Constitutional Provisions for Individual Human Rights (1977-1996): Are They More Than Mere
"Window Dressing?"
Author(s): Linda Camp Keith
Source: Political Research Quarterly, Vol. 55, No. 1, (Mar., 2002), pp. 111-143
Published by: Sage Publications, Inc. on behalf of the University of Utah
Stable URL: http://www.jstor.org/stable/3088068
Accessed: 14/05/2008 14:24
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=sage.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].
http://www.jstor.org
Constitutional Provisions for
Individual
Human
Rights
(1977-1996):
Are They More than Mere
"Window Dressing?"
COLLEGE
LINDA CAMP KEITH, COLLINCOUNTYCOMMUNITY
This global cross-nationalstudy seeks to build upon earlierstudies that
have tested the impact of constitutional provisions upon state human
rightsbehavior.I examine acrossa twenty-yearperiod the impact of constitutional provisions for six individual freedoms and four due process
rights on state abuse of the right to personalintegrity.Here I find statistical evidence that some constitutionalprovisions do matter,even when
controlling for democracy and for other factors known to influence
human rights behavior.While none of the constitutionalprovisions for
individualfreedomsis statisticallysignificant,two of the due processprovisions (provisionsfor fairand public trials)do decreasesubstantiallythe
likelihood that states will abuse their own citizens' human rights. The
other two due process provisions,which have become almost universal,
the ban against torture,and the writ of habeas corpus, are quite disappointing in that they do not produce the expected impact. Over the long
term,the trialprovisionswould lead to a decreaseof about one level in the
personal integrity abuse score, which is only somewhat less than the
impactproducedby othervariablesin the model, such as populationsize.
The fall of the Soviet Union and the dissolution of some Eastern Block states
have led to an unprecedented wave of newly drafted constitutions, many of
which incorporate bills of rights or individual provisions aimed at protecting
human rights. As of March 1997, 184 of the world's nation-states had promulgated formal constitutions, most of these being new creations (Flanz 1997). As
1993, more than 100 states had written bills of rights (Blaustein 1993). Some of
these bills of rights surpass even that of the United States in depth and breadth
PoliticalResearchQuarterly,Vol. 55, No. 1 (March2002): pp. 111-143
111
PoliticalResearch
Quarterly
of their protections(see Schwarzer1994). This wave of constitutionwriting follows upon a period of near universalacceptanceof an internationallyrecognized
set of minimum standards.The standardsare set forth in two documents, often
referredto as the InternationalBill of Rights:the UniversalDeclarationof Human
Rights (UNDHR) and the InternationalCovenant on Civil and Political Rights
(ICCPR).Scholarsargue that these documents representbroad consensus that
there is, in fact, a "coreset of human rights to which all humanityaspires"(Ibhawoh 2000: 838 and 843, emphasis added; also see Donnelly 1999; Linz 1992;
Martin 1991). Many of the newly drafted bills of rights in post-colonial and
formercommunist states explicitlyincorporatethe rights set forthin these international documents.
For many, these trends represent substantial progress and they engender
considerablepotential for future human rights protection.This optimism stems
from the general belief that the best way to safeguardindividual freedoms is
throughthe enumerationof rights,which extends the reachof the rule of law and
providesindividualsprotectionfrom the abuse of governmentpower (see Beatty
1994; Rosenthal1990). Presumablyregimesare less willing to abuse rights that
are clearlyand publicly promisedto their citizens in a legallybinding document
and that are supported by constitutionalmechanisms, such as an independent
judiciary.Even if constitutions do not serve as a charterof these fundamental
rights, they may still serve as a "bindingstatement of a people'saspirationsfor
themselves as a nation"(Murphy 1993: 10), and ultimatelymay foster the conditions under which a regime would give in to demands for these fundamental
protections.History,of course, has shown that many constitutions,such as those
of communistregimes,have often been mere "windowdressing"ratherthan substantiveprotectionfor individualhuman rights. Howard(1991: 3) arguesthat in
many of the world'scountries,the constitutionsare "worthlessscrapsof papers."
While the world's constitutionalexperts have been busy advising newly independent states and emerging democracies how best to provide constitutional
protectionsto their citizens, it appearsthat little scientificanalysishas been conducted that would support their advice or facilitatethe creationof new constitutions. Such analysiswould allow us not only to identify which elements of law
work, but also would allow us to determineunder what circumstancesor conditions these elements work best. This articleattempts to build upon recent constitutionalstudies by examining,acrosstwo decadesand a globalset of countries,
the question of whether constitutions make a differencein state human rights
behavior.This articlealso builds upon the growingbody of literaturethat seeks
to explain state abuse of the right to personalintegrity
CONSTITUTIONS AND HUMAN RIGHTS BEHAVIOR
Constitutionsare argued to be "the most important legal document for a
nation that subscribes to the rule of law" because they define the relationship
112
IndividualHumanRights(1977-1996)
between the nation'speople and their government(Siegan 1994: 72). Specifically,
constitutions impose constraintsupon government and protect the individual's
freedomsfrom arbitraryor abusive state action (see Andrews 1964; Finer 1974;
Stotsky 1993; Siegan 1994; Elster 1993; Finer, Bogdanor,and Rudden 1995).
The constitutional enumeration of such rights as those under study here is
believed to be the most important safeguardof the individual'sfreedoms (see
Rosenthal 1990). This enumerationis expected to improve states'human rights
practices in two ways. First, constitutions provide concrete standardsagainst
which the regime'sbehavior can be assessed not only by itself, but also by the
public and the world community (see Sartori 1962; Andrews 1964; Murphy
1993). Additionally,such provisions, and constitutions generally,may articulate
ideals and norms to which the regime aspires (Sartori 1962; Andrews 1964;
Finer 1974; Murphy 1993; Finer, Bogdanor,and Rudden 1995), and they may
"powerfullyshape popular culture"(Epp 1998: 13). In this sense, constitutions,
and bills of rightsin particular,may serve as a socializingtool that conditions the
expectationsof the public, promotingthe developmentof a rightsconsciousness
among the people (see Epp 1998; Martin 1991; MacGuigan 1965; Murphy
1993). Referringto bills of rights generally,James Madisonassertedthat "political truths declared in that solemn manner acquire by degrees the characterof
fundamentalmaxims of free Government,and as they become incorporatedwith
the national sentiment, counteractthe impulses of interest and passion"(Madison 1977: 298-99). In regardto the CanadianBills of Rights,MacGuigan(1965)
has arguedthat the bill of rights'real contributionwas, in fact, the way in which
it served as a stimulus to increaseCanadianrights consciousness and to further
the development of such rights.
Despite these theoreticalexpectationsconcerningconstitutions,some scholars (including some of the ones mentioned above) remain somewhat skeptical
about the actual influence of constitutions in the real world. They note significant instanceswhere constitutionshave been dishonoredor ignoredto the extent
that they were either reduced to "worthlessscraps of paper" or "convenient
screens"for tyrants to hide behind (see Sartori 1962; Andrews 1964; Howard
1991; Resenbrink 1991; Fruhling 1993; Murphy 1993; Schwarzer 1994; and
Epp 1998). Otherscholarssuch as Safran(1981) and Ludwikowski(1996) point
out that in many constitutions these personal freedoms are affordedprotection
only within the boundaries of the state'slaw and thereforemay be modified by
statutorylaw. Additionally,in some constitutions there has been an equal if not
greateremphasis placed on the individual'sdutiesto the nation. For example, in
the constitutions of socialist states, bills of rights were the norm; however, they
were usuallyjuxtaposed next to a list of duties that were consideredinseparable
from the list of rights (Ludwikowski1996: 227). Many constitutionsalso stipulate that any guaranteedrightsmust not be detrimentalto the nation'sbest interest or the public good. Thus, the level of protection promised by constitutions
113
PoliticalResearchOuarterlv
may be ratherlimited. These concerns have led some scholarsto ask if constitutions can ever be more than the mere "parchmentbarrier"that James Madison
sought to avoid. Ultimately,this is an empiricalquestion, one that has not been
examined fully by political scientists or legal scholars.
In part,both the skepticismand the lack of empiricalattentionto the question of whether constitutions matter are linked to the behavioristdismissal of
formalinstitutionsgenerally(see Sartori1962: 199, and Marchand Olsen 1984).
However, in political science there has been a renewed interest in institutions
Democracies,that
beginning with Powell'sgroundbreakingstudy, Contemporary
examined the impact of numerous institutionalfactorsthat affectedregime performance in regard to citizen participation,government stability,and political
violence. Numerous comparativestudies have followed Powell'swork, examining the impact or performanceof variousinstitutionalarrangementsacrossdemocraticsystems (e.g. Lijphart1984 and 1994; Diamond, Linz, and Lipset 1988;
Taageperaand Shugart1989; Shugartand Carey 1992; Stepanand Skach 1993;
Linz 1994; Bohrer1997). The wave of democratizationat the close of this century has further increased the scholarly attention to institutions. The recent
empirical interest in the relationshipbetween constitutions and human rights
behaviorfollows upon this resurgentinterestin institutionalism.Priorto the end
of the Cold Waronly two empiricalstudies looked at whetherconstitutionsmake
a differencein regardto state human rightsbehavior,but with the recent period
of constitution writing, the number of empiricalstudies have increased,as has
the level of scientific rigor.
EMPIRICALSTUDIES ASSESSINGTHE IMPACTOF CONSTITUTIONALPROVISIONS
The first known empiricalstudy of constitutionsand human rights is BoliBennett's(1976) bivariate analysis (comparison of means), which examined
states' human rights behavior in 1975 and found little correlationbetween his
human rightsviolationindices and his indices of rightsand duties, with only one
exception. Contraryto expectationshe found that the countriesthat less severely
violated human rightswere those in which the rightswere more constitutionally
restricted.'Pritchard's(1986) bivariateanalysisof states'human rights behavior
in 1974 produced counter-intuitiveresults in that greaterconstitutionalprotections were associated with lower levels of human rights protection.2Blasi and
1Boli-Bennett's measure of human violations is taken from an unpublished manuscriptby Winona
2
Hubrecht.It appearsto be a simple categorizationbased on AmnestyInternationalreports(the categories do not conform to any of the published variationsI have seen).
Her measure of human rights violations is somewhat problematicin that it combines civil and
politicalrightswith economic rightsas a single indicator.Theoreticallyone might expect that these
rights are not a single dimension of human rights but rather could be separate and competing
dimensions of human rights. Ensuringqualityof life may sometimes lead to the restrictionof civil
114
IndividualHumanRights(1977-1996)
Cingranelli's(1996) bivariatestudy of human rightsin 1993 is the firstpost-Cold
War study that assesses the impact of constitutions on human rights behavior;
however, their study focuses on constitutionalprovisions, which are not themselves direct promises of freedomsand rights. Ratherthey are institutionalprovisions that theoreticallywould be expected to facilitatethe provision of human
rights: an independent judiciary and a federal system of government. Overall,
their study finds only a weak direct link to the constitutionalprovisions. Blasi
and Cingranelli'sstudy has limitations similar to the two previous studies: it
encompassesonly a single year and is not a multivariatemodel that controls for
the numerous factors known to influence human rights behavior. Two recent
studies offer more rigorousstatisticalanalysisof this question.
Davenport's1996 study of 39 countriesacrossa 35-yearperiod (1948-1982)
is the most rigorousand comprehensiveanalysisto date. He identifies and analyzes four types of provisions of a constitution that may affect human rights
behavior:(1) constitutionalprovisionsthat explicitlymention rights;(2) constitutionalprovisionsthatexplicitlylimit rights;(3) stateof emergencyclauses;and (4)
restrictionson emergencysuspensions of constitutionalrights. Using Taylorand
Jodice's(1983) negativesanctionsas a measureof politicalrepression,he analyzes
the impact of these provisionswhile controllingfor level of democracy,domestic
conflict,and economic development.3He finds only three (out of fourteen)statisticallysignificantrelationships:(1) the constitutionalpromiseof freedomof press
reduces the likelihood of negative sanctions;(2) the constitutionalrestrictionof
press increasesthe likelihood of negativesanctions;and (3) a state of emergency
clause produces a negative impact on negative sanctions. While some constitutional provisionsdo demonstratean observableeffect on repression,their impact
is less than constitution-makersprobablywould expect. Davenport'sanalysis is
well executedand his findingsareclearlyrelevantto the analyticalpurposesof this
article.But, as Davenportnotes, his sample of countries is rathersmall and not
fully representativeof the world. In addition,his dependentvariableis arguablya
less severelevel of rightsabuse than the abuse of personalintegrityrightsthat are
under study here. Finally,because of limitationson the availabilityof his dependent variablemeasure, Davenport'sanalysis terminatedwell before the extensive
and intensive development of constitutions that has occurred during the Third
Wave of democratizationand the post-communistera.
Cross's(1999) study assesses the impact of constitutionalprotectionagainst
unreasonablesearchesand of judicial independenceupon human rightsbehavior
3
and political rights.Additionally,as Davenport(1996) points out, she also does not control for the
fact that some constitutions allow rights to be overriddenconstitutionally.
Davenportsummarizesnegative sanctions as "censorship(the limitationof news media) and political restrictions(variousconstraintsplaced upon individual citizens and political parties)"(1996,
636).
115
PoliticalResearch
Quarterly
in 54 to 58 countries. He finds no statisticallysignificantrelationshipwith his
constitutional provisions measures but does find that judicial independence
increases the probabilityof political rights and of the protectionagainst unreasonable searches and seizures. Cross'seffort to analyze the impact of judicial
independence and constitutionalprovisions on human rights is commendable.
However,there are severalaspects of his study that urge caution when drawing
conclusions. First,the two dependentvariablesin the analysesare either an average of each country's Freedom House political rights during the 1980s or
Humana'smeasureof searchand seizure protection.Averagingacrossa decade is
highly questionablein the areaof human rights,where there can be greatfluctuation fromyear to year.Second, his measureof judicial independenceand search
and seizurepracticesare the subjectiveratingof the late CharlesHumana(1992)
that provides no replicable operationalizationof either variable.4Finally, his
study is also limited to small sample of countries that is not necessarilyrepresentativeof the global set of countries.
While overall, the empiricalstudies suggest that constitutionsmay be less
importantthanlegalscholarsand humanrightsactivistswould hope, at leasttwo of
the studiessuggestthatsome elementsof the constitutionsdo, in fact,decreasethe
odds of a stateengagingin politicalrepressionor committinghumansrightsabuse.
CONSTITUTIONALPROVISIONSFOR INDIVIDUALFREEDOM
AND DUE PROCESSRIGHTS
Two types of constitutional protections are expected to affect personal
integrityabuse: (1) the individual freedomsassociatedwith constitutionalliberalism: speech, assembly,association, press, and religion and (2) the key due
process protectionsthat directlyrelateto personalintegrityabuse:writ of habeas
corpus(the prisoner'sright to be promptly informed of the charges),protection
againstarbitraryarrest,prohibitionagainsttortureor cruel and unusual punishment, and a fair and public trial. Both sets of constitutionalprotectionsare key
elements in the internationallyrecognizedminimum standardsfor human rights
and are modeled after the universallyaccepted InternationalBill of Rights discussed above. The overarchingresearchhypothesis for both sets of provisionsis
that they will improvehuman rightsprotectionbecause (1) they arebinding contracts between the regime and its citizens and (2) they are indicators of the
regime'swillingness to protectthese rights or freedoms.The overarchingalternative hypothesis then is that these promisesmay be only empty or unenforceable
promises. The theoreticaljustificationfor each set of constitutionalprovisionsis
somewhat different. The constitutional provisions for individual freedoms
4In fact, it is not clearwhetherthe measurecapturesformalor actualjudicial independence or some
combinationof the two.
116
IndividualHumanRights(1977-1996)
should allow persons to openly criticizeand perhapschallengethe regimeor the
statusquo. Additionally,as Poe and Tate (1994: 123-24) suggest, these types of
freedoms "may make it easier for citizens and opposition leaders to publicize
attempts at repression,therebybringing down on would-be abusive leaders the
weight of majorityor world opinion."
The constitutionalprovisionsfor due processprovideprotectionagainstarbitrary government action that directly relates to the abuse of personal integrity
rights. The writ of habeascorpusprotectsindividualsagainstarbitraryor political
imprisonmentin that the governmentis usuallyrequiredpromptlyto presentevidence sufficientto justify holding a prisonerand to make known the chargesleveled againsta prisoner-obviously precludingthe phenomenonof disappearances
and limiting the practice of political imprisonment.The promise of a fair and
public trialwould make it more difficultfor regimesto hold secrettrialsin which
persons could be convictedfor politicalreasonsor on trumpedup charges.Public
trialsalso facilitatethe disseminationof information,allowingthe pressand public
opinion to function as curbs on repressivebehavior.The ban on torture,when
practiced, directly eliminates one component of personal integrity abuse and
probablydecreasesthe number of extra-judicialkillings that result from torture.
PERSONALINTEGRITYABUSE MODEL
The abuse of the right to personal integrityis a narrowset of human rights
violations that include political imprisonment, torture, and killings or disappearances.While this set of rights does not include all of the rights promoted
under the internationalcovenants and some constitutions, it does focus on the
abuses that are considered to be the most "egregiousand severe crimes against
humanity,"and the ones that representabuses that "arethe sort that usually can
be avoided"(Poe and Tate 1994: 854). And more importantly,this set of rights
covers the core rights that would have to be fulfilled in order for the provision
of the other rights to be meaningful. A group of political scientists have developed and gathered data for standards-basedindices that are believed to be the
acceptable measure for this study: Stohl et al.'s Personal Integrity measure
(Stohl, Carleton, and Johnson 1984; Stohl and Carleton 1985; Carleton and
Stohl 1987; Gibney and Stohl 1988; Henderson 1991, 1993; Poe 1991, 1992;
Gibney,Dalton, and Vockrell1992; Poe and Sirirangsi1993, 1994; Poe and Tate
1994; Poe, Tate, and Keith 1999). These scales measure the abuse of personal
integrity rights on a scale of 1 to 5, with 1 representingstates with the least
amount of abuse and 5 representingstates with the highest level of abuse.5Two
5 The countries are
assigned a ratingaccordingto the following rules from Gastil(1980):
(1) Countries [are]under a secure rule of law, people are not imprisonedfor their views, and torture is rareor exceptional .... Politicalmurdersare extremelyrare.
117
PoliticalResearchQuarterly
sets of scales have been created:one based on State DepartmentCountryReports
on HumanRightsPracticesand one based on Amnesty InternationalReports. In
the Poe and Tatepersonalintegritymodels, parallelanalysishas been conducted
using each of these measures. Overall, the studies of personal integrity abuse
using these two measureshave produced consistent results across the two measures. The analyseshave led to the identificationof a set of eight factorsthat consistently demonstratesa substantiveand statisticallysignificant impact on personal integrity abuse: political democracy,population size, economic standing,
international war, civil war, and regime types-leftist regime and military
regime (Poe and Tate 1994; Poe, Tate, and Keith 1999; Keith 1999; Keith and
Poe 2000; Poe, Tate, Keith, and Lanier2000; Zanger2000). I follow Poe, Tate,
Keithand Lanier(2000) and employ only one of the personalintegrityscales in
this article-the one based on the Amnesty Internationalreports. Focusing on
one measuresimplifies the presentationof results and conserves space, allowing
me to extend the types of analysis conducted in this article.6BecauseAmnesty
Internationalreportstend to leave out countrieswith good human rights behavior, a potentialbias would ariseif I used only countrieson which Amnesty International reported. Therefore, consistent with Poe et al., in cases in which
Amnesty Internationaldid not issue a report but the State Departmentdid, a
score gained from applying the same set of standardsto the U.S. State Department reportswas substituted.
(2) Thereis a limited amount of imprisonmentfor nonviolent activity.However,few persons are
affected,tortureand beating are exceptional .... Politicalmurderis rare.
(3) There is extensive political imprisonment,or a recent history of such imprisonment.Execution or other political murdersand brutalitymay be common. Unlimited detention, with or
without trial, for political views is accepted.
(4) The practices of [level 3] are expanded to larger numbers. Murders,disappearancesare a
common partof life. ... In spite of its generality,onthis level terroraffectsprimarilythose who
interest themselvesin politics or ideas.
(5) The terrorsof [level 4] have been expanded to the whole population.... The leadersof these
societies place no limits on the means or thoroughnesswith which they pursue personal or
ideological goals. (Gastil 1980, quoted in Stohl and Carleton,1985)
6
Examplesof level 5 repressionwould be seen in Afghanistanin 1985 and Colombiain 1991. The
Philippinesin 1990 and Libyain 1987 are examples of level 4 countries. Cubain 1991 and Ghana
in 1982 are examples of level 3 countries. Congo in 1985 and Cameroonin 1987 are examples of
level 2 countries. Benin in 1991 and Oman in 1991 are examples of level 1 countries. Full sets of
examples and excerpts of country reportsfor each level of repressionare availablein Gibney and
Dalton (1996).
Following Poe et al. (2001) I use the Amnesty International-basedmeasure ratherthan the State
Department-basedmeasure.While there has been some concern of a State Departmentbias (i.e.,
anti-socialist,pro-U.S.ally), it should be noted that recent evidence suggests that StateDepartment
reports have become routinized and reliable over time, especially in the post-Cold War period
(Innes 1992, Poe, Vazquez,and Zanger 1998) and a recent article suggests that bias is minimal
(Poe, Vazquez,and Zanger2000).
118
IndividualHumanRights(1977-1996)
Since I am building upon previous work that examined the personal
integritymodels I control for the eight factorsmentioned above that have proven
to be statisticallyand substantivelysignificant.Since these variablesare described
in detail in previous studies, here I only brieflysummarizetheir theoreticallinks
and their operationalization.For fuller details see Poe and Tate(1994) and Poe,
Tate,and Keith (1999).7
PopulationSize:Henderson (1993) argues that states with large populations
may be tempted to "resortto repression as a coping mechanism"to deal with
threats that arise due to severely strained national resources and unfulfilled
public needs. Additionally,Henderson argues that we must statisticallycontrol
for population size since the laws of probabilitywould dictatethat as the number
of persons in a country grows so would the number of opportunitiesfor repressive actions. Recent empirical evidence has supported Henderson'shypothesis
(Poe and Tate 1994; Davenport 1995a; Poe, Tate, and Keith 1999, Keith 1999,
Zanger2000).8 The naturallogarithmof the total national population is used in
the model in order to deal with the skewed distributionof the population data.
EconomicStanding:Expectationsconcerningeconomic standingfollow those
of populationsize. Mitchelland McCormick(1988) and Henderson(1991) argue
that social and political tensions related to economic scarcity are likely to
increaseinstabilityin the poorest countriesand thus increasethe probabilitythat
the regimewould use repressivemeasuresto maintainorder;whereas,in wealthier countries the population will be satisfied and will be less likely to present a
threatto order that would triggerrepressivestate action. Empiricalevidence has
consistentlyshown that higher levels of economic developmentreduce the probability of political repression(Mitchelland McCormick1988; Davenport 1995;
Poe and Tate 1994; Poe, Tate,and Keith 1999; Keith 1999; Poe, Tate,Keith,and
Lanier2000; Zanger2000). Economic standing is operationalizedas the state's
per capita GNP (in dollar thousands).9
Civil WarExperience:A growing body of literaturehas demonstratedthat
governments faced with internal threats often resort to political repression to
restoreorder.Although the literatureon domestic threatshas mainly focused on
7 These are the same data
reportedin Poe, Tate,and Keith 1999 except that the authorhas extended
the datasetbeyond 1993 to 1996 and has replacedthe PolityIIImeasurewith the Polity98 measure.
8 Henderson also
hypothesized that growth in population size would increase the likelihood of
repression.While he did find evidence of this effect in his single-yearstudy,those resultshave not
been found to hold up in fuller studies of human rights abuse. Ratherwe have found population
size to be the significantfactor,both statisticallyand substantively
9 Despite some criticismof nationalGNP data (Summersand Heston 1984; Heston 1994), GNP has
continued to dominateas the most appropriatemeasureof economic development(Donnelly 1999,
623). For example, see Dahl 1992, Moaddel 1994, Cross 1999, Milner,Poe, and Leblang 1999.
Some studies have substitutedenergyconsumptionfor this measure(Henderson1991, 1993; Davenport 1996, 1998), but these data are not availableconsistentlyacrossa global set of countries.
119
PoliticalResearch
Quarterly
domestic threator conflict as a dependent variable(e.g. Gurr 1968, 1970; Feierabend and Feierabend1972; Gurrand Duvall 1973; Jenkins and Perrow 1977;
Tilly 1978; Lichbachand Gurr 1981; McAdam1982; Booth and Richard1996;
and Francisco 1995, 1996), recent studies have demonstrated that domestic
threat does increase the probabilityof state repression (e.g., Davis and Ward
1990; Alfatooni and Allen 1991; Poe and Tate 1994; Davenport 1995, 1996;
Krain 1997, 2000; Moore 1998; Poe, Tate, and Keith 1999; Keith 1999; Poe,
Tate,Keith,and Lanier2000). Civil war, which poses the most serious domestic
threat, is defined here following Small and Singer'sguidelines for identifying
instancesof civil war: (1) "thegovernment,as the centralauthorityin a country,
must be involved as a directparticipantin the war"and (2) "theremust be effective resistance,that is, either both sides must be 'organizedfor violent conflict"
or "theweakerside, although initiallyunprepared[mustbe] able to inflict upon
the strongeropponents at least five percentof the numberof fatalitiesis sustains"
(Small and Singer 1982: 215).
InternationalWarExperience:
Internationalwar is yet another serious threat
which may compel regimesto resortto political repressionas a tool to maintain
domestic order during such a state of emergency (see Gurr 1986). A growing
body of empiricalevidence has supportedthis hypothesis (Rasler1986; Poe and
Tate 1994; Poe, Tate,and Keith 1999; Keith 1999; Zanger2000). International
war is operationalizedfollowing Small and Singer's(1982: 50-55) guidelines for
where an internationalwar is one in which "(1) there was a total of a thousand
or more battle deaths sufferedby all of the participantsin the conflict, [and] (2)
the particularcountry sufferedat least a hundred fatalitiesor had a thousand or
more personnel taking part in the hostilities"(1982: 50, 55).
BritishColonialExperience:Mitchell and McCormickargue that the colonial
experiences that shaped the political culture of most states may have impacted
the states'respect for human rights. In particular,they note that Britishcolonial
rule is strongly associated with the development of postcolonial democracies;
whereas, other colonial experiences that presumablywere more authoritarian
may have left a legacy of greaterhuman rightsabuse (1988: 480). While Mitchell
and McCormickfound only slight evidence to support their hypothesis and Poe
and Tate (1994) found no evidence to support the hypothesis, Poe, Tate, and
Keith(1999) did find support for the hypothesis in their expandedstudy.Countries that have been territoriesof GreatBritainat some point during their history
are coded 1 and all other countriesare coded 0.
MilitaryRegime:As Poe and Tate(1994: 853) note, the role of the militaryis
one of the factorsfirstexploredby those interestedin human rights-relatedphenomena . Militaryregimesare believed to be more likely to resortto repression,
because armiesby their natureand by habit are preparedto use force as a means
of controlor coercion(e.g., Huntington 1964; McKinlayand Cohan 1975; Zwick
1984; Ziegenhagen 1986; Seligson 1987; Poe and Tate 1994; and Davenport
120
IndividualHumanRights(1977-1996)
1995a). Empiricalevidence of the impact of military regimes on human rights
behaviorhas been somewhatweak. In personalintegritystudies the resultshave
been ratherinconsistentcomparedto other factors.Poe and Tate(1994) and Poe,
Tate, and Keith (1999) examined the impact of military controlled regimes on
personalintegrityabuse. In these studies militarycontrolledregimesare defined
as either regimes that come to power "as a consequence of a successful coup
d'etat,led by the army,navy,or air force, that remainedin power with a military
person as the chief executive for at least six months in a given year"(McKinlay
and Cohan 1975: 1) or regimes "witheither a civilian as the chief executive and
severalmilitarypersons in the cabinet or militaryhead of governmentwho nominated a civilian as the head of government and himself worked behind the
scenes"(Madani1992: 61). Militarycontrolwas not statisticallysignificantin Poe
and Tate'sshorter study (1994) but was in their extended study (Poe, Tate, and
Keith 1999)." Davenport(1995a), who examined a differentform of repression,
negativesanctions,used three measuresof militaryinfluence:size of the military,
militarysector allocations, and direct representationof the militaryin the government.10Only militarysector allocationsdemonstrateda consistentstatistically
significantrelationshipwith repressionacrossall of his models.1 Here, I continue
to employ the Poe and Tatemeasureof militaryregime as defined above.
Originally,politicalscientistshypothesizedthatMarxist-Leninist
LeftistRegime:
controlled states would be more willing to use repressionto curb threatssince
their political ideology justified the dominationof the polity in the pursuit of an
ultimate political goal (Mitchelland McCormick1988; Poe and Tate 1994; Poe,
Tate,and Keith 1999). Initialglobal analysissupportedthis hypothesis, but only
in the case of abuse as reportedby the StateDepartment,a result that seemed to
suggest a possible bias in StateDepartmentreportsratherthan a true effect (Poe
and Tate 1994). However, when the global analysis was expanded beyond the
initial eight-yearperiod to a period of eighteen years, the evidence clearly contradictedexpectationsand suggested that leftist regimeswere actuallyless likely
to represspersonalintegrityrights than nonleftist regimes.While this result was
unexpected, it is not totally counter to the originaltheoreticalexpectation.First,
in leftist regimes, control of society and personal freedoms have often been so
complete that the regime might be less likely to need to engage in these more
severe abuses of personal integrity rights to maintain order than would be its
nonleftist counterparts.Second, as Duvall and Stohl (1983) and Lopez and Stohl
(1992) have argued, human rights repression may have an "'afterlife,'which
affects the behavior of people long after the observableuse of coercion by state
10 The Banks (1991) data are not availableacross the
global set of countries or the complete time
here.
under
study
period
Davenport'ssubsequentstudies have either not continued to control for militaryinfluence or have
not reportedthose results (Davenport 1996, 1998, and 1999).
121
PoliticalResearch
Quarterly
agents has ended"(Lopez and Stohl 1992: 218). Thus, past repressionin leftist
regimesmay actuallyreduce the need for futurerepressionor the need for more
severe forms of repression,such as those measuredby personal integrityrights
abuse. Of course, neither of these explanations is exclusive to leftist regimes.
Leftistregime is operationalizedas "thosegovernedby a socialist party or coalition that does not allow effectiveelectoralcompetition with non-socialist opposition"(Poe and Tate 1994: 858). Leftistregimeis a dichotomous variablewhere
leftist regimes are coded one and nonleftist regimes are coded zero. Because of
the conflicting expectations for this measure, it is assessed in this article at the
most strenuoustwo-tailedtest of significance.
PoliticalDemocracy:Democracy has consistently been shown to strongly
reduce the likelihood of various forms of political repression(for example, Diamond, Linz, and Lipset 1988 Henderson 1991; Fein 1995; and Davenport
1995a, 1995b, 1996 and 1998) including the abuse of personal integrityrights
(Poe and Tate 1994; Poe, Tate,and Keith 1999; Poe, Tate,Keithand Lanier2000;
Keith and Poe 2000; and Zanger2000). Theoreticalexplanationsof the nexus
between democracyand repressionhave tended to focus either on institutional
and structuralcharacteristicsand/or the values and norms in democracies.Political scientists argue that democraticprinciples emphasize bargaining,compromise, and elections as the appropriateor only fairmeans to resolvedisagreements
(e.g., Gurr 1986; Henderson 1991; Poe and Tate 1994 and 1995; Dixon 1994;
Rummel 1997). These norms socialize democratic leaders to resolve conflict
throughnon-violentmeans. Politicalinstitutionsreinforcethese norms, but more
importantly,scholars argue that political institutions provide the tools for the
public to hold government officials accountable for their actions (e.g., Gurr
1986; Schmitterand Karl1991; Poe and Tate 1994; Dixon 1994; Rummel 1997;
Zanger2000). For example, with fully participatoryand competitiveelections, a
potentiallyabusive leader might feel vulnerableto public discontent at the polls
and thus be curbed from abusivepractices.Or a potentiallyabusiveleadercould
be curbed by a system of checks and balances that place judicial or legislative
constraintson the executive'spowers.
When studyingthe relationshipbetween human rightsand democracypolitical scientistsmust carefullydefine and measuredemocracyin orderto preventa
tautologyin describingthis relationship.12 Poe and Tate(1994: 856) suggest that
in order for democracy to "function as an independent explanation for state
12
Some scholarshave expressed concerns that some elements of democracy,elections in particular,
may in certain circumstancesbe positively relatedto political repression(Hermanand Brodhead
1984; Beckett 1987; Haywardand Kandeh 1987; Weiner and Ozbudun 1987; and Booth and
Seligson 1989). Recentcross-nationalstudies have not supportedthis concern. Davenport(1998)
found that national elections tend to decreasenegative sanctions, regardlessof the political-economic context but Richards(1999) found no effect of elections on repression.
122
IndividualHuman Rights(1977-1996)
terrorismand the abuse of personalintegrity,it must be defined in terms or procedures and rights that do not themselves preclude repression"and "must be
defined in terms that allow independent operationalizationof the concept."Following Poe and Tate(1994) and other studies of personalintegrityabuse, I use a
definition of political democracy that primarilycaptures the institutional and
structuralaspects of democracythat are most appropriatehere. Politicaldemocracyis definedas "asystem of governancein which rulersareheld accountablefor
their actions in the public realmby citizens acting indirectlythroughthe competition and cooperationof their elected officials"(Schmitterand Karl1991: 76).13
Three measures of democracy have been used in personal integrity rights
models. In Poe and Tate's(1994) first study they employed Vanhanen's(1990)
index based on voter turnout and party competition. For cross-validity they
conducted a parallel analysis with Freedom House's political rights measure.
In their 1994 study Poe and Tate had noted that the Polity II (Gurr 1990)
measure of institutionalized democracy would have been their ideal measure
of democracy.They were unable to use the measure, however, because it had
not yet been expanded to encompass the entire time period covered in their
study. In their extended study, Poe et al. (1999) were able to switch to the preferred Polity measure of democracy,using Polity III (Jaggersand Gurr 1995).
Again, they conducted a parallel analysis with the Freedom House political
rights index. Here, I use the most recent Polity measure of democracy,Polity
13
There may be some concern that these constitutionalprovisions, which representconstitutional
liberalism,overlap with the concept (and my measure) of democracy In post Cold War period,
scholarshave been particularlyinterestedin disentanglingthe concepts of (and expectationsconcerning) constitutionalliberalismand political democracy Schmitter(1992) assertsthat while the
concept of constitutionalliberalismmay have coincidedwith the rise of democracy,"it has never
been immutablyor unambiguouslylinked to its practice."Zakaria(1997) distinguishesclearlythe
two concepts. He arguesthat the essence of democracyhas always been rule of the people while
the essence of constitutional liberalism has always been the natural (or 'inalienable')rights of
human beings secured againstgovernmentintrusion under basic law (26). He arguesthat constitutional liberalism". . . is not about the proceduresof selecting government,but rathergovernment'sgoals. It refersto the tradition,deep in Westernhistory,that seeks to protectan individual's
autonomyand dignityagainstcoercion,whateverthe sourcestate,church, or society"(25). Zakaria
furtherargues against defining democracy so broadly as to include an exhaustive list of social,
political, economic, and religiousrights,because such elevation "turnsdemocracyinto a badge of
honor ratherthan a descriptivecategory"and "rendersit analyticallyuseless"(25). As evidence of
this dichotomy,he points to the rising number of democraciesin the world today that in fact do
not practice constitutional liberalism but rather are what he calls "illiberaldemocracies."Of
course, it remainspossible that political democracywould be highly correlatedwith constitutional
liberalism,although, Zakaria'sand Huntington'sanalyses of democraciesand actual liberalismin
1990s certainlycast doubts on the probabilityof such a correlation.It should also be remembered
that my study measuresprotectionsthat are written in documents as opposed to the actual fulfillment of these protections. Statisticaldiagnostics, in fact, demonstratedonly minor correlation
between the individual constitutionalmeasuresand the democracymeasureused here.
123
PoliticalResearchQuarterly
98.14 The choice of the Polity measure is also consistent with other recent
studies of personal integrity abuses (Milner, Poe, and Leblang 1999; Zanger
2000) and negative sanctions (Davenport 1995b, 1995c, 1996, and 1998).
Polity's institutional democracy indicator is an 11-point additive index
coded along four dimensions using the following rules:
*
*
*
*
Competitivenessof PoliticalParticipation:competitive (3), transitional(2),
and factional(1);
Competitivenessof ExecutiveRecruitment:elective (2) and transitional(1);
Openness of Executive Recruitment:open election (1) or dual (hereditary
and election) (1); and
Constrainton Chief Executive:executive parityor subordinationto legislative or judicial branches (4), intermediateconstraints(constraintsthat fall
between parity/subordinationand substantial limitations) (3), substantial
limitations (2) and intermediateconstraints (constraintsthat fall between
substantiallimitationsand slight to moderatelimitations)(1).
A recent study (Gleditsch and Ward 1997: 380) examined empiricallythe
components of the Polity measure(Polity III). Their analysesdemonstratedthat
the measure is primarilydriven by its executive constraintdimension and that
the other dimensions, which measurepatternsof executive recruitmentand the
extent and competitiveness of participationdimensions, were "not especially
powerful in determiningthe degree of democracy"They concluded that to use
the data "summarilyto classifymodern polities as democraciesdirects attention
away from the actualdata that have been collected on authoritypatterns."Thus,
they urge scholars to move beyond discrete classification of "democracy"or
"autocracy"and to focus more on the subdimensions of this measure. Personal
integritystudies do not use the Polity data as a dependent variableand not do
use the data to divide countries into two discrete categories,but ratheruse the
index to measure the level of institutional democracy attained in the country.
While other human rights studies have continued to employ the composite
measure ratherthan the individual components (e.g., Davenport 1998; Milner,
14
In order to simplify and facilitatethe presentationof the analysis I have chosen to use only one
measure of democracyfor this study I have chosen to drop the Freedom House measurein this
analysisfor severalreasons.It has been criticizedfor being "highlyimpressionistic,being no more
than an estimateby a person who has collected a lot of seemingly relevantinformationon all of
the countries of the world"(McCamant1981, 132). While the measurementhas improved over
time and is considered,less impressionistic(see Poe and Tate 1994), the measuredoes include one
dimension that may overlap with the behavior capturedin the dependent variable(see Poe and
Tate 1994). Additionally,the measure is counter-intuitivein that the small score represents a
higher a level of political rights and the higher score a lower level of rights. The results produced
with this measurehave performednearlyidenticallyto both the Vanhanenmeasureand the Polity
measure.I did run the full model with the FreedomHouse measureand the resultsgenerallyheld.
124
IndividualHumanRights(1977-1996)
Poe, and Leblang 1999; Poe, Tate, and Keith 1999; Zanger2000), I decided to
conduct an additionalparallelanalysis to examine, as Gleditsch and Wardsuggest, the effects of the subdimensionsin my model.
Variables
Promisesand Restrictions
Constitutional
The key constitutional liberties and protections include: (1) freedom of
speech; (2) freedomof association"(3) freedomof assembly;(4) freedomof press:
(5) freedomof religion;(6) the rightto strike;(7) writ of habeascorpus(definedas
the requirementthat governmentpromptly present evidence before a judge or
magistratesufficientto justify holding a prisonerand to make known the crimes
with which a prisoneris being charged);(8) public trial;(9) fairtrial;and (10) ban
of tortureor cruel and usual punishment.I have coded the level of constitutional
provisionas: explicit guaranteeor mention in the constitution2; explicit guarantee or mention in the constitutionbut notes exceptions or qualifications,such as
a public interestclause 1; no mention in the constitution0. 5 This operationalization is similarto thatBlasiand Cingranell(1996) who measure"absolute"or "conditional" constitutional guarantees.16The data sources for these variables are
Blaustein(1993) and Flanz's(1997) Constitutions
of Countriesof theWorld.Flanz's
series containsEnglishtranslationsof most currentconstitutionsof the world and
in most cases contains informationon past constitutionsas well.17
METHODOLOGICAL
ISSUES
I have built a data set that is pooled acrossboth time (1976-1996) and space
(between 154 and 178 nation-states,dependingupon the year).Politicalscientists
15
When coding the constitutions I found nearly a dozen constitutions that restrictedthe constitutional protection in such a way that the provision would not apply to political opposition. For
example, in Algeria(1977-81) the promise of freedom of speech and assemblydoes not apply to
speech/assemblyagainstthe revolutionarycause and in Albaniathe promise of freedomof speech,
association,assembly,and press do not apply if there are in opposition to the socialistorder.Since
these clauses go well beyond public interest clauses, I have originallycoded the variablesas 0. I
have flaggedthe variablesso that I may adjust these codings as is necessaryduring my analysis.
16 I chose not to follow Davenport'sstrategyof creating for each freedom a separatedichotomous
variableto denote constitutional"promises"and constitutionalrestrictionsbecause it is likely that
these two measuresshare variance,thus reducing the statisticalsignificanceof the variables.
17 Therehas been some
difficultyin accessing all of the old reportsin this series. This publicationis
organized as series of looseleaf binders in which the outdated materialsare removed from the
looseleaf volumes and replacedby updated pages. Oceana Publicationsdoes not keep old copies
and only one U.S. universityhad the foresightto keep the older copies. I have been able to borrow
the volumes that Bryn Mawr had kept. Because they did not start their collection until around
1984, I was not able to get all old constitutions going back to 1976. ThereforeI do have some
missing years of constitutions, but relativelyfew, especially given that my data set contains 3738
country-years.The early years of the data set (1976-79) contain the most missing data on the
dependent variable.
125
PoliticalResearchQuarterly
such as Stimson (1985: 916) have noted that this design "canbe an extraordinarily robust researchdesign, allowing for the study of causal dynamics across
multiple cases, where the potential cause may even appearat differenttimes in
differentcases"(p. 916). The pooled cross-sectionaldesign eliminates the small
N problem of many comparativestudies, producing in this case a data set with
an N of 2,552 country-years.Twostatisticalproblemsare inherentin this particular design-the threatof heteroscedasticityfrom the unit effects and the threat
of autocorrelationfromthe time dimension, both of which may lead to the problem of unreliabletests of statisticalsignificanceand inferences(see Stimson 1985;
Ostrom 1990; Beck and Katz 1995). Following Poe and Tate (1994) and Poe,
Tate, and Keith (1999), I deal with the problem of heteroscedasticityby using
Beckand Katz'sPanelRobustStandardErrors(Beckand Katz 1995). Currentstatistical software (StataCorp1999:362) allows me to compute the Panel Robust
StandardErrorsusing an unbalanced data set (StataCorp1997: 619). To deal
with the problem of autocorrelation,I have also included a lagged dependent
variableto correctfor autocorrelation(Beck and Katz 1995).18
ANALYSISAND RESULTS
Preliminaryanalysisindicatedthat multicollinearitycould be problematicin
regardto four of the constitutionalprovisionsthefreedoms of speech, religion,
association, and assembly.19I tested a variety of approachesto amelioratethis
problemand ultimatelychose to deal with it by constructingan index that combined the four problematicvariablesinto an additive index, the Four Freedoms
Index. Since initial bivariateand multivariateanalysissuggested that some of the
relationshipsof the constitutionalprovisionswere signed in opposite directions,
18See Beck and Katz(1995) and Beck et al. (1993) for a more
complete descriptionand justification
of this approach.The inclusion of this variableis not only statisticallyjustified but also theoreticallyjustified since it has been shown that regimestend to use past decisions as a baseline for present decisions (Wildavsky1984).
19 Analysisof the correlationbetween each pair of variablessuggested that four of the variablesmay
be problematic:speech, assembly,association, and religion. Freedom of speech and freedom of
assembly are correlatedat .81 which is beyond the .80 limit set by Lewis-Beck(1980, 58-62).
Associationand assembly are correlatedat .77 and speech and association are correlatedat .77.
Religion and assembly are correlatedat .63 and religion and association are correlatedat .68.
While these pairs do not surpass Lewis-Beck's.80 cutoff, the correlationis high enough to merit
some concern;therefore,I ran Kleintests (each independentvariableis regressedupon the others)
on each of these variables.While the R-squareof each regressionis well below 1.0 as suggested
by Lewis-Beck, the r-square for speech on assembly is .66, which suggests that high multicollinearitycould still be a problem. The tests for the relationshipbetween religion and assembly
and between religion and associationare low enough to suggest that multicollinearitymay not be
a problem(r-squaresrangingfrom .39 to .45). The relationshipbetween associationand assembly
and between speech and associationfell into a fuzzy zone with r-squares,rangingfrom .59 to .60.
126
IndividualHuman Rights(1977-1996)
I was reluctantto construct a single additive index that combined the full set of
ten constitutionalvariables.This caution was justified in tests of such an additive
ten-variableindex; the index produced coefficientsthat were signed in the unexpected direction. Therefore,I ruled out the use of the full ten-variableindex. I
decided that a more acceptablesolution to the multicollinearityproblem would
be to construct a limited index that combined and isolated only the four problematic variables,which individuallyhad produced negative coefficientsin preliminary analysis. Such an index would not mask as many relationshipsas the
ten-variableindex did.
The results of the analysisfrom the full model are reportedin Table1. None
of the variablesdenoting the six individualconstitutionalfreedomsis statistically
significant.The Four Freedomsindex is signed properly,producinga small coefficient (-.01), but is statisticallysignificantonly at the .22 level. The other individual freedomprovisions (freedomof the press and the right to strike) produce
coefficientsthat are signed in the wrong directionand thus cannot be considered
statisticallysignificant.The due process variablesfarebetter than the individual
freedoms measures;two of the four due process provisions are statisticallysignificant.The public trialprovisionproduces a rathersmall coefficient(-.03) that
is statisticallysignificantat the .04 level. Thus, a change from a constitutionthat
does not include any provisionfor public trialsto one that includes qualifiedprovision or full provision of public trial would lead to an initial change of .03 or
.06 respectivelyin the level of personalintegrity.The fairtrialprovisionproduces
a larger coefficient (-.12) that is statisticallysignificant at the .0001 level. A
change from a constitution that does not include any provision for fair trials to
one that includes qualifiedprovision or full provision of fair trial would lead to
an initial change of .12 or .24 respectivelyin the level of personalintegrity.The
other two due process provisionsproduce coefficientsthat are signed incorrectly
and cannot be consideredstatisticallysignificant.20
20
I continued to explore other possibilities. I conducted an additional analysis that controlled for
overly restrictiveconstitutionalprovisions.As I had coded the constitutions it became clear that
some constitutionalpromises were restrictedwell beyond the standard"publicwelfare"caveats
that increasinglyappearin constitutions.A small set of constitutionsdenied using these freedoms
against the revolutionaryor socialist order/interest.For example, Albania'scommunist constitution says that the freedoms may not be exercised in opposition to the socialist order, and Iraq's
constitution says that the exercise of freedomsmust comply with the revolutionarytrend. These
restrictionsmay be so extensive that they render the promises useless or they may actuallylegitimize the abuse of such rights. To control for this possibility I createda dummy variableto mark
these country years (0 if the constitution contained one of the above clauses), and then I used a
dummy variableas a multiplierfor these countries, which convertedthe variablesto zeros. I also
createda multiplierwhere the overlyrestrictiveconstitutionswere coded (1) insteadof zero. I then
duplicated each of the above analyses and produced results that were almost identical to previously reportedresults. Thus, the impactof these more restrictiveclauses is not readilyobservable.
I examined one additionalpossibility:that there was not enough variationin the 10 measuresof
127
PoliticalResearchQuarterly
1.
TABLE
AND
FORINDIVIDUAL
FREEDOMS
PROVISIONS
THE IMPACTOF CONSTITUTIONAL
RIGHTS(1977-96)
DUE PROCESSON THEABUSEOF PERSONALINTEGRITY
withcompositePolity98
democracymeasure
Variable
Coef*
LaggedDependent
.64 (.04)
Variable
-.01 (.01)
Four FreedomsIndex
.05 (.03)
Freedomof Press
.03 (.03)
to
Strike
Right
.06 (.02)
HabeasCorpus
-.03 (.02)
Public Trial
-.12 (.04)
FairTrial
.04 (.02)
Torture
.48 (.07)
Civil War
.20 (.06)
InternationalWar
PoliticalDemocracy
-.04 (.01)
(Polity 98)
Constrainton Chief
Executive
Openness of Executive
Recruitment
Competitivenessof
ExecutiveRecruitment
Competitivenessof Political
Participation
.02 (.04)
MilitaryRegime
.11 (.05)
LeftistRegime
BritishColonial
.03 (.03)
Experience
.01 (.00)
Economic Standing
.07 (.01)
Population
.14 (.13)
Constant
N = 2552
z
p
withseparatedcomponentsof
Polity98 democracymeasure
Coef*
z
14.70
-.77
2.03
.93
3.28
-1.76
-3.21
2.33
7.10
3.64
.63 (.04) 14.31
.0001
.22
-.001 (.01) -.26
.04 (.02)
1.66
.02
.77
.02 (.03)
.16
.07 (.02) 3.47
.0001
.04
-.03 (.02) -1.70
.0001 -.11 (.03) -3.17
.03 (.02) 2.15
.01
.47 (.07) 7.09
.0001
.21 (.06) 3.66
.0001
-5.88
.0001
-.03
(.01) -1.88
p
.0001
.40
.05
.22
.0001
.04
.001
.02
.0001
.0001
.03
.01 (.01)
.73
.23
.04 (.03)
1.16
.12
.56
2.22
.29
.02
-.09 (.02) -4.30
1.28
.05 (.04)
-.14 (.05) -2.55
.0001
.11
.01
1.00
4.66
6.50
1.14
.17
.0001
.0001
.13
.01
-.01
.07
-.04
.40
(.04)
(.00) -4.56
(.01) 6.62
(.12) -.38
.35
.0001
.0001
.35
R2 = .73
Chi2 = 9149.98 p > .00001
R2 = .73
Chi2 =9885.33 p > .00001
*Mainentries are unstandardizedOLScoefficients,generatedusing STATA6.0. The robust standard
errors,which were used to control heteroscedasticityare in parentheses.
128
IndividualHumanRights(1977-1996)
Overall,the controlvariablesperformas they have in previousstudies. Each
of the controlvariablesis statisticallysignificantat the .0001 level, except for military regime and British colonial experience, which do not achieve acceptable
levels of statisticalsignificanceonce the constitutionalvariablesare added to the
model. These results are not surprising since these two variables performed
weakly and somewhat inconsistentlyin previous studies as well. Civil war produces the second largestimpact in the model, an increase of .48 in the level of
personal integrity abuse when a country moves into a period of civil war. The
impact of a change to internationalwar producesa smallerchange than does civil
war, an increase of .20 in the level of personal integrity abuse. As in previous
models, economic standingproducesa small positive impact (coefficientof-.01)
that is basicallyequal to zero when one considers that the measureis computed
in dollar thousands.For example, it would take an increasein per capitaGNP of
$20,000 to produce a .20 decreasein the level of repression.Politicaldemocracy
produces a seemingly small coefficient (-.04), but an increase from the lowest
level of democracy to the highest level of democracywould produce an initial
impact of -.40 in the level of personalintegrityabuse.
In deferenceto the concerns raisedby Wardand Gleditsch(1997) that were
discussed earlier,I have included a test of the disaggregatedmeasureof democracy Those resultsarereportedin the second half of Table1. The substantiveand
statisticalperformanceof the constitutionalvariablesand the controlmeasuresis
identicalto that achievedin the analysisusing the composite measure.Twoof the
four components of the democracymeasureproduce statisticallysignificantrelationships: constrainton the chief executive and the competitivenessof political
participation.21Executive constraint produces a coefficient of -.03 (p < .03)
which means that a one unit change in executive constraintwould producea .03
change in the level of human rights abuse. However, a maximum change from
unlimited executive power to executive parityor subordinationto the legislative
or judicial brancheswould produce a decreaseof .12 in the level of abuse. While
Ward and Gleditsch'sanalysis revealed that the Polity democracymeasure was
largely driven by executive constraint, in this analysis another component is
much more powerful-the competitivenessof political participation(coefficient
of .09 p < .0001). A maximum change in this variable from zero (suppressed
political participation)to three (competitiveelections) would produce a change
of-.27 in the level of repression.The two components that capturethe openness
and competitivenessof executive recruitmentare not statisticallysignificant.Not
constitutionalprovisionsto merit more than a dichotomous measurefor each provision.I recoded
the data where all 2s were coded 1. I achieved the same results with these dichtomous variables
and thus chose to reportthe results using my originalcoding.
21 Identicalresults were produced in the Poe-Tatemodel with the constitutionalvariablesremoved
from the model.
129
PoliticalResearch
Quarterly
surprisingly,the combined maximum impact of the two components (.39) is
approximatelyequal to that achieved by the composite measure (.40) achieved
in the firstanalysis.Regardlessof the constructionof the democracymeasure,its
impact is one of the largestin the model.
I conducted one final set of analyses to include interactionvariablesin the
model. Davenport'sstudy had found importantinteractionsbetween some of his
constitutionalprovisionsand conflict. He found that when nationalconstitutions
explicitlyprotectfreedomof press and domestic threatsoccur, the state will generally engage in lower levels of repressionthan those states that face domestic
threatswithout the explicit constitutionalprovision for this freedom. The interaction between the otherindividualfreedomsand threatwere not statisticallysignificant. Following Davenport,I tested for interactionbetween the two levels of
conflict included in my model (civil war and internationalwar) and the constitutional provisions. The results of these analysesare included in Table2. I first
tested the interactiveeffects with the civil war measure.These results appearin
the firsthalf of the table. The public trialand fairtrialprovisionsremainstatistically significanteven when controllingfor the interactiveeffects. The coefficient
for fairtrial remainsthe same (-.12) but the coefficientfor public trialincreases
somewhat(from-.03 to -.05). Only one of the interactivemeasuresachievesstatistical significance-the interaction between freedom of press and civil war.
When a state has a national constitution that provides for freedom of press and
the state experiencescivil war, the state is less likely to abuse human rights than
a state experiencing civil war that has no such constitutional provision. Even
though the level of conflict analyzed here (civil war) representsa more severe
threatto the regime than does the level of conflict measuredby Davenport(generalstrikes,antigovernmentdemonstrations,guerillawarfare,and riots), the relationship observedby Davenportholds. The resultsof the interactionvariablesfor
internationalwar are reportedin the second half of the table.None of these interaction variablesis statisticallysignificant.22Thus, it appearsthatwhen the regime
is threatened by conflict from outside its borders, the regime'ssurvival may
trump its constitutionalcommitments.
OVER-TIME EFFECTS OF THE INDEPENDENT VARIABLES
Becauseof the inclusion of the lagged dependent variablein my model, the
independent variablesin the model are expected to produce, through that variable, an additionalimpact on human rightsbehaviorfor severalyears. In Figure
1, I depict the overtime effect (from timet to timet_0) of substantialchanges in
22
These resultshold when all of the interactivevariablesare combined into a single model. Only the
interactionbetween freedomof press and civil war is statisticallysignificant.The impact of public
and fair trialremainsconsistent with the other models.
130
IndividualHumanRights(1977-1996)
TABLE2.
BETWEENCONSTITUTIONAL
PROVISIONS
AND
RELATIONSHIPS
MODELSOF INTERACTIVE
CONFLICT
(1977-96)
interaction with
international war
interaction with civil war
Variable
LaggedDependent
Variable
Four FreedomsIndex
Freedom of Press
Right to Strike
HabeasCorpus
Public Trial
FairTrial
Torture
Four FreedomsIndex
X Conflict
Freedom of Press X
Conflict
Right to Strike X
Conflict
z
Coef*
.63
-.01
.06
.02
.07
-.05
-.12
.04
p
Coef*
.0001 .64
.21
-.01
.05
.01
.03
.26
.002
.0
.01
-.03
.0001 -.11
.01
.04
z
p
14.70
-.67
1.93
.93
3.11
-1.85
-3.13
2.06
.0001
.25
.03
.18
.001
.04
.001
.02
(.04)
(.01)
(.03)
(.03)
(.02)
(.02)
(.04)
(.02)
14.55
-.81
2.24
.66
2.93
-2.54
-3.29
2.62
.02 (.02)
1.14
.13
(.11) -1.45
.07
.001 (.13)
.01
.50
-.16
(.04)
(.01)
(.03)
(.03)
(.02)
(.02)
(.04)
(.02)
-.003 (.04) -.08
.47
.01 (.15)
.04
.49
.03 (.26)
.11
.46
Conflict
.01 (.07)
Public TrialX
Conflict
.21 (.13)
FairTrialX Conflict
.01 (.24)
TortureX Conflict
.004 (.05)
.40 (.10)
Civil War
.20 (.06)
InternationalWar
PoliticalDemocracy
-.04 (.01)
(Polity 98)
.03 (.04)
MilitaryRegime
LeftistRegime
-.10 (.05)
BritishColonial
.04 (.03)
Experience
Economic Standing -.01 (.00)
.07 (.01)
Population
Constant
.16 (.12)
.11
.46
.07 (.13)
.58
.28
Habeas Corpus X
N = 2552
1.65
.04
.07
3.91
3.91
.05
.06
.49
-.08
.01
.47
.0001 .48
.0001 .18
.47
(.12)
(.15) -.56
.15
(.07)
7.00
(.07)
(.09) 1.95
.33
.29
.44
.0001
.03
-5.93
.67
-2.05
.0001 -.04 (.01) -5.85
.25
.02 (.04)
.58
.04
-.11 (.05) -2.20
.0001
.29
.03
1.16
-4.73
6.79
-1.29
.12
.03
.0001 -.01
.0001 .07
.10
-.14
.17
.0001
.0001
.13
R2 = .73
(.04) 1.00
(.00) -4.66
(.01) 6.50
(.12) -1.11
Chi2 = 20386.51
R2 = .73
Chi2 = 12814.36
p > .00001
p > .00001
*Mainentries are unstandardizedOLScoefficients,generatedusing STATA6.0. The robust standard
errors,which were used to control heteroscedasticityare in parentheses.
131
PoliticalResearchQuarterly
FIGURE
1.
OVERTIME
EFFECTS
0.00
0.00
,
.
,
_t+1
t+2
.
,
t+3
.
t+5
t+4
.
t+6
,
t+7
,
,
t+8
t+9
t+10
-0.20
-0.40
-0.60
-1.40
*~--Civil
-.0Democracy War
-1NP (1000s)
pc
or Internatonal War
Population (log)
-1.60
Year
--Public
Trial
4--Fair Trial
-U--Civil War
-4--Democracy
-*--GNP pc (1000s) or Intarnational War
.'! Population (log)
each of the independent variablesin the model that achieved statisticalsignificance at least at the .10 level. With the exception of two variables (per capita
GNPand population)maximumchange in the independentvariablefromtimet_l
to timetis assumed-for example, a change frombeing the highest to the lowest
level of democracy23In the case of per capita GNP a change of $20,000 dollars
is assumed and in the case of population a change of ten million people is
assumed. For ease of comparison,the figureassumes that each of the independent variablescauses a decreasein human rights abuse.
In Figure 1 we can see that the over-time effects of each of the variables
become asymptotic by the tenth lag. A change from a constitution with no
23 The effect at
timetis the coefficientmultiplied by the maximum variationpossible. For example,
the maximum amount the public trialvariablecan vary is twogoing from a score of 0 for no protection to a score of 2 for maximum protection.The effect of the changes at time +1is calculated
by multiplying effect of the change at time, by the coefficient for the lagged dependent variable
and adding the effect of independent variableat time +l.
132
IndividualHumanRights(1977-1996)
provision for public trial to a constitution with explicit provisions for fair trials
will produce, ceterisparibus,an initial impact of .06, that levels off at .17 by the
tenth lag. The impact of adding the fairtrialsprovisionto the constitutionis substantiallylarger,a change of .66 by the tenth lag. If we look at the combined
impact of the two constitutionalprovisionswe see an over-timeimprovementin
human rights of almost a full level (-.83). The impact of terminatinga civil war
(and staying at peace) reduces personal integrity abuse by .48 initially and by
1.32 at the tenth lag-the largestover-timeimpact. Terminationof involvement
in an internationalwar produces a much smaller impact, beginning at .20 and
leveling off at .55 by the tenth lag. Democracyproduces the thirdlargestimpact;
a country that progressesfrom a score of zero to a score of ten on the democracy
index would eventually see an impact of 1.10, ceterisparbus. The effect of a
reduction in population by 10 million persons produces the largest impact, a
decreaseof 1.35 in human rights abuse.24The effect of a gain of $20,000 in per
capita GNP produces the smallestimpact, a decreaseof.55 by the tenth lag.
It should be rememberedthat the theoreticalchanges that are inferredhere
in orderto calculatethe over-timeeffectsassumethe maximumvariation.Because
of differing operationalizationof variables,some variableshave a tremendous
range for variation-population and per capita GNP, for example. Thus, the
potentialfor variationincreasestheir estimatedover-timeeffect.Variablesthat are
dichotomous have a very limited range for change, and thus their over-time
impactwill be smallerthan the effectthat can be achievedwith variablesthathave
a full range of variation.And often assuming substantialvariationin those variables with a full range of variationis much less realisticthan assuminga change
of state in the dichotomousvariables.For example, the task of increasinga state's
per capita GNP by $20,000 (an amount approximatelyequivalentto the size of
Canada'sGNP in 1996) or reducingthe populationsize by ten million would be
ratherdifficultto achieve comparedto writing constitutionalprovisionsinto law.
Certainly,the willingnessof the stateto allow these actionsor to take steps toward
producing them would be necessary in order to bring about any of the three
changes. However,the willingness of the state to take action might not be sufficient to bring about such a reductionin the population (unless we assume genocide) or such an increasein the per capita GNP Thus from a practicalpoint of
view, pursuing better human rights through constitutional law offers greater
potentialthan pursuingit throughother less manipulablefactors.
24
Obviously,an increasein populationis the more likely scenarioand it would be expected that such
a sudden increasewould produce a 1.39 increase in human rights abuse. But again, in order to
comparethe overtimeeffects side-by-side, we must base the impacts upon the same assumptionproducing a decreasein human rights. Unfortunately,the ease of comparisoncreatesa somewhat
theoreticallyimplausiblescenariowith this variable.
133
PoliticalResearch
Quarterly
CONCLUSION
The purpose of this articlewas twofold: (1) to expand, across two decades
and across a global set of countries, the empiricalexaminationof the impact of
constitutionalprovisionson states human rightsbehavior,and (2) to build upon
the existingbody of literaturethat seeks to explain state abuse of the right to personal integrityWhile this study has identifiedtwo importantfactorsto add to the
list of known factorsthat improvethe likelihood of a state protectingthe right to
personalintegrity,the empiricaltest of the impact of these ten constitutionalprovisions is somewhatdiscouraging,given the effortat global constitutionbuilding.
In particular,the assessment of the impact of the constitutionalpromise of the
five basic freedoms (speech, assembly,association, religion, and press) and the
right to strike on human rights abuse is not very optimistic. None of these provisions produced an observableimprovement on human rights behavior. The
constitutionalprovision for freedomof press did demonstratethe expected relationship when the model controlled for the interactiveeffects with civil war.
Countries that face civil war and have a constitutionalpromise of freedom of
press are, indeed, less likely to abuse human rightsthan a state experiencingcivil
war that has no such provision.Futurestudies that continue to examinethe relationship between these promisesand human rightspracticesshould examine the
role of judicial independenceand legislativeconstraintson the executive, particularlyin regardto internaland externalthreats.
The analysis of the due process provisions produced perplexing results in
regardto the two variablesmost closely related to personal integrityabuse (the
ban on tortureand the promise of the writ of habeascorpus).While the levels of
statistical significance for these measures are quite high, the coefficients are
signed in the wrong direction. Clearly,if I had tested these variableswith a 2tailed level and not predicteda direction,the variableswould have been statistically significant. However, there is no theoreticaljustification to suggest that
these constitutionalpromises would increase the likelihood of political repression. Two of the due process provisionswere substantivelyand statisticallysignificant-the provisions for public and fair trials.The impact of simultaneously
adoptingboth of these constitutionalprovisionsis fairlylarge.Within ten years,
the state should experience an improvementof approximatelyone level in personal integrity scores, ceterisparibus. These two factors are inherently more
manipulablethan the other factorsknown to influence states'abuse of personal
integrityrights-civil war,internationalwar,populationsize, economic development, and political democracy-and thus from a policy perspective,may offer a
more feasiblepath to pursue in regardto human rights. Thus, while overall the
impact of constitutionsmay be less than would be expected by legal scholarsor
hoped for by human rights activists,the resultsof this study suggest that constitutions do mattermore than would be expected by their critics. It is clearlyprematureto dismiss the role of constitutionsin the realizationof personalintegrity
134
IndividualHumanRights(1977-1996)
or other human rights. Scholars should continue to explore the circumstances
and conditions under which constitutions can be more than "mereparchment
barriers"to human rights abuse.
REFERENCES
Alfatooni, Afra, and Michael PatrickAllen. 1991. "GovernmentSanctions and
CollectiveProtestin Peripheryand SemiperipheryStates."Journalof Political
and MilitarySociology19 (Summer):29-45.
and Constitutionalism.
New York:Van
Andrews, William G. 1964. Constitutions
Norstrand.
Banks, Arthur. 1991. CrossNationalTimeSeriesData Archive.Binghamton, NY:
Centerfor Social AnalysisPress.
Beatty,David. 1994. "HumanRights and the Rules of Law."In David M. Beatty,
ed., HumanRightsandJudicialReview.Dordrecht:MartinusNijhoff.
Beck, Nathaniel, and Jonathan N. Katz. 1995. "Whatto Do (and What Not to
Do) with Time-Series-Cross-SectionalData in ComparativePolitics."American PoliticalScienceReview89 (3): 634-47.
Beck, Nathaniel, Jonathan N. Katz, R. Michael Alvarez, GeoffreyGarrett,and
Peter Lange. 1993. "GovernmentPartisanship,Labor Organization, and
MacroeconomicPerformance:A Corrigendum."AmericanPoliticalScience
Review87 (4): 945-48.
Beckett, Paul. 1987. "Electionsand Democracy in Nigeria."In Fred Hayward,
ed., Electionsin Independent
Africa.Boulder,CO: WestviewPress.
Gerard
and
L.
David
Blasi,
J.,
Cingranelli. 1994. "A Model of Human Rights
Sequence."Unpublished manuscript.
_ . 1996. "Do Constitutionsand InstitutionsHelp ProtectHuman Rights?"In
David Cingranelli,ed., HumanRightsand DevelopingCountries.Greenwich,
CT:JAI.
Blaustein,Albert P 1993. Constitutions
of the World.Nashville, TN: Carmichael
and Carmichael.
Bohrer,R. E. II. 1997. "DeviationsfromProportionalityand Survivalin New ParliamentaryDemocracies."ElectoralStudies16: 217-26.
Boli-Bennett,John. 1976. "TheExpansionof NationalStates. 1870-1970." Ph.D.
dissertation.Palo Alto, CA:StanfordUniversity.
Booth, John, and PatriciaBayer Richard. 1996. "Repression,Participationand
DemocraticNorms in CentralAmerica."AmericanJournalof PoliticalScience
40:1205-32.
Booth, John, and Mitchell Seligson. 1989. Electionsand Democracyin Central
America.Chapel Hill: Universityof North CarolinaPress.
Carleton,David, and Michael Stohl. 1987. "TheRole of Human Rights in U.S.
ForeignAssistancePolicy"AmericanJournal
of PoliticalScience31: 1002-18.
135
PoliticalResearch
Quarterly
Cross, Frank. 1999. "TheRelevanceof Law in Human RightsProtection."InternationalReviewof Lawand Economics19: 87-98.
Dahl, Robert.1992. "Democracyand HumanRightsUnder DifferentConditions
of Development."In AsbjornEide and BerntHagtvet,eds. HumanRightsin
A GlobalAssessment.Oxford:BlackwellPublishers.
Perspective:
Christian.
1995a. "Assessingthe Military'sInfluence on Political
Davenport,
Repression."Journalof Politicaland MilitarySociology23: 119-44.
.1995b. "MultiDimensional Threat Perception and State Repression: An
Inquiry Into Why States Apply Negative Sanctions."AmericanJournalof
PoliticalScience39 (3): 683-713.
. 1995c. "Assessing the Military'sEffect on State Repression." Journal of Polit-
ical and MilitarySociology23 (Summer):119-44.
. 1996. "'ConstitutionalPromises'and RepressiveReality:A CrossNational
TimeSeriesInvestigationof Why Politicaland CivilLibertiesareSuppressed,'
Journalof Politics58(3): 627-654.
. 1998. "LiberalizingEvent or LethalEpisode?:An EmpiricalAssessmentof
How National ElectionsAffect the Suppressionof Politicaland Civil Liberties."SocialScienceQuarterly79: 321-40.
. 1999. "HumanRightsand the DemocraticProposition."Journalof Conflict
Resolution43 (1): 92-116.
Davis, David R., and Michael D. Ward. 1990. "Deathsand the Disappearedin
ContemporaryChile."Journalof ConflictResolution34 (3): 449-75.
. 1990. "They Dance Alone: Deaths and the Disappeared in Contemporary
Chile."Journalof ConflictResolution34: 449-75.
Diamond, Larry,Juan Linz, and SeymourLipset. 1988. Democraciesin Developing
Countries,vol. 4. Boulder,CO: LynneReinner.
Dixon, WilliamJ. 1994. "Democracyand the PeacefulSettlementof International
Conflict."AmericanPoliticalScienceReview88: 14-32.
Donnelly,Jack. 1999. "HumanRights, Democracy,and Development."Human
RightsQuarterly21: 608-32.
Duvall, Raymond D., and Michael Stohl, 1983. "Governanceby Terror."In
MichaelStohl, ed., ThePoliticsof Terrorism,
3rd ed., pp. 179-219. New York:
MarcelDekker.
Elster,Jon. 1993. "Introduction."In Jon Elster and Rune, eds., Slagstad.Constitutionalismand Democracy. Cambridge, MA: Cambridge University
Press.
Lawyers,ActivistsandSupremeCourts
Epp, CharlesR. 1998. TheRightsRevolution:
in Comparative
Perspective.Chicago:Universityof ChicagoPress.
Feierabend,Ivo D., and RosalindL. Feierabend. 1972. "SystematicConditions
and PoliticalAggressions:An Applicationof FrustrationAggressionTheory."
In Ivo Feierabendet al., eds., Anger,Violenceand Politics.Englewood Cliffs,
NJ:Prentice-Hall.
136
IndividualHumanRights(1977-1996)
Fein, Helen. 1995. "MoreMurderin the Middle: Life IntegrityViolations and
Democracyin the World, 1987." HumanRightsQuarterly17: 170-91.
Government.
Middlesex,England:Penguin Books.
Finer,S. E. 1974. Comparative
Finer, S. E. 1995. Veron Bogdanor,and BernardRudden. ComparingConstitutions.Oxford:ClaredonPress.
in Crisis:PoliticalViolenceand the Ruleof Law.
Finn, John E. 1991. Constitutions
New York:OxfordUniversityPress.
Flanz, Gisbert H. 1997. Constitutions of Countries of the World. New York:
OceanaPublications.
Francisco,RonaldA. 1995. "TheRelationshipBetweenCoercionand Protest:An
EmpiricalEvaluationin the Three CoerciveStates."Journalof ConflictResolution39: 263-82.
. 1996. "Coercionand Protest:An EmpiricalTestin TwoDemocraticStates."
AmericanJournalof PoliticalScience40: 1179-1204.
Fruhling,Hugo E. 1993. "HumanRightsin ConstitutionalOrderand in Political
Practicein LatinAmerica."In Douglas Greenberg,et al., eds. Constitutionalin the Contemporary
World.New York:Oxford
ism and Democracy:Transition
Press.
University
Gartner,Scott Sigmund, and PatrickM. Regan, 1996. "Threatand Repression:
The NonLinear Relationship between Government and Opposition Violence."Journalof PeaceResearch33 (3): 273.
Gastil,RaymondD. 1980. Freedomin the World:PoliticalRightsand CivilLiberties
1980. New Brunswick,NJ:TransactionBooks.
. 1988. Freedomin theWorld.1987. Lanham,MD:UniversityPressof America.
. 1990. "The ComparativeSurvey of Freedom: Experiences and SuggesInternational
tions."Studiesin Comparative
Development25: 25-50.
"The
PoliticalTerrorScale."In Stuart
and
Matthew
Dalton.
1996.
Gibney,Mark,
S. Nagel and David Louis Cingranelli,eds., HumanRightsand Developing
Country.Greenwich,CT:JAIPress.
Gibney,Mark,VanessaDalton, and MarcVockell, 1992. "USARefugeePolicy:A
Human RightsAnalysisUpdate."Journalof RefugeeStudies5 (1): 33-46.
Gibney, Mark, and Michael Stohl. 1988. "Human Rights and U.S. Refugee
Policy."'In MarkGibney,ed., OpenBorder?ClosedSocieties?TheEthicaland
PoliticalIssues.Westport,CT:GreenwoodPress.
Gleditsch,KristianS., and MichaelD. Ward. 1997. "DoubleTake:A Reexamination of Democracy and AutocracyPatternsin Modern Politics."Journalof
ConflictResolution41: 361-83.
Gurr,Ted Robert.1968. "ACausalModel of Civil Strife:A ComparativeAnalysis
Using New Indices."AmericanPoliticalScienceReview62 (4): 1104-24.
. 1970. WhyMenRebel.Princeton,NJ:PrincetonUniversityPress.
. 1986. "ThePolitical Origins of State Violence and Terror:A Theoretical
Analysis."In MichaelStohl and GeorgeA. Lopez, eds., GovernmentViolence
137
PoliticalResearch
Quarterly
and Repression:
AnAgendaforResearch.Westport,CT:GreenwoodPress (4571).
. 1990. Polity II: PoliticalStructuresand RegimeChange:1800-1986. Ann
Arbor,MI:Inter-UniversityConsortiumfor Politicaland Social Research.
Gurr,Ted Robert,and RaymondDuvall. 1973. "CivilConflict in the 1960s: A
Political
ReciprocalTheoreticalSystem with ParameterValues."Comparative
Studies6 (1): 135-70.
Hayward,Fred, and Jimmy Kandeh. 1987. "Perspectiveson Twenty-fiveYearsof
Elections in SierraLeone."In Fred Hayward,ed., Electionsin Independent
Africa.Boulder,CO: WestviewPress.
Henderson, Conway W 1993. "PopulationPressuresand Political Repression."
SocialScienceQuarterly74 (2): 322-33.
. 1991. "ConditionsAffecting the Use of Political Repression."Journalof
ConflictResolution35 (1): 120-42.
Herman, Edward, and Frank Brodhead. 1984. DemonstrationElections:U.S.StagedElectionsin the DominicanRepublic,Vietnam,and El Salvador.Boston,
MA:South End Press.
Heston, Alan. 1994. "A Brief Review of Some Problems in Using National
Accounts Data in Level of Output Comparisonsand GrowthStudies."Journal of DevelopmentEconomics44: 52.
CausalAnalyHibbs, DouglasA., Jr. 1973. MassPoliticalViolence:A CrossNational
sis. New York:John Wiley
Howard,A. E. Dick. 1991. "TheEssence of Constitutionalism."In KennethW
and Human
Thompson and Rett T Ludwikowski, eds., Constitutionalism
Rights:America,Poland,and France.New York:Lanham.
Humana, Charles. 1992. WorldHumanRightsGuide.New York:Oxford University Press.
Huntington,SamuelP 1964. TheSoldierandtheState.New York:RandomHouse.
. 1993. The Third Wave: Democratization in the Late Twentieth Century.
Norman:Universityof OklahomaPress.
Ibhawoh, Bonny 2000. "BetweenCultureand Constitution:Evaluatingthe CulturalLegitimacyof Human Rightsin the AfricanState."HumanRightsQuarterly22: 838-60.
Innes, Judith Eleanor. 1992. "HumanRights Reporting as a Policy Tool: An
Examinationof the StateDepartmentCountryReports."In ThomasB.Jabine
and RichardP Claude, eds., HumanRightsand Statistics:Gettingthe Record
Straight.Philadelphia,PA:Universityof PennsylvaniaPress.
Jaggers,Keith, and Ted R. Gurr. 1995. 'TrackingDemocracy'sThird Wave with
the Polity III Data."Journalof PeaceResearch32 (4): 469-82.
Jenkins,J. Craig,and CharlesPerrow.1977. "Insurgencyof the Powerless:The
FarmWorkersMovement,1946-1972."AmericanSociological
Review42 Kenneth W Thompson and RettT. Ludwikowski,eds. (2): 249-68.
138
IndividualHumanRights(1977-1996)
Karatnycky,Aaron. 1994. "TheComparativeSurveyof Freedom:1994." Freedom
Review25: 413.
Keith, LindaCamp. 1999. "TheUnited Nations InternationalCovenanton Civil
and PoliticalRights:Does It Makea Differencein Human RightsBehavior?"
Journalof PeaceResearch36 (1): 95-118.
Keith, Linda Camp, and Steven C. Poe. 2000. "TheU.S., the I.M.E,and Human
Rights."In David P Forsythe,ed., TheUnitedStatesand HumanRights:Looking Inwardand Outward.Lincoln:Universityof NebraskaPress, 2000.
Krain, Matthew 1997. "The Onset and Severity of Genocides an Politicides."
Journalof ConflictResolution41: 331-60.
Krain, Matthew. 2000. "Democracy,Internal War, and State Sponsored Mass
Murder."HumanRightsReview1: 40-48.
States.New
in Post-Revolutionary
andAccommodation
.forthcoming.Repression
York:St. Martin'sPress.
An Introduction.
CA: Newbury
Lewis-Beck,Michael S. 1980. AppliedRegression:
Park.
Lichbach,Mark,and Ted RobertGurr.1981. "TheConflictProcess:A Self-GeneratingModel."Journalof ConflictResolution31 (1): 266-97.
Lijphart,Arend. 1984. Democracies:Patternsof Majoritarianand ConsensusGovernmentin TwentyOne Countries.New Haven, CT:YaleUniversityPress.
1994. "Democracies:Forms, Performance,and ConstitutionalEngineering."EuropeanJournalof PoliticalResearch25: 1-17.
Linz. Juan J. 1992. "Typesof Political Regimes and Respect for Human Rights:
Historical and Cross-NationalPerspectives."In Asbjorn Edie and Bernt
Hagtvet, eds., Human Rightsin Perspective:A GlobalAssessment.Oxford:
BlackwellPublishers.
or ParliamentaryDemocracy:Does It MakeA Dif_ . 1994. "Presidentialism
ference?"In JuanJ. Linz and ArturoValenzuela,eds., TheFailureof Presidential Democracy.Baltimore,MD:Johns Hopkins UniversityPress.
Lopez, George, and Michael Stohl. 1992. "Problemsof Concept and Measurement in the Study of Human Rights."'In Thomas B. Jabine and RichardP
Claude,eds., HumanRightsandStatistics:GettingtheRecordStraight,pp. 216234. Philadelphia:Universityof PennsylvaniaPress.
in the Regionof FormerSoviet
Ludwikowski, Rett R. 1996. Constitutionmaking
Press.
Dominance.Durham:Duke University
MacGuigan, Mark. 1965. "The Development of Civil Liberties in Canada."
Queen'sQuarterly27: 273.
Madani,Hamed. 1992. "SocioeconomicDevelopment and MilitaryPolicy Consequences of ThirdWorldMilitaryand CivilianRegimes, 1965-1985." Ph.D.
dissertation,Universityof North Texas.
Madison,James. 1977. ThePapersofJamesMadison,vol. 11. Charlottesville:University Press of Virginia.
139
PoliticalResearch
Quarterly
March,JamesG., andJohanP Olsen. 1984. "TheNew Institutionalism:OrganizationalFactorsin PoliticalLife."AmericanPoliticalScienceReview78: 1212-41.
Martin,David. 1991. "RoundtableDiscussion."In Kenneth W Thompson and
Rett T. Ludwikowski, eds., Constitutionalism
and Human Rights:America,
Poland,and France,New York:Lanham.
McAdam,Douglas. 1982. PoliticalProcessand theDevelopment
of BlackInsurgency.
of
Press.
Chicago:University Chicago
McCamant,John. E 1981. "A Critique of Present Measuresof Human Rights
Development and an Alternative.In Ved P Nanda, James R. Scarritt,and
George W Shepard,eds., GlobalHumanRights:PublicPolicies,Comparative
Measures,and NGO Strategies.Boulder,CO: Westview.
McColm, R. Bruce (survey coordinator). 1990. Freedomin the World:Political
Rightsand Civil Liberties,19891990. Lanham, MD: University of America
Press.
McCormick,James M., and Neil J. Mitchell. 1988. "IsU.S. Aid ReallyLinkedto
Human Rightsin LatinAmerica?"AmericanJournalof PoliticalScience32 (2):
231-39.
_ . 1989. "HumanRights and ForeignAssistance:An Update."SocialScience
Quarterly70: 969-79.
McKinlay,R. D., and A. S. Cohan. 1975. "AComparativeAnalysis of the Political and EconomicPerformanceof Militaryand CivilianRegimes."ComparativePolitics8: 130.
Milner,Wesley,Steven C. Poe, and David Leblang.1999. "SecurityRights,Subsistence Rights and Liberties:A TheoreticalSurvey of the EmpiricalLandscape."HumanRightsQuarterly21: 403-43.
Mitchell, Neil J., and James M. McCormick. 1988. "Economicand Political
Explanationsof Human RightsViolations."WorldPolitics40 (4): 476-98.
Moaddel, Mansoor. 1994. "PoliticalConflict in the World Economy: A CrossNationalAnalysisof Modernizationand World-SystemTheories."American
SociologicalReview59: 276-303.
Muller,Edward. 1985. "IncomeInequality,Regime Repressivenessand Political
Violence."AmericanSociologicalReview50 (1): 47-61.
Murphy,WalterE 1993. "Constitutions,Constitutionalism,and Democracy"In
and Democracy:Transition
in
Douglas Greenberget al., eds., Constitutionalism
the Contemporary
World.New York:OxfordUniversityPress.
Ostrom, CharlesW 1990. TimeSeriesAnalysis:RegressionTechniques,2nd. ed.
SageUniversitySeriesPaperson QuantitativeApplicationsin the Social Sciences, no. 9. BeverlyHills: Sage.
EconomicDoctrineand PolitPion-Berlin,David. 1989. TheIdeologyof StateTerror:
in Argentinaand Peru.Boulder,CO: Rienner.
ical Repression
Poe, Steven C., 1991. "HumanRightsand the Allocationof U.S. MilitaryAssistance."Journalof PeaceResearch28: 205-216.
140
IndividualHumanRights(1977-1996)
. 1992. "HumanRightsand EconomicAid Allocationunder RonaldReagan
and Jimmy Carter."AmericanJournalof PoliticalScience36: 14767.
Poe, Steven C., and RangsimaSirirangsi.1993. "HumanRights and U.S. Economic Aid to Africa."InterationalInteractions18 (4): 114.
_ . 1994. "HumanRights and U.S. Economic Aid During the ReaganYears."
SocialScienceQuarterly75 (3): 494509.
Poe, Steven C., and C. Neal Tate. 1994. "Repressionof HumanRightsto Personal
Integrityin the 1980s: A GlobalAnalysis."AmericanPoliticalScienceReview
88 (4): 853-72.
Poe, Steven C., C. Neal Tate, and Linda Camp Keith. 1999. "Repressionof
Human Rights to Personal Integrity Revisited: A Global Cross-National
Study Covering the Years 1976-1993." InternationalStudiesQuarterly43:
291-313.
Poe, Steven C, C. Neal Tate, Linda Camp Keith, and Drew Lanier.2000. "The
Continuity of Suffering:Domestic Threatand Human RightsAbuse Across
HumanRightsand
Time."In ChristianDavenport,ed., Pathsto StateRepression:
Roman
Littlefield.
CO:
Politicsin Comparative
Contentious
Boulder,
Perspective.
Are
1998.
"How
these Picand
Sabine
Steven
C., TanyaVasquez,
Poe,
Zanger.
tures Different?An EmpiricalComparisonof the U.S. StateDepartmentand
AmnestyInternationalHumanRightsReports,1976-1995." Paperpresented
at the meeting of the InternationalStudiesAssociation,Minneapolis.
_ . 2001. "How are these PicturesDiffferent:Assessing the Biases in the U.S.
State Department'sCountry Reports on Human Rights Practices."Human
RightsQuarterly23 (3) 650-77.
Democracies:
Powell, G. Bingham. 1982. Contemporary
Stabilityand
Participation,
Violence.Washington,DC: C.Q. Press.
Pritchard,Kathleen. 1986. "ComparativeHuman Rights:An IntegrativeExplanation."PolicyStudiesJournal15 (1): 110-28.
Rasler,Karen, 1986. "War,Accommodation,and Violence in the United States,
18901970." AmericanPoliticalScienceReview80 (3): 921-45.
Regan,PatrickM. 1995. "U.S.EconomicAid and PoliticalRepression:An Empirical
Evaluationof US ForeignPolicy"PoliticalResearchQuarterly48 (3): 613-28.
Resenbrink,John. 1991. "RoundtableDiscussion." In Kenneth W Thompson
and HumanRights:America,
and RettT. Ludwikowski,eds., Constitutionalism
Poland,and France.New York:Lanham.
Richards,David L. 1999. "PerilousProxy: Human Rights and the Presence of
National Elections."SocialScienceQuarterly80: 648-65.
In Louis Henkin and AlbertJ. Rosenthal,
Rosenthal,AlbertJ. 1990. "Afterword."
and
the
Constitutionalism
eds,
Rights: Influenceof the UnitedStatesConstitution
Abroad.New York:ColumbiaUniversityPress
Rummel, Rudolph J. 1997. PowerKills:Democracyas a Methodof Nonviolence.
New Brunswick,NJ:TransactionPress.
141
PoliticalResearch
Quarterly
Safran,William. 1981. "CivilLibertiesin Democracies:ConstitutionalNorms,
Practices,and Problemsof Comparison."In Ved P.Nanda,James R. Scarritt,
and GeorgeW Shepherd,Jr., eds., GlobalHumanRights:PublicPolicies,ComparataiveMeasures,and NGOStrategies.Boulder,CO: WestviewPress.
Sartori,Giovanni. 1962. "Constitutionalism:A PreliminaryDiscussion."American PoliticalScienceReview56:853-64.
Schmitter,Philippe C. 1992. "The Irony of Modem Democracy and Effortsto
ImproveIts Practice."Politicsand Society20: 507-12.
Schmitter,Philippe C., and TerryLynnKarl.1991. "WhatDemocracyIs...and Is
Not."Journalof Democracy2: 75-88.
Schwarzer,WilliamW 1994. "Civiland HumanRightsand the CourtsUnder the
New Constitutionof the Russian Federation."TheInternationalLawyer28:
825-34.
in LatinAmerica:The CurrentCycle"
Seligson,Mitchell. 1987. "Democratization
In James M. Malloy and Mitchell Seligson, eds., Authoritarians
and Democrats:RegimeTransitionin LatinAmerica.Pittsburgh,PA:University of Pittsburgh Press.
Shugart, Matthew Soberg, and John S. Carey 1993. Presidentsand Assemblies.
New York:CambridgeUniversityPress.
for a Nationor RepublicEmerging
Siegan, Bernard.1994. Draftinga Constitution
intoFreedom.Fairfax,VA:GeorgeMasonUniversityPress.
Small, Melvin, and J. David Singer. 1982. Resortto Arms:Internationaland Civil
Wars,1816-1980. BeverlyHills, CA:Sage.
State Corporation.1999. StataReferenceManual,Release6. College Station, TX:
StataPress
Stepan,Alfred,and Cindy Skach. 1993. "ConstitutionalFrameworksand Democratic Consolidation:Parlimentarianismversus Presidentialism.WorldPolitics46: 1-22.
Stimson, James A. 1985. "Regressionin Space and Time: A StatisticalEssay.'
AmericanJournalof PoliticalScience29 (4): 914-47.
Stohl, Michael,and David Carleton.1985. "TheForeignPolicy of HumanRights:
Rhetoricand RealityfromJimmy Carterto Ronald Reagan."HumanRights
Quarterly7 (2): 205-29.
Stohl, Michael, David Carleton,and Steven E. Johnson. 1984. "HumanRights
and U.S. ForeignAssistance:FromNixon to Carter."
Journalof PeaceResearch
21 (3): 215-26.
Stohl, Michael, David Carleton, George Lopez, and Stephens Samuels. 1986.
"StateViolation of Human Rights: Issues and Problems of Measurement."
HumanRightsQuarterly8 (4): 592-606.
Stotzky,Irwin P 1993. "TheTraditionof ConstitutionalAdjudication."In Irwin
P Stotzky,ed., Transitions
to Democracyin LatinAmerica:TheRoleof theJudiCO:
Westview.
ciary.Boulder,
142
IndividualHumanRights(1977-1996)
Summers,Robert,and Alan Heston. 1984. "ImprovedInternationalComparisons
of Real Product and its Composition: (1950-1980). Reviewof Incomeand
Wealth30: 207-62.
Sunstein, Cass R. 1993. "TheNegative Constitution:Transitionsin LatinAmerto Democracyin LatinAmerica:The
ica." In Irwin P Stotzky,ed., Transitions
Roleof theJudiciary,ed. Boulder,CO: Westview.
Taagapera,Rein, and MatthewSobergShugart.1989. Seatsand Votes:TheEffects
of Determinantsof ElectoralSystems.New Haven, CT:YaleUniversityPress.
Taylor,CharlesLewis, and David Jodice. 1983. WorldHandbookof Politicaland
SocialIndicatorsIII.New Haven, CT:YaleUniversityPress.
Tilly, Charles. 1978. From Mobilizationto Revolution.Reading, MA: AddisonWesley
A Comparative
Vanhanen,Tatu. 1990. TheProcessof Democratization:
Studyof 147
States,1980-88. New York:CraneRussak.
Wesson, Robert,ed. 1987. Democracy:A Worldwide
Survey.New York:Praeger.
ConsistentCovarianceMatrixEstimaWhite, Halbert. 1980. "Heteroscedasticity
Econometrica
48: 817-38.
tor and a Direct Test for Heteroscedasticity."
in Developeds.
1987.
Elections
and
Ozbudun,
Weiner,Myron,
Ergun
Competitive
Press.
NC:
Duke
Countries.
Durham,
University
ing
Wildavsky,Aaron. 1984. ThePoliticsof the BudgetaryProcess.Boston, MA:Little,
Brown.
Zakaria,Fareed. 1997. "TheRise of IlliberalDemocracy"ForeignAffairs76 (6):
22-43.
Zanger, Sabine C. 2000. "A Global Analysis of the Effect of Political Regime
Changeson Life IntegrityViolations, 1977-93."Journalof PeaceResearch37:
213-33.
Researchon PoliticalProtest."
Ziegenhagen,EduardC. 1986. Macro-comparative
In Ted RobertGurr,ed., Handbookof PoliticalConflict:Theoryand Research.
New York:Free Press.
Zwick,James. 1984. "Militarismand Repressionin the Philippines."In Michael
Stohl and GeorgeLopez, eds., TheStateas Terrorist:
the Dynamicsof GovernmentViolenceand Repression.
Westport,CT:Greenwood.
Received:February24, 2000
Accepted for Publication:May 1, 2001
[email protected]
143