%
30%
40%
50%
60%
70%
1946-
1950
1951-
1960
1961-
1970
1971-
1980
1981-
1990
1991-
2000
2001-
2010
2011-
2020
FIGURE 2.
Percentage of New Justices Elected by Decade, 1946–2020.
Appointed or Elected 11
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

In my study of changes in state supreme court elections between 1946 and 2012 (Kritzer
2015), I found that changes in election patterns were most pronounced in the southern
states. Is there a difference between southern and nonsouthern states in the percentage
of justices initially elected?20 Looking at the entire period 1946–2020 the answer is
not much: 47.6 percent elected in the South versus 52.3 percent in the non-
South.21 Figure 4a shows the comparison across the decades. Very clearly, there is a
convergence, with the South differing from the non-South in the first three periods
but differing little after 1970. Figure 4b extends the analysis of regions by adding
the type of election, collapsing semi-partisan and partisan elections because no southern
state employs, or previously employed, semi-partisan elections.22 As the figure shows,
there were no nonpartisan elections in the South until the 1970s.23 Although the dif-
ference between the South and non-South generally declines over time, election is
more likely in non-South states for every decade for both partisan and nonpartisan
elections with the single exception of the 1991–2000 decade.24
20%
30%
40%
50%
60%
70%
1946-
1950
1951-
1960
1961-
1970
1971-
1980
1981-
1990
1991-
2000
2001-
2010
2011-
2020
Partisan Semi-Partisan Nonpartisan
FIGURE 3.
Percentage Elected by Decade and Election Type.
20. I do not include border states, only the nine states of the Confederacy that used popular elections
for some or all of the period under study.
21. Chi-square =2.19, p =.139.
22. The figure looks very similar if rather than semi-partisan and partisan election states being com-
bined, semi-partisan election states are omitted.
23. Florida had nonpartisan elections for a few years in the 1970s before adopting a Missouri Plan
system for appellate judges. Georgia adopted nonpartisan elections in the 1980s, but all new justices during
that decade, including the three joining the bench after the switch to nonpartisan elections, were appointed.
24. If I omit semi-partisan election states rather than combining them with partisan election states, the
single exception disappears.
12 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

Impact of Changing Type of Election
Figure 1b included two symbols for the seven states that switched from partisan to
nonpartisan (and in North Carolina, back to partisan) between 1946 and 2020. With
the exception of Utah, where there were only two new justices during the period of
partisan elections and five new justices during the nonpartisan period, there was not
a lot of difference in the likelihood of initial election when a state changed. This is
20%
30%
40%
50%
60%
70%
80%
1946-
1950
1951-
1960
1961-
1970
1971-
1970
1981-
1990
1991-
2000
2001-
2010
2011-
2020
Non-South South
By Region Only
20%
30%
40%
50%
60%
70%
80%
1946-
1950
1951-
1960
1961-
1970
1971-
1970
1981-
1990
1991-
2000
2001-
2010
2011-
2020
Non-South Partisan & Semi-Partisan Non-South Nonpartisan
South Partisan South Nonpartisan
By Election Type and Region
(a)
(b)
FIGURE 4.
Percentage Elected by Decade and Region.
Appointed or Elected 13
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

clear in Table 1, which shows the percentage elected under each election system for the
six states other than Utah. In no state is there a statistically significant decrease in the
percentage initially elected when a state switched to nonpartisan elections.25 In fact, a
higher percentage of justices were elected in North Carolina during the period (2004–
2016) the state used nonpartisan elections.
FURTHER EXPLANATIONS: MANDATORY RETIREMENT AND
PARTY COMPETITION
The substantial variation among the states, even after accounting for type of elec-
tion system and region, raises the question of whether any other state or situational
characteristics might explain whether a new justice joins a state’s supreme court by elec-
tion or appointment. The Introduction noted several possible variables beyond election
system. This section considers two of those variables: mandatory retirement rules and
party competitiveness.
TABLE 1.
Percentage Elected in States that Changed from Partisan to Nonpartisan Elections
Percentage Elected Election Type
Partisan Nonpartisan
Florida 35.7 (14)b25.0 (4)
Georgia 17.4 (23) 6.3 (16)
Kentucky 77.3 (22) 75.8 (33)
Mississippi 40.0 (35) 40.0 (20)
North Carolinaa37.5 (56) 50.0 (10)
West Virginia 63.3 (30) 40.0 (5)
All 44.4 (180) 47.7 (88)
aNorth Carolina switched to nonpartisan elections effective 2004 and then back to partisan elections
effective 2018.
bThe numbers in parentheses are the numbers on which the percentages are based.
25. Chi-square tests for the six states and overall never produced a p-value less than .30. This was not
necessarily the case in earlier periods. For example, according to my data, in the period prior to the adoption
of nonpartisan election and omitting the first three justices elected in 1857 after the Minnesota Supreme
Court was created, fourteen of twenty (70 percent) of justices were initially elected, compared to only six of
twenty-two (30 percent) in the pre-1946 period after judicial elections became nonpartisan starting in 1912
(chi-square =11.32, p<.001). Bowers (1990, 6) reports that in Nevada, fifteen of twenty (75.0 percent)
justices joining the Nevada Supreme Court between 1964 and 1915 were elected compared to nine of
twenty-one (42.9 percent) between 1916 (when elections became nonpartisan) and 1988 (chi-square
=4.36, p =.036).
14 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

Mandatory Retirement26
Over the period of this study, twenty-four states had mandatory retirement or other
maximum age rules for some part of the period.27 Only four states had such rules as of
1946; twenty states imposed mandatory retirement or age limits after 1946 with fifteen
doing so between 1950 and 1974; as of 2020, twenty-two states had rules mandating
judicial retirement or limiting maximum age.28 The maximum age falls between seventy
and seventy-five for all states using elections,29 with fourteen states currently setting the
age at seventy, six at seventy-five, three at seventy-two, and one at seventy-four.30 Of
more importance than the specific age is the range of age limitation rules that exist,
which differ in their implication for whether new justices would tend to be appointed
rather than elected. The variants are:
•Must retire on reaching mandatory retirement age or shortly thereafter—for example, at the
end of the month (“immediate mandatory retirement”).
•Must retire at the end of the court term or end of the year in which the justice reaches the
mandatory retirement age (“delayed mandatory retirement”).
•Must retire at end of the term of office (i.e., “ineligible for reelection due to age”) during
which the justice reaches retirement age.
•Other: (a) must retire at retirement age but may complete the term of office if less than 50
percent of the current term remains when retirement age is reached (Florida), or (b) must
retire at the end of the court term unless the justice reaches retirement age during the first
four years of the justice’s term, in which case the justice must retire at the end of the fourth
year of the term (Texas).
The first two, immediate mandatory retirement and delayed mandatory retirement, are
most likely to create a vacancy to be filled by appointment. Under the third variant,
ineligible for reelection, there is no reason to expect an increase in the likelihood of a
vacancy to be filled by appointment; rather, it might increase the likelihood of open-
seat elections due to justices choosing to serve out their final terms. The “other”variants
would fall somewhere between the ineligible for reelection variant and the two date-
certain variants.
Column aof Table 2 shows the percentage of new justices initially elected when
there is no mandatory retirement and under each of the mandatory retirement regimes.
26. I include in the discussion of mandatory retirement laws that specify a maximum age at which a
person can run for the state supreme court, which bar incumbents from seeking reelection once over that
maximum age.
27. The information on mandatory retirement comes from several sources; the two most extensive are
Ballotpedia (https://ballotpedia.org/Mandatory_retirement) and the National Center for State Courts
(https://www.ncsc.org/information-and-resources/trending-topics/trending-topics-landing-pg/mandatory-
judicial-retirement).
28. Wisconsin had a mandatory retirement age of seventy in effect 1955–1978 and Illinois originally
had a mandatory retirement age of seventy, later raised to seventy-five, from 1965 until 2009, when it was
struck down by the Illinois Supreme Court as not authorized by the state constitution (Maddux v.
Blagojevich, 911 N.E.2d 508 (Ill. 2009)).
29. In Vermont, which does not use elections, the mandatory retirement age is ninety (Vermont
Statutes, Title IV, Chapter 15, § 609(a)).
30. Those are the mandatory retirement ages as of 2020. A few states increased their mandatory retire-
ment ages during the period of this study, but most proposals to do so failed (see Rafferty 2015, 6).
Appointed or Elected 15
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

Overall, justices are least likely to be elected in states requiring immediate retirement.
However, justices in states with no mandatory retirement are no more likely to have
been initially elected than justices in states with delayed mandatory retirement or jus-
tices in states where justices become ineligible for reelection due to age. Actually, a
greater percentage of justices were initially elected where justices become ineligible
for reelection due to age than where there was no mandatory retirement or where there
was delayed mandatory retirement.31
Columns band cof Table 2 show what initially appear to be interesting variations
after controlling for type of election system. In states using partisan and semi-partisan
elections, the percentage of justices initially elected was the same in states without man-
datory retirement as in states with immediate mandatory retirement; a higher percent-
age were initially elected in states with other types of mandatory retirement rules,
particularly where there was delayed mandatory retirement. There are effectively only
three mandatory retirement regimes in nonpartisan election states: no mandatory retire-
ment, immediate mandatory retirement, and delayed mandatory retirement. There is no
difference in the likelihood of having been elected for justices in states with no man-
datory retirement and states with delayed mandatory retirement, but sharply fewer jus-
tices were initially elected in nonpartisan election states with immediate mandatory
retirement rules.
TABLE 2.
Likelihood of Initial Election by Mandatory Retirement and Election Type
Percentage Initially Elected
All Election States Omitting Minnesota
Mandatory Retirement
Regime
All
Partisan &
Semi-Partisan Nonpartisan All Nonpartisan
(a) (b) (c) (d) (e)
No mandatory
retirement
49.0% (649)a53.1% (356) 44.0% (293) 49.7% (636) 45.4% (280)
Immediate
mandatory
retirement
40.3 (139) 53.1 (81) 22.4 (58) 51.9 (106) 48.0 (25)
Delayed mandatory
retirement
55.1 (138) 71.2 (59) 43.0 (79) 55.1 (138) 46.0 (9)
Ineligible for
reelection due to
age
58.5 (94) 58.5 (94) –(0) 58.5 (94) –(0)
Other 63.0 (100) 64.6 (96) (1) (4b) 63.0 (100) (1) (4b)
Chi-square 16.19 (4 df)
(p =.003)
9.71 (4 df)
(p =.042)
9.96 (3 df)
(p =.019)
8.18 (4 df)
(p =.085)
0.875 (3 df)
(p =.832)
aThe numbers in parentheses are the numbers on which the percentages are computed.
b“Other”includes only Florida and Texas; Florida briefly used nonpartisan elections during part of the
1970s.
31. The chi-square recomputed omitting the “Immediate”category remains statistically significant
(chi-square 9.30, df =3, p =.026).
16 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

The latter pattern must be treated with caution because a single state, Minnesota,
which adopted mandatory retirement effective in 1974, is driving much of what is in
column cof Table 2. Of the forty-three new justices in nonpartisan election states with
mandatory retirement, thirty-three came from Minnesota and thirty-two of the new
Minnesota justices were initially appointed. As shown in column eof Table 2, omitting
Minnesota justices, the 22.4 percent elected in nonpartisan states with immediate
retirement becomes 48.0 percent, and the chi-square drops to a nonsignificant
0.872. Moreover, omitting Minnesota and ignoring election type, the percentage
elected in states with immediate retirement increases to 51.9, which is actually higher
than the 49.7 percent in states without mandatory retirement.32 Thus, after omitting
Minnesota, what appeared to be an interesting pattern disappears.
Although it seems logical that mandatory retirement would be at least a partial
explanation for state supreme court justices often being initially appointed rather than being
initially elected, that does not appear to be the case. Likelihood of initial appointment is
lower under some types of mandatory retirement rules than when there is no mandatory
retirement. Moreover, there does not appear to be a sharp difference between no mandatory
retirement and the immediate form of mandatory retirement that one might expect to have
the greatest effect once one considers the pattern in Minnesota that long predates that
state’s adoption of an immediate mandatory retirement rule.33
One reason that mandatory retirement may not have the effect hypothesized is that
the opportunity for appointments may arise frequently in states without mandatory retire-
ment due to death or serious health conditions. There were 331 appointments in states
without a mandatory retirement rule. The opportunity for appointment in 116 (35.0 per-
cent) of those arose due to the death (104) or illness (12) of the appointee’s predecessor.34
Party Competitiveness
A second potential covariate is party competitiveness in a state. One might
hypothesize that justices’decisions whether to step down prior to when they will have
to stand for election will be affected by who they want to choose their replacement, the
governor or the voters, and party competitiveness could easily play a role in that choice.
In some states, party competitiveness is strong and in others it is weak. The extreme
version of weak party competitiveness was the overwhelming dominance of the
Democratic Party in the southern states during the early years being examined (see
Ranney 1965,65–67).
32. The chi-square drops to 8.18 (p =.085) but the lingering marginal statistical significance does not
relate to a decrease in initial elections in the presence of the kinds of mandatory retirement rules that would
be expected to produce such an increase.
33. Even prior to the imposition of a mandatory retirement rule, most new justices in Minnesota were
initially appointed: twenty-five of thirty-three between 1912 (when Minnesota started using nonpartisan
judicial elections) and 1972. Three of those initially elected were elected during World War II (all to open
seats) and three others during the first four years of nonpartisan judicial elections (one to an open seat and
two defeated recent appointees). Between 1946 and 2020, there were two open-seat elections, one before
and one after the adoption of mandatory retirement, and in one other election a recently appointed incum-
bent was defeated.
34. Information on the reasons for the vacancies came from news coverage of the appointments; two
databases were used for these searches: https://www.newspapers.com and https://www.newsbank.com.
Appointed or Elected 17
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

To assess the impact of party competitiveness I adapted a version of an index of
party dominance created by Austin Ranney (1965,63–64).35 Ranney’s measure com-
bines the partisan divisions in the two chambers of the state legislature, the most recent
partisan division in the vote for governor, and whether one party controls the gover-
norship and both chambers of the legislature. From these indicators, a running average
is created. The resulting index runs from 0 for extreme Republican control to 1 for
extreme Democratic control; .5 constitutes maximum competitiveness. For my analysis
I created a competitiveness measure by “folding”the four-year running average version
of the Ranney index at .5, and then algebraically adjusting the result to run from zero to
ten, with zero meaning no competition (i.e., the equivalent of either 0 or 1 on the origi-
nal scale) and ten meaning maximum competition (.5 on the original scale). The actual
values range from 0 to 9.99 with a mean of 6.53 (median 6.98) and a standard deviation
of 2.60. Because the dependent variable is dichotomous, election versus appointment,
logistic regression is an appropriate statistical methodology.
A series of logistic regression models, see Table A1 in the online appendix, showed
that party competitiveness had an effect, but only where partisan (including semi-par-
tisan) elections were used. Table 3 provides an easy way to see the effect of competi-
tiveness in partisan/semi-partisan election states and the lack of effect in nonpartisan
election states. I split the data at the three observed quartiles of party competitiveness
(5.07, 6.98, 8.77) and determined the observed percentage of new justices elected in
each of the four quartiles, both across election types and within the two election-type
categories.36 As column aof the table shows, ignoring the type of election used, there is
an inconsistent tendency for the percentage of justices elected to increase across the
quartiles. For nonpartisan election states (column c), there is no pattern, and three
of the four quartiles show virtually the same likelihood of election. There is a strong,
clear pattern of increase in partisan/semi-partisan election states (column b), with 43.4
percent of justices elected in the least competitive quartile, rising to 71.8 percent in the
most competitive.
This suggests another question: is the pattern shown in Table 3 attributable to the
frequency of extreme one-party dominance in the southern states? Columns dand eof
Table 3 show the pattern for partisan and semi-partisan states separately for the South
and non-South. There is a strongly significant relationship for the South and at best a
very marginal relationship for the non-South.37 This is largely confirmed by another set
of logistic regression models included in the online appendix (see Table A2 and
35. Most of the data concerning the Ranney Index came from the Correlates of State Policy Project
dataset (http://ippsr.msu.edu/public-policy/correlates-state-policy); for the last few years, not included in that
dataset, I computed the index based on information compiled from various sources such as the Book of States
and the Congressional Quarterly elections database.
36. The distribution among the overall quartiles differs for partisan/semi-partisan states and nonparti-
san states:
Q1 Q2 Q3 Q4
partisan/semi-partisan 33.2% 27.8% 23.9% 15.0%
nonpartisan 12.0 20.2 27.2 40.6
37. There is no statistically significant relationship between competitiveness and likelihood of initial
election in nonpartisan elections in either the South or the non-South, although there is a pattern of
increase in the South as one goes from the lowest quartile of competitiveness to the highest, which would
have been statistically significant if there were substantially more cases (there were only seven observations
in the lowest quartile and only six in the highest).
18 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

accompanying discussion). The one difference that appears in the logistic regression is
that when state fixed effects are included, competitiveness is statistically significant in
partisan election states in both the South and the non-South, and the coefficient for the
non-South is larger (but not to a statistically significant degree) than the coefficient for
the South. From this it appears that the relationship between the likelihood of initial
election and party competitiveness appears to be largely a southern phenomenon, par-
ticularly in the context of partisan elections. Perhaps of more importance is that the set
of state fixed effects is statistically significant.38 This indicates that there may be system-
atic characteristics of the states that have not yet been accounted for.
STATE NORMS
What might be hidden in the state fixed effects included in the logistic models
examining competitiveness? Recall that despite the broad pattern of a lower likelihood
of initial election in states using nonpartisan elections, there were no statistically sig-
nificant changes in that likelihood in the states that switched from partisan to nonpar-
tisan elections during the period under study. This suggests that there are probably state
norms that persist despite this shift. In this section I explore several ways of assessing the
role of state norms:
•The method of selection of a justice’s predecessor: are those whose predecessor was elected
more likely to be elected?
•Appointees’likelihood of facing opposition at their first elections: does the likelihood of an
appointee being opposed increase as the percentage of justices initially elected to open seats
increases?
•Appointees’success in standing for reelection: are appointees less successful in their first elec-
tion as the percentage of justices initially elected to open seats increases?
TABLE 3.
Likelihood of Election by Quartiles of Competitiveness
Percentage Elected
Quartiles
All
Partisan & Semi-
Partisan Nonpartisan
Partisan/Semi-partisan only
South Non-South
(a) (b) (c) (d) (e)
Q1 (least competitive) 43.4% (279) 43.4% (228) 43.1% (51) 36.7% (166) 61.3% (62)
Q2 50.9 (277) 58.1 (191) 34.9 (86) 62.3 (77) 55.3 (114)
Q3 55.7 (280) 65.2 (164) 42.2 (116) 67.9 (56) 63.9 (108)
Q4 (most competitive) 52.5 (276) 71.8 (103) 41.0 (173) 68.4 (19) 72.6 (84)
Chi-square (3 df) 9.19
(p =.027)
31.06 (p<.001) 1.44
(p =.697)
26.06
(p<.001)
6.35
(p =.096)
38. Chi-square =85.53 (30 df), p<.001.
Appointed or Elected 19
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

•The role of state political culture: is there a relationship between the method of initial selec-
tion in one or more measures of state political culture?
Method of Predecessor’s Selection39
A possible indicator of the influence of state norms is the manner of selection of a
justice’s predecessor (ignoring any strictly interim appointees).40 Table 4 shows that
new justices whose predecessor was elected were 12–13 percentage points more likely
to have been elected themselves than if their predecessor was appointed. Logistic regres-
sions reported in the online appendix (Table A3) show that method of the predecessor’s
selection continues to have an effect after controlling for election type and party
competitiveness.
Appointees’First Elections
Likelihood Appointees Will Be Opposed at Their First Election
In a state where the norm is that state supreme court justices should be elected one
could expect that appointees will be more likely to face opposition at their first primary
and/or general election than in states where the norm is that justices will initially be
appointed. Where there is such a norm, justices will be more likely to step down at the
end of their terms so that their successors can be chosen by the electorate. A possible
indicator of a norm that justices should be elected is the percentage of justices initially
elected to open seats (PJIEOS) between 1946 and 2020. The circles in Figure 5a show
the percentage of appointees who were opposed in their first election (circles) in each of
the twenty-eight states where there were five or more appointees; the figure also shows
PJIEOS for those twenty-eight states (squares). The circles and squares tend to increase
together (i.e., move to the right together in the figure). Figure 5b shows a traditional
scatterplot and regression line predicting the probability of appointees being opposed in
their first election by PJIEOS. The regression indicates that PJIEOS accounts for more
than a third (r2=.374, p<.001) of the variation in the percentage of appointed justices
facing opposition at their first election. Moreover, for each one-point increase in
PJIEOS, the percentage of appointees facing opposition increases 0.84 percentage
points—not far from a one-to-one relationship.
One limitation of this simple aggregate analysis is that it does not account for other
factors that might be important. For example, in a study of nonsouthern state supreme
court elections between 1948 and 1974, Dubois (1980, 141) found that appointees were
more likely to be opposed in partisan and semi-partisan elections (86.6 percent) than in
nonpartisan elections (55.0 percent).41 The same is true for the period and states
39. There were twenty-eight new state supreme court positions created during the period under study;
the justices who initially filled those positions are omitted from the analysis reported in this section.
40. In the case of those running for an open seat where an interim appointee chose not to run, I used
the method of selection of the justice whom the interim appointee replaced.
41. These are my calculations based in part on information from Dubois (1978, 267).
20 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

included in this analysis, although the gap is not as large: 64.1 percent for partisan and
semi-partisan elections and 52.9 percent for nonpartisan elections. The lower number
for partisan and semi-partisan elections is largely due to partisan elections in the South
(where there have been no semi-partisan elections);42 only 48.0 percent of appointed
incumbents in partisan elections in the South faced opposition when they stood for
their first election compared to 87.3 percent outside the South.
There are a variety of other factors that one could hypothesize would affect the
likelihood that appointees would face opposition at the first election following their
appointment. As previously discussed, that likelihood might be expected to be related
to the political competitiveness in the state, increasing as competitiveness increased. It
might be expected to increase as the term length increased because those interested in a
seat on the court would have to wait longer for an opportunity to run as the term length
increased. Finally, one could also expect that the longer the time from the initial
appointment until the election, the more established the appointee would become
as a justice, which might then decrease the likelihood of opposition; this variable is
scaled from one to six with one indicating that the election was occurring the same
year as the justice’s appointment and six that the election was occurring during the sixth
year since the justice’s appointment. Based on results so far, one could also expect that
competitiveness would be conditioned on partisan (including semi-partisan) versus
nonpartisan elections, and the same might be true for the effect of PJIEOS as a predictor
of opposition.
Table A4 in the online appendix reports the results of a probit analysis, which is
similar to logistic regression except for the scaling of the coefficients.43 I used probit
analysis because it will be needed for the analysis of electoral success of appointees dis-
cussed in the next section.44 The results show a significant effect for PJIEOS in states
TABLE 4.
Effect of How Predecessor Was Selected
All Elections
Partisan & Semi-Partisan
Elections Nonpartisan Elections
Predecessor was Predecessor was Predecessor was
Elected Appointed Elected Appointed Elected Appointed
Percentage
Elected (n)
57.9 (539) 44.7 (553) 63.0 (357) 51.8 (311) 47.8 (182) 35.5 (242)
Chi-square (1 df) 19.09 (p<.0001) 8.63 (p =.0003) 6.47 (p =.0011)
42. Dubois reported 88.2 percent for partisan and 84.2 percent for semi-partisan; I find 60.9 (n =230)
percent for partisan and 75.8 (n =65) percent for semi-partisan.
43. Logistic regression uses the natural logarithm of the odds: log-odds =log
e
[p/(1-9)] as the depen-
dent variable. Probit analysis transforms the probability using the standard normal distribution, which means
that the number of standard deviations from the mean corresponds to the probability (e.g., a probability of
.025 becomes −1.96 and a probability of .975 becomes +1.96). Recovering the probability from the probit
estimates is the reverse (e.g., +1.96 becomes .025).
44. The appendix table includes results for both probit and logistic regression to make clear their
equivalence.
Appointed or Elected 21
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

using partisan elections but no such effect where nonpartisan elections are used.
Similarly, competitiveness has a strong effect in partisan election states and only a mar-
ginal effect in nonpartisan election states. Years on the court has the expected effect
with those having been on the court longer before their initial elections less likely
to be opposed in those elections. Term length has a marginal effect in the expected
direction. The effects of PJIEOS and competitiveness are shown graphically in
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
TN (9,5)
SD (12,1)
MD (16,0)
OR (27,14)
GA (34,5)
ID (22,9)
MS (33,13)
ND (17,6)
NV (13,16)
FL (12,5)
MT (17,20)
WA (30,18)
AL (18,25)
MN (43,2)
NC (40,11)
OK (7,7)
NM (15,9)
WI (22,14)
TX (37,49)
AZ (5,7)
NY (5,12)
IL (6,31)
PA (9,8)
OH (20,19)
KY (13,27)
CO (5,6)
MI (24,14)
WV (14,10*)
(a)
Percentage of Appointees Opposed at First Election PJIEOS**
*First number by state abbreviation is number of vacancies filled by appointment; second
number is number of vacancies filled by open-seat elections.
**Percentage of Justices Initially Elected to Open Seats
%Opposed = 27.52 + 0.85 × PJIEOS
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Appointees Opposed at First Election
0% 10% 20% 30% 40% 50% 60% 70% 80%
PJIEOS**
**Percentage of Justices Initially Elected to Open Seats
(b)
FIGURE 5.
Appointees Opposed at First Election.
22 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

Figure 6;45 the absence of an effect of PJIEOS in nonpartisan states is clear, as is the
smaller effect of competitiveness in those states.
Appointees’Success in Standing for Election
A third way to see the role of a possible state norm is to consider the likelihood
that appointees will survive the subsequent election. As he did regarding the likelihood
that recently appointed justices would face opposition in their first election, Dubois
(1980, 141–43) reported on the success of appointees in those elections. He found that
those appointees who were opposed were considerably less likely to lose the election in
nonpartisan states (18.8 percent) than were appointees in partisan/semi-partisan states
(40.3 percent). Including appointees who were unopposed at their first election, only
4.1 percent in nonpartisan states lost compared to 32.8 in partisan/semi-partisan states.
In the current study, the comparable figures are 38.7 percent for opposed appointees and
25.6 percent for all appointees in partisan/semi-partisan states, and 15.4 percent for
opposed and 8.2 percent for all appointees in nonpartisan states.
Figure 7 shows the aggregate relationship between the percentage of appointees
defeated and PJIEOS. The circles and solid lines represent state percentages of all
appointees defeated (n =28); the diamonds and broken line represent the state per-
centages of opposed appointees defeated (n =23). The slopes of the two lines are sim-
ilar (0.43 for all appointees and 0.38 for opposed appointees) but the regression fits (r2s)
are quite different (0.24 and 0.10 respectively); the relationship for opposed appointees
does not meet the criterion for statistical significance (p =.134).
PJIEOS**
0
Competitive
20
Pr(Opposed)
40 60 80 100
0
2
4
6
8
10
0
.2
.4
.6
.8
1
Partisan-Election States*
PJIEOS**
0
Competitive
20
Pr(Opposed)
40 60 80 100
0
2
4
6
8
10
0
.2
.4
.6
.8
1
Nonpartisan-Election States
Partisan-election states include states with semi-partisan systems.
**Percentage of Justices Initially Elected to Open Seats
FIGURE 6.
Likelihood Appointees Are Opposed at First Election by Competitiveness and
Percentage of Justices Initially Elected to Open Seats (PJIEOS).
45. To generate Figure 6, the values for term length and years since appointment were set to their
median values: 6 for term length and 1 for years since appointment.
Appointed or Elected 23
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

Except for term length, the same variables that I hypothesized would influence
whether appointees would be opposed at their first election would also potentially influ-
ence whether an opposed appointee won that election. Because whether an appointee is
opposed is not entirely random, failing to take into account whether there is an oppo-
nent would produce a biased analysis (Heckman 1979; Berk and Ray 1982). For this
reason, the appropriate approach is to use a model that accounts for the selection pro-
cess. The estimating method I use employs probit analysis to simultaneously estimate
both the “selection”equation and the “outcome”equation.46 I again condition the
effects of PJIEOS and competitiveness on type of election (collapsing partisan and
semi-partisan). The probit results are shown in online appendix Table A5 and indicate
that only one of the predictors has a statistically significant effect on election outcomes
of the opposed appointees, and that is PJIEOS in partisan election states; however, the
sign of the coefficient is in the wrong direction: as PJIEOS increases, the likelihood of
an opposed appointee in a partisan election state being defeated declines. Figure 8 pro-
vides two visual representations of the predicted results based on the probit model, vary-
ing PJIEOS and party competitiveness.47 What appears as almost a straight line in the
0%
10%
20%
30%
40%
50%
60%
70%
Percent Appointees Defeated
0% 10% 20% 30% 40% 50% 60% 70% 80%
PJIEOS*
% All Appointees Defeated % Opposed Appointees Defeated
Regression All Appointees Regression Opposed Appointees
*Percentage of Justices Initially Elected to Open Seats
FIGURE 7.
Success of Appointees Standing for Election.
46. The method is described by Van de Ven and Van Praag (1982).
47. To generate Figure 8, the value years since appointment was set to its median value (1).
24 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

top-right panel of the figure reflects the perspective from which the three-dimensional
representation is presented; the lower two panels provide perspectives that exchange
the axes of the two predictors.48
***
To summarize, in states where new justices tend to be elected in open-seat elec-
tions rather than being appointed (or defeating an incumbent), a justice who is
appointed is more likely to be opposed at the justice’sfirst election, but only in partisan
election states. The likelihood of opposition increases as party competitiveness increases
regardless of election type, but more strongly in partisan election states.49 In contrast,
the outcome of elections in which an appointee is opposed does not have the hypothe-
sized relationship with PJIEOS.
State Political Culture
There is a body of literature that focuses on state political culture. The best known
is Elazar’s(1966,85–94) tripartite distinction: individualistic, moralistic, and tradition-
alistic. According to Elazar’s analysis (97, 122), some states reflect only one of these
cultures, others two of them, and one state (Illinois) all three (122). There is a hint
PJIEOS**
0
Competitive
20
Pr(Losing|Opposed)
40 60 80 100 0246810
0
.2
.4
.6
.8
1
Partisan-Election States*
PJIEOS**
0
Competitive
20
Pr(Losing|Opposed)
40 60 80 100 0246810
0
.2
.4
.6
.8
1
Nonpartisan-Election States
Competitive
0
PJIEOS**
2
Pr(Losing|Opposed)
46810
020 40 60 80 100
0
.2
.4
.6
.8
1
Partisan-Election States*
Competitive
0
PJIEOS**
2
Pr(Losing|Opposed)
46810
020 40 60 80 100
0
.2
.4
.6
.8
1
Nonpartisan-Election States
*Partisan-election states include states with semi-partisan systems.
**Percentage of Justices Initially Elected to Open Seats
FIGURE 8.
Likelihood Opposed Appointees Are Defeated at First Election by Competitiveness
and Percentage of Justices Initially Elected to Open Seats (PJIEOS).
48. Ideally, one would have greater control over the perspective, but such control is not available in
the Stata procedure (surface) that I am using.
49. One might hypothesize that some of the relationship with party competitiveness reflects the pat-
tern in the old one-party Democratic South; however, an interaction term between region and party com-
petitiveness does not reach statistical significance.
Appointed or Elected 25
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

that there might be something of a continuum, running from moralistic through indi-
vidualistic to traditionalistic, but Elazar did not try to use this as more than a way of
grouping states (110). Not surprisingly, the cultures varied regionally with the South
and Southwest heavily traditionalist, northern states from Illinois to the East Coast
heavily individualistic, and the Midwest through the West Coast heavily moralistic.
Several scholars have made stabs at producing quantitative measures of Elazar’s idea.
Sharkansky (1969) proposed a single dimension and produced a scale ranging from
one to nine. Johnson (1976, 496–97) combined perspectives from another book by
Elazar (1970, 475–76) linking various religious groups to the three political cultures
with state-level census data on religious affiliation from the first half of the twentieth
century to propose a set of three measures, one for each culture; his scales range from 0
to .999.50
Correlating these four indicators with the percentage of new justices who were
elected in the thirty-four states using elections for at least part of the period under study
did not produce evidence of statistically significant relationships. The strongest corre-
lation, 0.32 with the individualistic dimension, did approach statistical significance (p
=.0695). Correlations with the other three measures of Elazar’s concept of state polit-
ical culture were not even close to statistical significance: −0.14, −0.16, and 0.02 for
moralistic, traditionalistic, and Sharkansky’s single dimension respectively. A regression
equation combining the four measures did account for 22 percent of the variation, but
given the small number of observations the equation was not overall statistically signifi-
cant (p =.1139) and none of the individual coefficients was statistically significant at
the .05 (two-tailed) level.
Lieske (2012) sought to update Elazar’s work using county-level data from the 2000
census and a 2000 survey of American church bodies.51 Rather than replicating Elazar’s
approach, Lieske applied two statistical techniques, factor analysis and cluster analysis,
to fifty state-level cultural indices and derived eleven relatively homogeneous political
subcultures (541–43): Heartland, Latino, Nordic, Border, Mormon, Global, Blackbelt,
Native American, Germanic, Rurban, and Anglo-French. To provide a measure of each
dimension, Lieske aggregated “the proportion of statewide population that are under the
influence of each regional subculture”; for each state, those proportions add to one.
Only the Nordic dimension among Lieske’s eleven dimensions has a statistically signif-
icant correlation (-.41, p =.0172) with the percentage of justices initially elected; the
strong negative correlation is not surprising given that the dimension is the strongest of
the cultural dimensions in Minnesota (.77 out of 1.00), where only three of forty-one
justices were initially elected between 1946 and 2020.
The overall lack of relationship is not surprising. In a study I did some years ago
(Kritzer 1979), I found that the Sharkansky and Johnson measures of political culture
were unrelated to several indicators of state judicial/legal systems. Thus, it appears that
Elazar’s notion of state political culture fails to capture a dimension of culture relevant
for understanding important legal phenomena and leaves the question of whether there
50. The correlations of Johnson’s measures with Sharkansky’s single scale were .67 for moralistic, −.54
for individualistic, and .81 for traditionalistic (Kritzer 1979, 499).
51. Lieske (1993) had previously reported a similar analysis using 1980 data; the overall patterns are
similar.
26 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

is something that could be described as a distinct dimension of state “legal culture”that
relates to the state norms about how judges should be selected.
Friedman (1969, 34) provides a succinct definition of the general concept of a legal
culture:“the network of values and attitudes relating to law.”Although the concept of
legal culture has most often been used to understand cross-national variations (see
Ehrmann 1976; Cotterrell 2006; Nelken 2012), Silbey (2010, 475–76) points out that
there are several threads of discourse on legal culture, including legal ideology, legal
consciousness, cultures of legality, and the structure of legality. However, little of this
research has focused on variation in legal culture among the fifty US states. The one
approach that has applied the idea of legal culture to specific geographic areas in the
United States employs the idea of “local legal culture,”defined as “common practitioner
norms governing case handling and participant behavior in court”(Church 1985, 449).
This concept has been applied in research on a range of specific topics,52 but none of the
studies have produced generalizable measure of local legal culture that could be applied
at the state level.
In my earlier study (Kritzer 1979) that found little connection between measures
derived from Elazar’s work and several legal variables, I did find a dimension that the I
labeled the “role of the legal system.”That dimension related to three indicators:
Walker’s measure of state innovation (Walker 1969, 883), Vines and Jacob’s measure
of legal professionalism (Vines and Jacob 1971, 292), and Glick and Vines’s measure of
the modernization of court organization (Glick and Vines 1973, 30). However, neither
those individual indicators nor a scale formed by combining three indicators had a sta-
tistically significant correlation with the percentage of justices initially elected.53
Thus, the analyses in this subsection fail to demonstrate a relationship with state
political or legal culture as I have been able to find or create possible measures of these
concepts. Nonetheless, it remains an open question whether there might be a way to
operationalize the notion of state legal culture that could help explain the variation in
the likelihood of initial election. Whether there is such a concept that could be oper-
ationalized to produce a valid measure must be left to future research.
The Role of State Norms in Election versus Appointment
In this section, I have attempted to assess the role of state norms regarding how
judges of a state’s highest court should in practice be selected. The fact that for all
appointees considered in this analysis their state constitutions called for election rep-
resents the presumed ideal as to how justices should be selected. The question of interest
concerns understanding the working norm regarding selection. To try to assess the pos-
sible role of a working norm, I presented several analyses, two of which support the
argument that norms are important and one providing conditional support.
52. Topics covered in these studies include court delay (Church et al. 1978), plea bargaining (Church
1995), bankruptcy proceedings (Sullivan, Warren, and Westbrook 1994; LoPucki 1996), domestic violence
(Currul-Dykeman 2014), and the control of litigation in federal courts (Kritzer and Zemans 1993).
53. The scale used was a factor score based on the three individual indicators; that score accounted for
a little over two-thirds of the variation among the three indicators.
Appointed or Elected 27
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

First, given that the likelihood of initial election is clearly higher in states using
partisan rather than nonpartisan elections, one would expect that a state shifting from
partisan elections to nonpartisan elections would see a drop in the percentage of new
justices elected. The fact that such a drop did not occur supports the idea that some-
thing akin to a state norm is playing a role. Second, the fact that how a justice’s prede-
cessor was selected predicts the method of the justice’s own selection supports the idea
that norms play a role. Third, an indirect indicator of a norm that justices should be
elected rather than appointed is the percentage of justices initially elected to open seats
(PJIEOS). If such a norm exists, I hypothesized that appointees would be more likely to
be opposed at their first elections and more likely to lose that election in states where a
high percentage of new justices were initially elected to open seats; more specifically,
Ifitted models predicting opposition and defeat conditional on opposition that included
as a predictor PJIEOS. The models produced mixed results. Opposition did increase
with PJIEOS, but the relationship was much weaker in nonpartisan states than in par-
tisan/semi-partisan states. In contrast, there was no relationship between PJIEOS and
the defeat of opposed appointees; in fact, the relationship in partisan states was inverse
to what was expected. It is noteworthy that various measures of state political and legal
culture were uncorrelated with the percentage of justices in a state that were initially
elected.
DISCUSSION
Although the constitutions of twenty-two of the American states currently call for
justices of the state’s highest court(s) to be elected, as did another twelve states for some
part of the period between 1946 and 2020, only about half of the persons who joined
those courts as more than strictly temporary members over those seventy-five years ini-
tially obtained their seats by being elected to vacancies created by death, retirement,
resignation, or the creation of new positions. This article examined various factors that
might explain the pattern of election versus appointment.
The analysis confirms earlier research showing that the likelihood of initial selec-
tion by election rather than appointment to a midterm vacancy is related to election
format, with election most likely in partisan election states and least likely in nonparti-
san election states. There is also some evidence that party competitiveness tends to
increase initial selection through election. Somewhat surprisingly, laws requiring justi-
ces who reach a mandatory retirement age to vacate their positions before the end of a
term fail to account for a greater likelihood of appointment. Most important, there
seems to be ongoing state practice or norms. That is, in some states (e.g., Minnesota
and Georgia54) there is a strong norm that incumbents leave before the end of their
terms, allowing their successors to be appointed, rather than leaving at the end of a
term so that there is an open-seat election. In Georgia, the norm was not affected
by the switch from partisan to nonpartisan elections. At the other end one finds
54. I noted previously that Minnesota’s norm of usually appointing new justices goes back at least since
the adoption of nonpartisan elections. In Georgia, this norm seems to have been in existence since at least
the beginning of the twentieth century: between 1900 and 1945, nineteen of the twenty-five new justices
joining the Georgia Supreme Court were appointed to fill vacancies.
28 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

Illinois and Kentucky, where over 70 percent of justices come to their seats by election,
and the pattern in Kentucky did not change when that state switched from partisan to
nonpartisan elections.
As noted in the Introduction, Shugerman (2012) argues that a key motivation for
replacing gubernatorial appointment with popular election in the mid-nineteenth cen-
tury was to increase the independence of state judges from the elected officials who had
previously appointed the judges, but it is not clear that that goal was initially achieved
given the continuing role of political party leaders in nominating candidates. Today the
choice between appointment and selection is more often discussed as reflecting a ten-
sion between independence and accountability, with popular elections more on the
accountability side of the ledger. However, if as noted in the Introduction the concern
is accountability, initial selection is less important than how judges are retained. Even
though accountability is more a matter of retention than selection, the initial selection
serves at least two purposes.
First, initial selection speaks to the symbolic politics question of who we want to
choose our judges. In the United States, there is a strong norm that public officials of all
sorts should be elected by the people, even though we know that the people typically pay
relatively little attention to down-ballot elections. Although voters may not want to take
the time necessary to evaluate significant numbers of candidates for low-level office, they
apparently take some satisfaction from believing they have had a say. A common form of
disparagement in the United States is to say that a person “couldn’t even be elected dog-
catcher.”In fact, there are at least two places where the dogcatcher (“animal control offi-
cer”) is an elected position. Moreover, in one of those places, a small town in New
Hampshire, the citizens in 2018 voted to reinstate the position as an elected town posi-
tion several years after it had been changed to an appointed position (Giddings 2018). If
Americans are going to want to elect the dogcatcher, or the library board, or the water
and sewer commission, it is not surprising that they will want to elect judges.
A second purpose that initial selection can serve is aiding in achieving descriptive
representation by the inclusion of otherwise underrepresented groups. There is a substan-
tial literature on whether the formal selection system for state supreme court justices or
other judges affects the characteristics of the justices serving on those courts (Jacob 1964;
Canon 1972; Glick and Emmert 1987). There is a more specific literature that has exam-
ined the impact of formal selection systems on the representation of women (Alozie 1996;
Hurwitz and Lanier 2003,340–42; Williams 2007; Frederick and Streb 2008), Blacks
(Alozie 1988), Hispanics (Alozie 1990), 318–19), and racial and ethnic minorities more
generally (Hurwitz and Lanier 2003,338–40) as judges on state courts. Most of this
research concludes that the formal selection system has relatively little impact.
There is a more limited literature that examines whether how individual justices
are actually selected matters for inclusion of women and minorities. In that literature
justices are categorized as “elected”only if they initially came to their positions by win-
ning an open seat or defeating an incumbent. Two studies demonstrate that Blacks or
other non-Whites are less likely to come to be on state supreme courts through election
than are Whites (Martin and Pyle 2002; Holmes and Emrey 2006, 7). In contrast, a
study looking at all levels of courts found that Blacks were more likely to be selected
through election systems than through appointive systems (Graham 1990, 328–30).
Regarding state supreme courts, the first Black justice to sit on a state supreme court
Appointed or Elected 29
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

since Reconstruction was Otis Smith who was appointed to the Michigan Supreme Court
in 1961;55 it was another ten years before the next Black justice was appointed, Robert
Morton Duncan in Ohio. The first Black justice who was initially elected to a state supreme
court was Robert N. C. Nix in 1972 in Pennsylvania; not until 1990 did another Black
justice gain a seat on a state supreme court by election, Charles Freeman in Illinois. By
my count, there have been seventy Black justices post-Reconstruction. Five of them served
as strictly interim justices in Arkansas, Louisiana, and Pennsylvania. Of the remaining sixty-
five,56 twenty-nine served in states using systems that exclusively used appointment for ini-
tial selection. Of the thirty-six Black justices who served in states where the constitution
specified initial election, only ten (29 percent) were initially elected.
Regarding women on the bench, the results are less clear (Martin and Pyle 2002;
Holmes and Emrey 2006, 7). The first woman to serve on a state supreme court,
Florence Allen, was elected to the Ohio Supreme Court in 1922.57 Next came Rhoda
Lewis thirty-seven years later when she was appointed to the Hawaii Supreme Court in
1959. Three women joined state supreme courts in the early 1960s, one elected (Lorna
Lockwood in Arizona) and two appointed to midterm vacancies (Anne Alpert in
Pennsylvania and Susie Sharp in North Carolina). As this is written, 318 women have
served on state supreme courts.58 Only eighty-two (25.8 percent) have come to their
positions by popular election. Omitting those selected in states that do not use popular
election leaves 172 justices, eighty-two (47.7 percent) of whom were elected. Perhaps of
more importance is that since 1990 just over half (51.7 percent of 143 justices) were
elected, but up until 1990 only a quarter (25.0 percent of twenty-eight) were elected.59
Descriptive representation can overlap with “substantive representation”if demo-
graphic groups have specific interests. One issue in assessing whether judges with certain
demographic characteristics, such as race and gender, differ from their colleagues in the
decisions that they make is that one must take into account other factors related to
decisional propensity, particularly political ideology. This is a challenge when looking
at the race of the judge because minority judges are overwhelmingly Democrats, but it is
less of a problem for gender. Also, one might expect the effects of gender or race to be
more prominent in cases that raise issues related to gender or race. There are studies
showing that women judges are more pro-plaintiff in discrimination cases than are
men, even controlling for party or ideology (Davis, Haire, and Songer 1993, 132;
Songer, Davis, and Haire 1994, 435), and in sexual harassment cases prior to the
Clarence Thomas hearings (McCall 2003,90–93), and there is one study showing that
women are more pro-victim (i.e., anti-defendant) in domestic violence cases when the
victim is female (McCall 2008, 288). However, gender effects can show up in other types
55. I constructed a list of Black justices who served (or are serving) on state supreme courts from two
websites: https://www.blackpast.org/special-features/african-american-state-supreme-court-justices-since-
1870/ and https://en.wikipedia.org/wiki/List_of_African-American_jurists.
56. This includes judges who served on the Texas and Oklahoma Courts of Criminal Appeals.
57. Allen later became the first Article III judge in the federal court system when she was appointed to
the US Court of Appeals (Sixth Circuit) in 1934.
58. This count, and the discussion that follows, is based on a list accessed January 19, 2022, at https://
en.wikipedia.org/wiki/List_of_female_state_supreme_court_justices, that I supplemented by adding the
eleven women who have served on either the Texas Court of Criminal Appeals or the Oklahoma
Court of Criminal Appeals.
59. Chi-square =6.72, p =.009.
30 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

of cases; one study found that women judges on state supreme courts were more likely to
vote in a liberal direction than were men in obscenity cases and death penalty cases after
taking into account the judge’s political party (Songer and Crews-Meyer 2000). Other
research, focused on the federal Court of Appeals, has shown that the presence of a
woman on a three-judge panel can affect the decisions of her male colleagues in discrim-
ination cases (Farhang and Wawro 2004;Peresie2005); a similar effect has been found for
the presence of a Black judge on a federal Court of Appeals panel (Kastellec 2013).60
Thus, even though it is best to think in terms of retention when the focus is on account-
ability, the initial selection process can and does play a role in some types of cases.
In conclusion, how judges are actually selected is important. It has implications for
how people think about the courts, and it affects who actually serves as judges and jus-
tices. Who serves has potential effects on the decisions that courts make. Although this
article has specifically focused on selection of state supreme court judges, the issue may
be as important, or more important, for who serves on lower-level state courts. In those
courts, a much larger proportion of judges may have been initially appointed in states
where the formal system specifies election than is the case for the state supreme courts.
Interestingly, in several states where appellate judges are appointed, trial judges have
continued to be elected (Kritzer 2020), at least in theory. If in practice in those states
most trial judges are initially appointed, one might ask why the electorate has not been
more willing to shift to a formal system of initial appointment. It might well be the case
that most people in election states incorrectly believe that most of their trial judges got
their positions through election even though that is not in fact the case. It would be
useful to know the degree to which people know how their state judges initially obtain
their positions, but I am not aware of any research that addresses this question.
I could conclude here with a discussion of the normative issue of how judges should
be selected. That question has been debated since the founding of the country, and there
is no definitive way to answer it. The choice of popular elections as the initial selection
mechanism reflects a set of goals, independence in Shugerman’s analysis or accountability
as argued by many contemporary observers. However, the system may not work to accom-
plish the goals sought because of how the system works in actual practice. Some might
argue that deviations from the sought goals reflect intentional efforts to undermine the
goals. However, it is equally likely that the system produces unanticipated consequences
that short-circuit the goals. Certainly, there are justices and judges who step down stra-
tegically, not just at the federal level (e.g., Justice Breyer’s retirement), but also at the state
level—that is, departures from the bench midterm to permit the governor to fill a vacancy
(see Curry and Hurwitz 2016).61 However, it is also likely that there are unanticipated or
60. See Harvey and Yntiso (2021) for a somewhat similar pattern in the appellate division of the New
York Supreme Court (New York’s intermediate appellate court). A study of panel decisions in federal plead-
ing cases found no effect of having one woman on the panel but there was an effect of having two women
(Burbank and Farhang 2021, 2257–58).
61. Just as I was finishing the final draft of this article (early April 2022), Regina Chu, the Minnesota
judge who presided over the trial of the female police officer who grabbed her handgun rather than her taser
and fatally shot Daunte Wright in 2021 during a traffic stop and attempted arrest for an outstanding warrant,
announced her retirement. Chu had two years until reaching the mandatory retirement age. Her current
term expires at the end of the current (2022) year. She could easily have served out her term and allowed
her position to be filled by the voters. However, by retiring in early May, she allows the governor to name
her replacement (see Walsh 2022).
Appointed or Elected 31
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

inevitable effects. Those who viewed judicial elections as producing independence from
the political branch failed to consider nominating politics. The adoption of mandatory
retirement by necessity produces the need to fill midterm vacancies, although as my anal-
ysis shows, the presence of mandatory retirement does not explain variation among states
in the likelihood that justices are initially elected. Importantly, the absence of mandatory
retirement also produces midterm vacancies due to ill health or death as justices age
beyond what might have been a mandatory retirement age. These kinds of unanticipated
or inevitable consequences can, and not infrequently do, prevent intended goals from
being fully reached.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/
lsi.2022.23
REFERENCES
Achury, Susan, Jason P. Casellas, Scott J. Hofer, and Matthew Ward. “The Impact of Racial
Representation on Judicial Legitimacy: White Rections to Latinos on the Bench.”Political
Research Quarterly (forthcoming).
Adamany, David, and Philip Dubois. “Electing State Judges.”Wisconsin Law Review 1976,
no. 3 (1976): 731–79.
Alozie, Nicholas O. “Black Representation on State Judiciaries.”Social Science Quarterly 69, no. 4
(1988): 979–86.
—— “Distribution of Women and Minority Judges: The Effects of Judicial Selection Methods.”Social
Science Quarterly 71, no. 2 (1990): 315–25.
—— “Selection Methods and the Recruitment of Women to State Courts of Last Resort.”Social
Science Quarterly 77, no. 1 (1996): 110–26.
Atkins, Burton M. “Judicial Elections - What the Evidence Shows.”Florida Bar Journal 50, no. 3
(1976): 152–57.
Atkins, Burton M., and Henry R. Glick. “Formal Judicial Recruitment and State Supreme Court
Decisions.”American Politics Quarterly 2, no. 4 (1974): 427–49.
Auman, Francis R. “The Selection, Tenure, Retirement and Compensation of Judges in Ohio.”
University of Cincinnati Law Review 5, no. 4 (1931): 408–28.
Berk, Richard A., and Subhash C. Ray. “Selection Biases in Sociological Data.”Social Science Research
11 (1982): 352–98.
Berry, Kate, and Cathleen Lisk. Appointed and Advantaged: How Interim Vacancies Shape State
Courts. New York: Brennan Center for Justice, New York University School of Law, 2017.
https://www.brennancenter.org/sites/default/files/analysis/Appointed_and_Advantaged_How_
Interim_Appointments_Shape_State_Courts.pdf.
Bonneau, Chris W., and Melinda Gann Hall. In Defense of Judicial Elections. New York: Routledge,
2009.
Bowers, Michael W. “The Impact of Judicial Selection Methods in Nevada: Some Empirical
Observations.”Nevada Public Affairs Review, no. 2 (1990): 3–8.
Brace, Paul, and Brent D. Boyea. “State Public Opinion, the Death Penalty, and the Practice of
Electing Judges.”American Journal of Political Science 52, no. 2 (2008): 360–72.
Burbank, Stephen B., and Sean Farhang. “Politics, Identity, and Pleading Decisions of the U.S. Courts
of Appeals.”University of Pennsylvania Law Review 169, no. 8 (2021): 2225–89.
32 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

Canes-Wrone, Brandice, Tom S. Clark, and Jason P. Kelly. “Judicial Selection and Death Penalty
Decisions.”American Political Science Review 108, no. 1 (2014): 23–39.
Canes-Wrone, Brandice, Tom S. Clark, and Jee-Kwang Park. “Judicial Independence and Retention
Elections.”Journal of Law, Economics, and Organization 28, no. 2 (2012): 211–34.
Canes-Wrone, Brandice, Tom S. Clark, and Amy Semet. “Judicial Elections, Public Opinion, and
Decisions on Lower-Salience Issues.”Journal of Empirical Legal Studies 15, no. 4 (2018): 672–707.
Canon, Bradley C. “The Impact of Formal Selection Processes on the Characteristics of Judges—
Reconsidered.”Law & Society Review 6, no. 4 (1972): 579–93.
Church, Thomas W., Jr. “Examining Local Legal Culture.”American Bar Foundation Research Journal
1985, no. 3 (1985): 449–518.
Church, Thomas W., Jr. “Plea Bargaining and Local Legal Culture.”In Contemplating Courts, edited by
Lee Epstein, 132–54. Washington, DC: CQ Press, 1995.
Church, Thomas W., Jr., Alan Carlson, Jo-Lynne Lee, and Teresa Tan. Justice Delayed: The Pace of
Litigation in Urban Trial Courts. Williamsburg, VA: National Center for State Courts, 1978.
Cook, Beverly Blair. “Role Lag in Urban Trial Courts.”Western Political Quarterly 25, no. 2 (1972):
234–48.
Cotterrell, Roger. “Comparative Law and Legal Culture.”In The Oxford Handbook of Comparative Law,
edited by Mathias Reimann and Reinhard Zimmermann, 709–37. Oxford, UK: Oxford
University Press, 2006.
Currul-Dykeman, Kathleen. “Domestic Violence Case Processing: A Matter of Local Legal Culture.”
Contemporary Justice Review 17, no. 2 (2014): 250–72.
Curry, Todd A., and Mark S. Hurwitz. “Strategic Retirements of Elected and Appointed Justices:
A Hazard Model Approach.”The Journal of Politics 78, no. 4 (2016): 1061–75.
Davis, Sue, Susan Haire, and Donald R. Songer. “Voting Behavior and Gender on the U.S. Courts of
Appeals.”Judicature 77, no. 3 (1993): 129–33.
DeBow, Michale, Diane Brey, Erick Kaardal, John Soroko, Frank Strickland, and Michael Wallace.
“The Case for Partisan Judicial Elections.”University of Toledo Law Review 33, no. 2 (2002):
393–409.
Driscoll, Amanda, and Michael J. Nelson. “Judicial Selection and the Democratization of Justice:
Lessons from the Bolivian Judicial Elections.”Journal of Law and Courts 3, no. 1 (2015):
115–48.
Dubois, Philip L. “Judicial Elections in the States: Patterns and Consequences.”PhD diss., University
of Wisconsin, 1978.
—— From Bench to Ballot: Judicial Elections and the Quest for Accountability. Austin, TX: University of
Texas Press, 1980.
Ehrmann, Henry W. Comparative Legal Cultures. Englewood Cliffs, NJ: Prentice-Hall, 1976.
Elazar, Danial J. American Federalism: A View from the States. New York: Thomas Y. Crowell, 1966.
—— Cities of the Prairie. New York: Basic Books, 1970.
Farhang, Sean, and Gregory Wawro. “Institutional Dynamics on the U.S. Court of Appeals: Minority
Representation under Panel Decision Making.”Journal of Law, Economics, and Organization 20,
no. 2 (2004): 299–330.
Flango, Victor E., and Craig R. Ducat. “What Difference Does Method of Judicial Selection Make?
Selection Procedures in State Courts of Last Resort.”Justice System Journal 5, no. 1 (1979):
25–44.
Frederick, Brian, and Matthew J. Streb. “Women Running for Judge: The Impact of Sex on Candidate
Success in State Intermediate Appellate Court Elections.”Social Science Quarterly 89, no. 4
(2008): 937–54.
Friedman, Lawrence M. “Legal Culture and Social Development.”Law & Society Review 4, no. 1
(1969): 29–44.
Giddings, Kristen. “Danville: O’Neil and Johannesen Remain Selectmen.”Eagle-Tribune, March 13,
2018. https://www.eagletribune.com/news/danville-oneil-and-johannesen-remain-selectmen/
article_71b629bb-337a-55a1-83e2-2b5c97dadcfb.html.
Glick, Henry R., and Craig F. Emmert. “Selection Systems and Judicial Characteristics: The
Recruitment of State Supreme Court Judges.”Judicature 70, no. 3 (1987): 228–35.
Appointed or Elected 33
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

Glick, Henry Robert, and Kenneth N. Vines. State Court Systems. Englewood Cliffs, NJ: Prentice-Hall,
1973.
Graham, Barbara Luck. “Do Judicial Selection Systems Matter? A Study of Black Representation on
State Courts.”American Politics Quarterly 18, no. 3 (1990): 316–36.
Hall, Kermit L. “Progressive Reform and the Decline of Democratic Accountability: The Popular
Election of State Supreme Court Judges, 1850-1920.”American Bar Foundation Research
Journal 1984 (1984): 345–69.
Harvey, Anna, and Sidak Yntiso. “Judicial Accountability and Racial Disparity in Criminal Appeals.”
Journal of Legal Studies 50, no. 2 (2021): 261–302.
Heckman, James J. “Sample Selection Bias as a Specification Error.”Econometrica 48, no. 1 (1979):
153–61.
Herndon, James. “Appointment as a Means of Initial Accession to Elective Courts of Last Resort.”
North Dakota Law Review 38, no. 1 (1962): 60–73.
Holmes, Lisa M., and Jolly A Emrey. “Court Diversification: Staffing the State Courts of Last Resort
through Interim Appointment.”Justice System Journal 27, no. 1 (2006): 1–13.
Howell, William G. Power without Persuasion: A Theory of Presidential Action. Princeton, NJ: Princeton
University Press, 2003.
Hurwitz, Mark S., and Drew Noble Lanier. “Explaining Judicial Diversity: The Differential Ability of
Women and Minorities to Attain Seats on State Supreme and Appellate Courts.”State Politics &
Policy Quarterly 3, no. 4 (2003): 329–52.
Jacob, Herbert. “The Effect of Institutional Differences in the Recruitment Process, Case of State
Judges.”Journal of Public Law 13, no. 1 (1964): 104–19.
Johnson, Charles A. “Political Culture in American States: Elazar’s Formulation Examined.”American
Journal of Political Science 20, no. 3 (1976): 491–509.
Kastellec, Jonathan P. “Racial Diversity and Judicial Influence on Appellate Courts.”American Journal
of Political Science 57, no. 1 (2013): 167–83.
Kritzer, Herbert M. “Political Cultures, Trial Courts, and Criminal Cases.”In The Study of Criminal
Courts: Political Perspectives, edited by Peter F. Nardulli, 131. Cambridge, MA: Ballinger
Publishing Company, 1979.
—— Justices on the Ballot: Continuity and Change in State Supreme Court Elections. New York:
Cambridge University Press, 2015.
—— “Judicial Elections in the 2010s.”DePaul Law Review 67, no. 2 (2018): 387–423.
—— Judicial Selection in the States: Politics and the Struggle for Reform. New York: Cambridge University
Press, 2020.
—— “Polarization and Partisanship in State Supreme Court Elections.”Judicature 105, no. 3
(2021–22): 65–75.
Kritzer, Herbert M., and Frances Kahn Zemans. “Local Legal Culture and the Control of Litigation.”
Law & Society Review 27, no. 3 (1993): 535–58.
Lieske, Joel. “Regional Subcultures of the United States.”Journal of Politics 55, no. 4 (1993): 888–913.
——.“American State Cultures: Testing a New Measure and Theory.”Publius 42, no. 1 (2012):
108–33.
Lindquist, Stefanie A. “Judicial Activism in State Supreme Courts: Institutional Design and Judicial
Behavior.”Stanford Law & Policy Review 28, no. 1 (2017): 61–108.
LoPucki, Lynn M. “Legal Culture, Legal Strategy, and the Law in Lawyers’Heads.”Northwestern
University Law Review 90, no. 4 (1996): 1498–1556.
Martin, Elaine, and Barry Pyle. “Gender and Racial Diversification of State Supreme Courts.”
Women & Politics 24, no. 2 (2002): 35–52.
McCall, Madhavi. “Gender, Judicial Dissent, and Issue Salience: The Voting Behavior of State
Supreme Court Justices in Sexual Harassment Cases, 1980–1998.”Social Science Journal 40,
no. 1 (2003): 79–97.
——.“Structuring Gender’s Impact: Judicial Voting across Criminal Justice Cases.”American Politics
Research 36, no. 2 (2008): 264–96.
34 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

Mendicino, Anthon. “Sally Updyke Mundy Appointed to Pennsylvania Supreme Court.”Pittsburgh
Post-Gazette, June 28, 2016. https://www.post-gazette.com/news/state/2016/06/28/Justice-Mundy-
confirmed-to-Pennsylvania-Supreme-Court/stories/201606280123.
Nelken, David. “Legal Cultures.”In Comparative Law and Society, edited by David S. Clark, 312–27.
Cheltenham, UK: Edward Elgar, 2012.
Nelson, Michael J. “Judicial Elections and Support for State Courts.”In Judicial Elections in the 21st
Century, edited by Chris W. Bonneau and Melinda Gann Hall, 217–32. New York: Routledge,
2017.
O’Brien, David M. “The Politics of Judicial Selection and Appointments in Japan and Ten South and
Southeast Asian Countries.”In Appointing Judges in the Age of Judicial Power: Critical Perspectives,
edited by Kate Malleson and Peter Russell, 355–74. Toronto: University of Toronto Press, 2006.
O’Malley, Deobrah O. A Defense of the Elected Judiciary. Heritage Foundation, Legal Memorandum No.
57, 2010. http://thf_media.s3.amazonaws.com/2010/pdf/lm0057.pdf.
Overby, L. Marvin, Robert D. Brown, John M. Bruce, Charles E. Smith, Jr., and John W. Winkle, III,
“Race, Political Empowerment, and Minority Perceptions of Judicial Fairness.”Social Science
Quarterly 86, no. 2 (2005): 444–62.
Peltason, Jack W. “The Missouri Plan for the Selection of Judges.”University of Missouri Studies 20, no.
2 (1945).
Peresie, Jennifer L. “Female Judges Matter: Gender and Collegial Decision-Making in the Federal
Applellate Court.”Yale Law Journal 114, no. 7 (2005): 1759–90.
Pitkin, Hannah. The Concept of Representation. Berkeley, CA: University of California Press, 1967.
Rafferty, William E. “Happy Birthday! Now Get Out.”Judicature 90, no. 3 (2015): 5–6.
Ranney, Austin. “Parties in State Politics.”In Politics in the American States: A Comparative Analysis,
edited by Herbert Jacobs and Kenneth N. Vines, 61–99. Boston: Little, Brown and Company,
1965.
Salyer, Lucy E. “The California Supreme Court in an Age of Reform, 1910–1940.”In Constitutional
Governance and Judicial Power: The History of the California Supreme Court, edited by Harry N.
Scheiber, 141–243. Berkeley, CA: Berkeley Public Policy Press, Institute of Governmental
Studies, University of California, 2016.
Scherer, Nancy. Diversifying the Bench: Race, Gender and the Legitimacy of the Courts. New York: New
York University Press (forthcoming).
Scherer, Nancy, and Brett Curry. “Does Descriptive Race Representation Enhance Institutional
Legitimacy? The Case of the U.S. Courts.”The Journal of Politics 72, no. 1 (2010): 90–104.
Schlesinger, Joseph. Ambitions and Politics: Political Careers in the United States. Chicago: Rand
McNally, 1966.
Sharkansky, Ira. “The Utility of Elazar’s Political Culture: A Research Note.”Polity 2, no. 1 (1969):
66–83.
Sheldon, Charles H., and Linda S. Maule. Choosing Justice: The Recruitment of State and Federal Judges.
Pullman, WA: Washington State University Press, 1997.
Shepherd, Joanna M. “The Influence of Retention Politics on Judges’Voting.”Journal of Legal Studies
38, no. 1 (2009): 169–206.
Shugerman, Jed Handelsman. The People’s Courts: Pursuing Judicial Independence in America.
Cambridge, MA: Harvard University Press, 2012.
Silbey, Susan S. “Legal Culture and the Cultures of Legality.”In Handbook of Cultural Sociology, edited
by John R. Hall, Linda Grindstaff, and Ming-Cheng Lo, 470–79. New York: Routledge, 2010.
Songer, Donald R., and Kelley A. Crews-Meyer. “Does Judge Gender Matter? Decision Making in
State Supreme Courts.”Social Science Quarterly 81, no. 3 (2000): 750–62.
Songer, Donald R., Sue Davis, and Susan Haire. “A Reappraisal of Diversification in the Federal
Courts: Gender Effects in the Courts of Appeals.”Journal of Politics 56, no. 2 (1994): 425–39.
Sullivan, Teresa A., Elzabeth Warren, and Jay Lawrence Westbrook. “The Persistence of Local Legal
Culture: Twenty Years of Evidence from the Federal Bankruptcy Courts.”Harvard Joural of Law &
Public Policy 17, no. 3 (1994): 801–65.
Appointed or Elected 35
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press

Van de Ven, Wynand P. M. M., and Bernard M. S. Van Praag. “The Demand for Deductibles in
Private Health Insurance: A Probit Model with Sample Selection.”Journal of Econometrics 17,
no. 2 (1982): 229–52.
Vines, Kenneth N., and Herbert Jacob. “State Courts.”In Politics in the American States: A Comparative
Analysis, edited by Herbert Jacob and Kenneth N. Vines, 272–311. Boston: Little Brown, 1971.
Walker, Jack. “The Diffusion of Innovations among the American States.”American Political Science
Review 63, no. 3 (1969): 880–99.
Walsh, Paul. “Fresh out of the Spotlight, Judge Chu Says She’ll Retire.”Minneapolis Star Tribune, April
2, 2022.
Weinschenk, Aaron, Mandi Baker, Zoe Betancourt, Vanessa Depies, Nathan Erck, Quinne Herolt,
Amanda Loehrke, Cameron Makurat, Hannah Malmberg, Clarice Martell, Jared Novitzke,
Bradley Riddle, Tara Sellen, Leah Tauferner, and Emily Zilliox. “Have State Supreme Court
Elections Nationalized?”Justice System Journal 41, no. 4 (2020): 313–22.
Williams, Margaret. “Women’s Representation on State Trial and Appellate Courts.”Social Science
Quarterly 88, no. 5 (2007): 1192–1204.
Winters, Glen R. “One-Man Judicial Selection.”Journal of the American Judicature Society 45, no. 9
(1962): 198–203.
Wipf, Thomas. “Swiss Cantons.”In Legal Systems of the World, vol. IV, edited by Herbert M. Kritzer,
1567–69. Santa Barbara, CA: ABC-CLIO, 2002.
CASES AND CONSTITUTIONS CITED
Maddux v. Blagojevich, 911 N.E.2d 508 (Ill. 2009).
Arkansas Constitution of 1874, Amendment 29, §2.
Louisiana Constitution of 1974, Article V, §22(B)
36 LAW & SOCIAL INQUIRY
https://doi.org/10.1017/lsi.2022.23 Published online by Cambridge University Press