Sie sind auf Seite 1von 26

Comparative Political Studies

http://cps.sagepub.com Toward a Unified Theory of Causality


James Mahoney Comparative Political Studies 2008; 41; 412 originally published online Jan 31, 2008; DOI: 10.1177/0010414007313115 The online version of this article can be found at: http://cps.sagepub.com/cgi/content/abstract/41/4-5/412

Published by:
http://www.sagepublications.com

Additional services and information for Comparative Political Studies can be found at: Email Alerts: http://cps.sagepub.com/cgi/alerts Subscriptions: http://cps.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations (this article cites 53 articles hosted on the SAGE Journals Online and HighWire Press platforms): http://cps.sagepub.com/cgi/content/refs/41/4-5/412

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Toward a Unified Theory of Causality


James Mahoney
Northwestern University, Evanston, Illinois

Comparative Political Studies Volume 41 Number 4/5 April/May 2008 412-436 2008 Sage Publications 10.1177/0010414007313115 http://cps.sagepub.com hosted at http://online.sagepub.com

In comparative research, analysts conceptualize causation in contrasting ways when they pursue explanation in particular cases (case-oriented research) versus large populations (population-oriented research). With case-oriented research, they understand causation in terms of necessary, sufficient, INUS, and SUIN causes. With population-oriented research, by contrast, they understand causation as mean causal effects. This article explores whether it is possible to translate the kind of causal language that is used in caseoriented research into the kind of causal language that is used in populationoriented research (and vice versa). The article suggests that such translation is possible, because certain types of INUS causes manifest themselves as variables that exhibit partial effects when studied in population-oriented research. The article concludes that the conception of causation adopted in case-oriented research is appropriate for the population level, whereas the conception of causation used in population-oriented research is valuable for making predictions in the face of uncertainty. Keywords: causality; logic; probabilities; cases; populations

omparative analysts use distinct methods and conceptualize causation in contrasting ways when they carry out explanation in particular cases versus large populations. When analyzing individual cases, comparativists seek to identify the specific values of variables that enabled and/or generated outcomes of interest. They pursue a genetic and sequential mode of explanation that is implemented in part through narrative prose. By contrast, when comparativists study large populations, they seek to estimate the average effects of independent variables of interest. They carry out statistical analysis and report results as coefficients that apply to populations as wholes rather than particular observations.
Authors Note: For helpful insights, I thank James Caporaso, Robert J. Franzese Jr., John Gerring, Edward Gibson, Gary Goertz, Erin Kimball, Kendra Koivu, Damon B. Palmer, Jason Seawright, Larkin Terrie, David Waldner, Erik Wibbels, and Steven I. Wilkinson. All errors are mine. 412
Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

413

These two approaches can be called, for convenience, case-oriented and population-oriented research.1 Case-oriented researchers seek to identify the causes of particular outcomes in specific cases. They may find causal patterns that apply broadly, but their primary concern is with causation in the specific cases under analysis. By contrast, population-oriented researchers seek to identify typical causal effects in overall populations. They may sometimes investigate whether a particular case follows a general casual pattern, but their main interest is to say something about the larger population pattern. To illustrate these differences, consider research on democratization. Both case-oriented and population-oriented researchers are interested in the question, What causes democracy? However, they normally address this question in different ways. The case-oriented researcher pursues it by asking what caused democracy in certain cases, such as a set of democratic and nondemocratic countries that are homogeneous on contextual variables. It is common, for example, to ask why democracy occurred among some but not other countries of a particular region (e.g., Yashar, 1997) or a particular historical epoch (e.g., R. B. Collier, 1999). The analyst may explain the cases under investigation without being able to generalize this explanation to a broad range of places and times. The population-oriented researcher, by contrast, addresses the question by asking about the partial effects of independent variables of interest on the distribution of democracy in a large number of cases. For example, cross-national research on democracy is often concerned with producing good estimates of the average effect of variables such as economic development (e.g., Londregan & Poole, 1996) or the number of political parties (e.g., Mainwaring, 1993). In these studies, the analyst may accurately report the typical effects of particular independent variables without providing a comprehensive explanation of democracy for any particular case. Perhaps most comparativists believe that both case-oriented and populationoriented forms of research are worthy. However, their many differences are not well appreciated, which fosters misunderstandings (Mahoney & Goertz, 2006). Of these differences, perhaps the most basic one concerns conceptions of causality. Case-oriented and population-oriented approaches diverge in their fundamental understandings of causation, leaving the comparative field without a unified theory of causality. The challenge of unification involves reconciling the apparently contradictory claims about causation endorsed in the two approaches. How can causation be both a process that enables or generates specific outcomes in cases and a statistical likelihood that operates probabilistically within a population? A unified theory should tell us why there is no contradiction here and why (and if) we need such seemingly different understandings of causation at the case level and the

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

414

Comparative Political Studies

population level. It should in fact provide the tools for translating the kind of causal language that is used at the case level into the kind of causal language that is used at the population level (and vice versa). In this article, I explore how to unify the case-oriented and populationoriented approaches to causation. The theory for unification that I propose is reductionistit assumes that causal effects at the population level manifest themselves only insofar as causal processes are operating in the individual cases.2 The population level does not exhibit any emergent properties that cannot be reduced to (i.e., explained in terms of) processes that occur in the individual cases. Causation at the population level is thus epiphenomenal; case-level causation is ontologically prior to population-level causation. Although to some this bottom-up approach will seem obvious, the more common approach to achieve unification has been top-down: Scholars try to understand causation at the level of the individual cases using ideas that apply to the population level. As we shall see, this approach is fraught with problems that are corrected by starting with a firm understanding of causation at the level of the individual cases and then extending this understanding to the level of the population. In developing a unified theory of causation, I do not explore all differences between the case-oriented and population-oriented approaches. Perhaps most notably, I do not consider here important questions related to temporality, causal mechanisms, and the transmission of causal effects over time. Nevertheless, the issues covered in this article provide a necessary foundation for the further explication of these and other differences between the two approaches.

The Case-Oriented Approach


Case-oriented researchers seek to explain particular outcomes in specific cases. They ask questions such as, Why did European countries develop either liberal-democratic, social-democratic, or fascist regimes during the interwar period? (Luebbert, 1991); Why did the United States provide generous benefits for Civil War veterans and their dependents?(Skocpol, 1992); and Why did Korea and Taiwan but not Syria and Turkey experience sustained economic development after World War II? (Waldner, 1999). In addressing these questions, the researchers try to identify the values on variables that actually caused the particular outcomes in the specific cases. This research goal seems straightforward enough. But what concretely do case-study researchers mean when they assert that a given variable value

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

415

is a cause of a particular outcome? For example, what does Skocpol (1992) mean when she argues that competitive patronage democracy was a cause of the Civil War pensions? There are different possible answers to these questions, some more useful than others.

Probability-Based Definitions
Perhaps the most common approach to defining causation in the individual case is to assume that it works like causation at the level of the population. Here, causality is conceived in terms of likelihoods and probabilities. A cause is specifically a value on a variable that makes an outcome more likely; a cause increases the probability that an outcome will take place. This definition of cause as probability raiser stems from the probability theory that underpins population-oriented research.3 Yet at the level of the individual case, there are various problems with the idea that causes are probability raisers. For one thing, probabilities must be derived from populations. And causes that increase the probability of a given outcome in a population need not increase the probability of that outcome in any particular case. Indeed, some factors that one would want to call causes of an outcome in an individual case actually decrease the probability of the outcome in a larger population. Philosophers use stylized scenarios to illustrate the idea,4 but students of comparative politics will be aware of examples. For instance, although much research suggests that economic development increases the probability of democracy (or democratic stability), it has long been treated as a cause of the breakdown of democracy and the emergence of authoritarianism in South America during the 1960s the 1970s (e.g., D. Collier, 1979; ODonnell, 1973; also see Mainwaring & Prez-Lin, 2003). Likewise, the election of leftist and labor-based governments historically makes the initiation of radical market-oriented reforms less likely on average, but these governments have been treated as a cause of such reforms in several contemporary developing countries (Levitsky & Way, 1998). What is more, the very idea of viewing causation in terms of probabilities when N = 1 is problematic. At the individual case level, the ex post (objective) probability of a specific outcome occurring is either 1 or 0; that is, either the outcome will occur or it will not. This view is consistent with nearly all interpretations of probabilities. Thus, according to frequency probability approaches: It makes no sense to speak of a single-case probability . . . . On a frequentist interpretation probabilities are properties of the whole ensemble, not of the individual events (Appleby, 2004, p. 449). Other realist

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

416

Comparative Political Studies

approaches that see probabilities as objective features also treat single-case probabilities as meaningless (Milne, 1986). Likewise, for different reasons, Bayesian subjectivists reject the idea of single-case probabilities (see Albert, 2005).5 To be sure, the ex ante (subjective) probability of an outcome occurring in a given case can be estimated in terms of some fraction. But the real probability of the outcome is always equal to its ex post probability, which is 1 or 0. For example, across many trials, the probability of a craps player rolling snake eyes (1s on two six-sided dice) is 1 in 36. Yet for any individual roll, the real probability of snake eyes at the time of the release of the dice is equal to 1 or 0. The outcome for the particular role is affected not only by the fact that each dye has six sides but also by initial position, velocity, angular momentum, and so forthfactors that determine infallibly and, in principle, predictably (not haphazardly, not quantum randomly) whether the dice will come up snake eyes.6 We can make these ideas more concrete with an example from comparative politics: the emergence of democracy in India. Scholars who work on this question routinely note that Indias long-standing democracy is surprising given what we know about population-level probabilities (e.g., Kohli, 1988, 2001; Rudolph & Rudolph, 1959; Varshney, 1998; Waldner, 2006). After all, India has many features (e.g., poverty, ethnic heterogeneity, inequality, regionalism) that are believed to severely decrease the likelihood of democracy. Yet given that democracy did happen, the challenge for case-oriented researchers has been to explain it (as opposed to assuming that it was a product of random chance). Among other things, most case-oriented researchers point to the Congress Party as a key promoter and stabilizer of democracy. Of course, in other contexts, mass-incorporating parties of this nature have triggered stable authoritarianism and even longlived totalitarian regimes (Huntington, 1968). Why should the party have been a cause of democracy in India? Interestingly, case-oriented researchers suggest that some of the very factors that made democracy in India so subjectively surprising also explain why the Congress Party promoted democracy. Thus, hierarchical social structures and ethnic divisions led the Congress Party to adopt a rule-based internal functioning in the course of the nationalist movement (Varshney, 1998). The presence of a heterogeneous society also encouraged elites to incorporate into the party middle and rich farmersactors who occupied a strategic position between landed elites and landless workers. In turn, internal bureaucratic functioning and a broad accommodationist alliance meant that the Congress Party was well poised to lead a peaceful and sustainable transition to democracy with independence. The crucial point for us is that the conditions that

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

417

caused democracy in India included some of the very factors that probabilistically decrease the likelihood of democracy in general. Without all of these obstacles, Indias Congress Party may not have played its historic role, and the country may not have developed a stable democratic regime.

Necessity and Sufficiency Definitions


Rather than thinking about case-level causes as probability raisers, we will do better to adopt an orientation grounded in the philosophy of logic. From this orientation, it is common to define cause in the individual case as a variable value that is necessary and/or sufficient for an outcome. The idea that causality should be viewed as specifically necessary causation follows a long tradition going back to Hume (Goertz, 2003; Lewis, 1986). The approach captures the intuition that a cause is something thatwhen counterfactually taken away under ceteris paribus conditionsfosters a different outcome. For instance, Moore (1966) famously quips no bourgeoisie, no democracy (p. 418) to highlight the idea that a strong and independent bourgeoisie was necessary for a path to democracy in early modern cases. More recently, Weyland (1998) argues that profound crises were necessary for leaders to initiate shock plans of economic adjustment in the contemporary developing world; Tarrow (1989) asserts that deep, visible structural cleavages and political opportunities have been necessary causes of historical protest cycles; and Amsden (1992) contends that a relatively equitable distribution of income has been necessary for successful late industrialization. In these examples, the author believes that, in the absence of the cause, the outcome of interest would not have happened. Although useful, this approach cannot accommodate sufficient causes. With sufficient causes, the counterfactual absence of the cause may not change the outcome in an individual case and thus could be interpreted as not exerting an effect under the necessary cause definition.7 There is in fact a large philosophical literature, also going back to Hume, which defines cause in terms of sufficiency rather than necessity.8 Some comparativists at least implicitly use this approach. For instance, Goldhagen (1997) argues that a culture of virulent anti-Semitism was sufficient to cause Germans to be motivated to kill Jews. Alternative causal factors, such as the rise of the Nazis and Hitler, were not necessary. Germanys anti-Semitism was enough by itself to deliver the motivational basis for the Holocaust. The real problem with these frameworks is not that they fail to speak of causes, but rather that they are not exhaustive. Individual causes do not take the form of only necessary and/or sufficient causes. One can easily think of

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

418

Comparative Political Studies

examples of causes that were neither necessary nor sufficient. In each of these examples, one will find, the cause works together in combination with other causes to produce an outcome. It is the larger combination that generates the outcome; or stated differently, the values on two or more variablesnot any single variableare jointly sufficient (or possibly jointly necessary) for the outcome. Let us examine these causes.

INUS and SUIN Causes


Case-oriented researchers rarely treat any single variable value as sufficient for a given outcome. Instead, when attempting to identify the causes that generate a specific outcome, they usually treat individual variable values as INUS causes. The acronym INUS is derived by Mackie (1965) as follows: the so-called cause is, and is known to be, an insufficient but necessary part of a condition which is itself unnecessary but sufficient for the result (p. 246; also see Mackie, 1980). A stylized illustration of INUS causation is the idea that a building can burn down (Y1) either because of a short circuit (A1) combined with wooden framing (B1) or because of a gasoline can (C1) combined with a furnace (D1); that is, Y1 = (A1 & B1) v (C1 & D1), where & represents the Logical AND, the v represents the Logical OR, and = represents sufficiency. There is no necessary cause for the outcome; instead, alternative combinations are each sufficient. INUS causes are commonplace in case-oriented research. For instance, consider Moores (1966) argument that democratic pathways in the early modern world required both a strong bourgeoisie and an aristocracy that either aligned with this bourgeoisie or was historically weakened. In the argument, there are two combinations that generate democracy: (a) a strong bourgeoisie that is allied with aristocratic elites, and (b) a strong bourgeoisie and a weak aristocracy. Although a strong bourgeoisie is a necessary cause, the bourgeoisaristocratic alliance cause and the weak aristocracy cause are INUS causes. We can summarize the argument as follows:
Y1 = X1 & (A1 v B1), (1)

where Y1 = democratic pathway; X1 = strong bourgeoisie; A1= alliance between bourgeoisie and aristocracy; and B1 = weak aristocracy. Neither A1 nor B1 is individually necessary or individually sufficient. Instead, they are INUS causes that combine with X1 to form two possible combinations that are sufficient for Y1. Many other studies that use Qualitative Comparative Analysis (QCA) in either its dichotomous or its fuzzy-set version assume

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

419

INUS causation and seek to identify the different causal pathways to given outcomes.9 An alternative kind of combinatorial cause, one that is linked to necessity, is a SUIN cause. A SUIN cause is, to make a parallel with Mackies INUS cause, a sufficient but unnecessary part of a factor that is insufficient but necessary for an outcome (Mahoney, Kimball, & Koivu, 2007). Here the constitutive attributes of a necessary cause are treated as causes themselves. For example, a well-known theory holds that nondemocracy (i.e., a nondemocratic dyad) is necessary for war. There are various attributes that by themselves would constitute nondemocracy, including fraudulent elections, high levels of repression, and severe suffrage exclusions. By virtue of constituting nondemocracy all by themselves, each of these conditions is a SUIN cause of war. Fraudulent elections, repression, and suffrage exclusions are neither necessary nor sufficient for war, but rather factors that can constitute nondemocracy, which in turn is necessary for war. Or consider again Moores (1966) argument. We saw how a bourgeoisaristocratic alliance and a weak aristocracy are INUS causes. If we now imagine that Moore were to treat these two factors as each examples of a third factor politically subordinate aristocracythen they would be SUIN causes: constitutive parts sufficient for a necessary cause (politically subordinate aristocracy) of democracy. We can summarize this idea as follows:
Y1 = X1 & Z1; Z1 = A1 v B1 , (2)

where Y1 = democratic pathway; X1 = strong bourgeoisie; Z1 = politically subordinate aristocracy; A1 = alliance between bourgeoisie and aristocracy; and B1 = weak aristocracy. Here A1 and B1 are SUIN causes of a democratic path. Logically speaking, then, there are five kinds of causes that can be used in case-oriented research. A cause may be (a) necessary but not sufficient, (b) sufficient but not necessary, (c) necessary and sufficient, (d) INUS, or (e) SUIN. This list appears to be mutually exclusive and collectively exhaustive (Mahoney et al., 2007). Of these five causes, the prototypical one is a necessary and sufficient cause. All other causes are derivative of a necessary and sufficient cause. In fact, the other four causes become more important to the extent that that they approach the threshold of being a necessary and sufficient cause (Mahoney et al., 2007; also see Braumoeller & Goertz, 2000; Goertz, 2006; Hart & Honor, 1959; Ragin, 2006). Thus, necessary causes are increasingly important as they approach causal sufficiency. Sufficient causes are more important as they approach causal necessity. And INUS and SUIN causes gain importance as they become closer to being individually necessary and sufficient.
Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

420

Comparative Political Studies

To embrace necessary and/or sufficient causation (including INUS and SUIN causation) is to assume that the world of human beings is deterministic.10 This assumption does not of course mean that a given researcher will be successful at identifying the causes of any outcome. Nor does this assumption mean that the researcher will achieve valid measurement or carry out good research. The assumption of an ontologically deterministic world in no way implies that researchers will successfully analyze causal processes in this world. But it does mean that randomness and chance appear only because of limitations in theories, models, measurement, and data. The only alternative to ontological determinism is to assume that, at least in part, things just happen; that is, to assume that truly stochastic factors whatever those may be (see Humphreys, 1989)randomly produce outcomes. The assumption of a genuinely probabilistic world finds its best defense with indeterministic relations in quantum mechanics. Yet whether or not subatomic processes are truly indeterministic is still debated among physicists. Moreover, as Waldner (2002) suggests, quantum theorists argue that the kinds of issues addressed in the social sciences do not work like quantum mechanics. Randomness at the subatomic level seems inappropriate when applied to the world of human beings and their objects, assuming that these entities function like all other known nonparticle entities. Before concluding this section, I want to note that various methods exist for analyzing necessity and sufficiency (including INUS and SUIN causes). Indeed, all of the major small-N cross-case methods used in this field Mills methods of agreement and difference, typological theories, counterfactual analysis, Boolean algebra, and fuzzy-set analysisassume these kinds of causes (e.g., Elman, 2005; George & Bennett, 2005; Goertz & Levy, 2007; Mahoney, 1999; Ragin, 1987, 2000). Here is not the place to review the huge literature on these methods, including critiques of them, but the methods are out there, and they are a basic part of the tool kit of case-oriented research in comparative politics.

The Population-Oriented Approach


Population-oriented researchers do not have as their primary goal the explanation of any particular case. Instead, they want to accurately estimate the effects of variables within whole populations. They ask questions such as, What is the effect of presidentialism on the consolidation of democracy? (Cheibub, 2007); To what extent is globalization a cause of inegalitarian distribution? (Garrett, 1998); and Does regime type matter for economic growth? (Gerring, Bond, Barndt, & Moreno, 2005). In addressing these

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

421

questions, the researchers want to generalize about the typical effects of the particular causal factor(s) of interest.

Mean Causal Effects


The population-oriented approach in many ways parallels experimental research.11 Experiments are usually not designed to provide comprehensive explanations of outcomes. Nor are they designed to explain particular cases. Rather, they are intended to accurately assess the typical effects of particular treatments for populations. Experimenters want to know what difference (if any) one or more treatments has for an outcome on average. The approach is quite consistent with the idea that causes are probability raisers. One can in fact say that a treatment is a cause when its presence raises the probability of an outcome occurring in any given case of a larger population. The experimental template underpins the definition of causality used in the population-oriented approach, a point well illustrated by King, Keohane, and Verbas (1994, pp. 76-81) famous discussion of mean causal effect (). They present this concept using the example of the effect of incumbency on the vote fraction received by a candidate running for office. The mean causal effect in the example is the difference between the mean distribution of the vote fraction received by an incumbent candidate across multiple hypothetical elections and the mean distribution of the vote fraction received by a nonincumbent across multiple hypothetical elections. In their notation, = Ii _ N , where X is the mean of the distribution of the depeni i dent variable under condition X and I and N are incumbent and nonincumbent, respectively. This understanding of as being equal to the difference between the means of two distributions is shared by many leading statistical methodologists (Braumoeller, 2006).12 Population-oriented comparativists often identify causes by estimating coefficients in multivariate statistical models. They are especially concerned with two aspects of . The first is substantive significance, which derives from the size of the coefficient and may be evaluated in various ways. The second is statistical significance, which summarizes the likelihood that the observed finding is a product of sampling error or other random error and is affected by the number of variables and cases used in the statistical test. For example, Gerring et al.s (2005) argument that democratic stock is an important cause of economic growth is based on the substantive and statistical significance of for the democracy stock variable in their multivariate models. In contemporary comparative politics, it is worth noting, independent variables of theoretical

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

422

Comparative Political Studies

interest that meet a specified level of statistical significance (e.g., p .05) are often considered important causes even if the size of their effect is modest. Indeed, the total explanatory power of most statistical models is modest when evaluated using a measure such as adjusted R2. This underscores the experimental underpinning of this tradition: The goal is to assess the average effects (if any) of individual variables rather than comprehensively explain variation on an outcome.

Net Causal Effects


Causal research in the population-oriented tradition has for decades been concerned with distinguishing genuine causation from spurious covariation (e.g., Pearl, 2000, pp. 78-85, 173-200). When one seeks to estimate the mean effect of X on Y in the context of an observational study, the question immediately arises as to which other variables should be held constant. Both the size and statistical significance of an effect can change radically depending on the other variables that are included. These other variables known as covariates, concomitants, or confoundersmust be controlled for (conditioned on) if one wishes to correctly estimate the net effect of X on Y. Net effects are essentially calculated by partitioning the population into subgroups that are homogenous on the control variables, measuring the effect of values of X on the distribution of Y in the subgroups, and then averaging the results (e.g., Turner, 1997). The goal of this form of multivariate analysis is not to deny that the control variables are causes of Y, but to rule out the possibility that they account for the observed effect of X on Y. Explanation in population-oriented research makes various specification assumptions to mitigate the potential problems that arise from the use of observational data rather than experimental data (e.g., Shadish, Cook, & Campbell, 2002). With an experiment, of course, cases are randomly assigned to different values of an independent variable, such that in theory their prior characteristics are unrelated to those values. But this kind of independence is almost never true with observational data. Hence, population-oriented researchers need to control for variables using tools such as stratification to achieve conditional independence on average (mean conditional independence). Conditional independence is, in effect, a way of trying to meet the counterfactual assumptions of causation embedded in this approach. Population-oriented researchers make other familiar and related assumptions, each of which can be justified to varying degrees, depending on the study in question. These include the following:

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

423

1. Stable Unit Treatment Value Assumptionthe idea that influences among observations do not affect outcomes and the notion that a given value on an independent variable is stable across cases (for an excellent recent discussion, see Brady & Seawright, 2004). 2. Absence of endogeneityin comparative politics, endogeneity is often discussed in conjunction with the specific problems that arise from reciprocal causation. Strictly speaking, though, endogeneity arises whenever an independent variable is correlated with the error term. 3. Temporal stabilitythe assumption that causal effects are stable over time. Unrecognized differences across time can produce missing variable bias. 4. Measurement transiencethe assumption that the act of observing cases and measuring variables does not alter causal effects or, if it does, that this impact can be modeled.

The proliferation of more and more advanced statistical techniques is part of an ongoing quest to meet these various assumptions. How successful contemporary research is at doing so is debated.13 For our purposes, this debate need not distract us any more than analogous debates over the merits of small-N methods mentioned above. Rather, we need to explore how this approach to causal analysis, which thinks about causation in terms of probability theory, can be unified with the case-oriented approach discussed above, which thinks about causation in terms of logic.

From the Cases to the Population (and Back): Toward a Unified Theory
The starting point for a unified theory is the recognition that, at the population level, causation occurs almost exclusively through INUS causes. To see why this is true, we need to explore the relationship between INUS causes and variables that exert partial effects in statistical models. As we shall see, these variables have the effects they do because they are INUS causes.

Causation at the Population Level


If one accepts that causation at the population level derives from causation at the case level, then the five kinds of causes that can apply to individual casesnecessary, sufficient, necessary and sufficient, INUS, and SUIN causesare the only kinds of causes that can operate at the population level. Four of these five types, however, are usually not relevant when analyzing populations. We saw below how two typessufficient causes and

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

424

Comparative Political Studies

necessary and sufficient causesare very infrequently used at the case level. As one generalizes to the population level, these types will, if anything, be even less common. By contrast, necessary causes are standard in case-oriented research. Yet when we shift to the population level, necessary causes become hard to find. For example, causes that were necessary for high unemployment in Germany may not have been necessary for high unemployment in France; necessary causes for particular cases often do not generalize. Indeed, with only a few possible exceptions, such as the hypothesis that nondemocracy (i.e., a dyad with a nondemocratic state) is necessary for war, comparativists have not found many necessary causes that apply to broad populations. Because SUIN causes work through necessary causes, they will also not often be discovered at the population level. This leaves INUS causes as the main type of cause that applies at the population level. If INUS causes are present, different sets of variable values lead to a particular outcome. The presence of multiple combinations of variable values that produce the same outcome is sometimes called equifinality (George & Bennett, 2005). Although equifinality is usually associated with QCA techniques (Ragin, 1987, 2000), its emphasis on multiple paths to a given outcome is implicitly present in most population-oriented research, including the additive linear models that comparativists frequently use. There are countless ways to arrive at an outcome (i.e., a particular range of values on the dependent variable) in an additive linear model. Each independent variable exerts its own effect, and each independent variable can potentially compensate for any other. One case may have the outcome of interest because it has high values on certain variables, whereas a different case arrives at the same outcome because it has high values on other variables. No variable value is necessary, but different variable values (in conjunction with the error term) are sufficient to produce the outcome. Equifinality is thus omnipresent in mainstream population-oriented research (Mahoney & Goertz, 2006). Equifinality is not the only issue to which analysts of INUS causes must attend. Another key assumption underpinning INUS causation is that variables often do not exert effects independently of one another. Instead, the effect of one variable depends on the values of one or more other variables. Variables work together as packages, not isolated factors. Boolean algebra is the usual way in which case-oriented researchers try to formally model this kind of causation (Ragin, 1987, 2000). Some statistical methodologists have developed broadly parallel tools such as Boolean probit and Boolean logit (Braumoeller, 2003). Moreover, the study of interaction effects, which is not uncommon (e.g., perhaps 25% of quantitative comparative articles now include at least one interaction term), has moved statistical research in the direction

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

425

of combinatorial causation (Kam & Franzese, 2007). Insofar as statistical comparativists start to use saturated interaction models in which all possible interactions are assessed and simplified in a top-down manner, we would essentially see an integration of QCA techniques and statistical methods. The key point here is that there are various methods that can, in principle, be used to study INUS causes and combinatorial causation across broad populations. Indeed, some leading scholars suggest that many or most outcomes at the population level need to be explained in terms of combinatorial causation (e.g., Achen, 2005; Franzese, 2003; Hall, 2003; Ragin, 1987). If they are right, one might wonder if thinking about causation in terms of the mean effects of individual variablesthe traditional focus of populationoriented researchis useful at all. Doubts are reinforced by recent writings on interaction effects in the population-oriented tradition. Methodologists have made it clear that one should include lower order constitutive terms when testing multiplicative interaction models, but one cannot interpret the coefficient on these terms as an average effect (Brambor, Clark, & Golder, 2005; Braumoeller, 2004; Kam & Franzese, 2007). As the number of interaction terms increases, the idea of thinking about coefficients as reporting the effects of individual variables starts to drop out altogether. If one were to use a saturated interaction model, in fact, one could say virtually nothing about the individual variables. With QCA models, analogously, the focus is on the causal combinations, not the individual factors (with the exception of causes that are individually necessary or sufficient). How, then, is the notion of a mean causal effect for an individual variablestill the dominant way of conceptualizing causation in political sciencerelevant to thinking about causation?

Mean Causal Effects as Symptoms of INUS Causes


The answer to this question is that independent variables that exert partial effects in well-specified statistical models are INUS causes; they are normally, in fact, important INUS causes. However, with only a few partial exceptions (Brady & Seawright, 2004; Marini & Singer, 1988; Shadish et al., 2002), the possibility that mean causal effects are symptoms of INUS causes has not been developed in the methodological literature. The relationship between a mean causal effect and an INUS cause can be better understood if we first define an important INUS cause. The general criterion is that an INUS cause becomes more important as it becomes closer to being a necessary and/or sufficient cause; beyond this, it must also be a nontrivial cause.14 Borrowing from the language of probability theory,

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

426

Comparative Political Studies

we can say that important INUS causes are probabilistically necessary and/or probabilistically sufficient causes at the population level. The extent to which they are probabilistically necessary or probabilistically sufficient can be assessed in practice with one or more available techniques (see Braumoeller & Goertz, 2000; Clark, Gilligan, & Golder, 2006; Dion, 1998; Eliason & Stryker, in press; Ragin, 2000). INUS causes that are probabilistically necessary are factors that usually or almost always have to be present for the outcome to occur. They are thus normally present in the different combinations of variable values that can each generate the outcome of interest. Arguably, smoking as a cause of lung cancer is a good example. Smoking is an INUS cause because it must combine with other environmental and biological factors to generate the cancer. At the same time, smoking may well be probabilistically necessary for lung cancerthat is, it is usually present in the various distinct combinations of variable values that generate lung cancer.15 Smoking, of course, exerts a partial effect on lung cancer in population-oriented research (i.e., it raises the probability of the outcome). The reason why smoking exerts this partial effect and thus can be treated as a probability raiser is precisely because it is an important INUS cause. The causal properties of the variable (i.e., its status as a probabilistically necessary INUS cause) lead it to manifest the population-level statistical effect. Probabilistically necessary INUS causes are often implicitly discussed in the comparative politics literature as well. An example is the presence of an economic crisis as a cause of third wave transitions to democracy. As Bermeo (1990) remarked in her review of ODonnell, Schmitter, and Whiteheads (1986) Transitions from Authoritarian Rule: It is surely not coincidental that economic crises accompanied every transformation reviewed here. The pattern suggests that economic crises might be a necessary though not sufficient incentive for the breakdown of authoritarian regimes (p. 366). Although one can now easily find cases of recent democratic transition without economic crisis (e.g., Chile), it seems fair to conclude that crisis was a probabilistically necessary cause of a third wave democratic transition (i.e., countries rarely experienced transitions during economic good times; for systematic evidence, see Haggard & Kaufman, 1995, pp. 32-36). Statistical models that show an individual effect for economic crisis on democratic transition at the population level (e.g., Gasiorowski, 1995) are picking up this variables status as a probabilistically necessary INUS cause. The other kind of important INUS causes are those that are probabilistically sufficient. These causes do not have to combine with many other (nontrivial) causes to generate the outcome of interest. Rather, probabilistically

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

427

sufficient causes can almost produce the outcome by themselves (but need some help from other variables). For instance, lung cancer is probabilistically sufficient for death. Most deaths do not involve lung cancer, but when lung cancer is present, it will generate death under most circumstances (though there are some exceptions). Again, INUS causes that are probabilistically sufficient will manifest themselves as variables that exert partial effects when included in correctly specified statistical models. Probabilistically sufficient INUS causes are also readily found in the comparative literature. For example, one of the most celebrated empirical findings in the field concerns the relationship between economic development and democracy. Although the specific workings of this relationship are subject to debate, much evidence suggests that a high level of economic development is probabilistically sufficient for either democracy or democratic stability (e.g., Boix & Stokes, 2003; Przeworski, Alvarez, Cheibub, & Limongi, 2000). Rich countries are thus almost always democratic. There are some exceptions, and certainly one would not want to say that wealth alone is fully sufficient for democracy (or democratic stability). However, when significant wealth is present, one can be nearly certain that democracy (or democratic stability) will also be present. Statistical results that show that economic development has a significant effect on democracy (e.g., Londregan & Poole, 1996) are likely an artifact of the probabilistically sufficient status of this variable. Again, the implication is that the variable shows up as a probability raiser because it is an important INUS cause. We may conclude, then, that individual variables that reveal partial effects in correctly specified statistical models are causes. But the effects they exhibit as summarized by a statistical coefficient are not what make them causes. They are causes by virtue of being important INUS causes (i.e., they are probabilistically necessary and/or sufficient for outcomes). And their exhibited effects are just symptoms of this underlying logical status.16

Translating Back and Forth


Treating variables that exhibit mean causal effects in statistical models as INUS causes is helpful in several ways. For one thing, it is substantively enlightening to know whether a variable exerts its effect because it is probabilistically necessary or because it is probabilistically sufficient (or both). For example, the finding that a higher level of economic development is almost always sufficient for democracy is of obvious substantive importance. It tells us that the promotion of economic growth beyond a certain level nearly guarantees the promotion of human freedom. The same is

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

428

Comparative Political Studies

true of the finding that high levels of cumulative left governance are almost always necessary for high levels of public day care (Stryker, Eliason, & Tranby, in press). The implication is that public day care programs will be dislodged only with a long-term shift to the right in government. Although these kinds of findings are not about partial effects as traditionally understood, they are clearly useful if correct. Moreover, by translating an initial statistical finding into the language of INUS causation, the researcher can better theorize how the factor should be included in a subsequent model that assesses combinatorial causation. In the case of INUS causes that are probabilistically necessary, they must be included in nearly all causal combinations of a QCA model (and would normally need to appear in many terms in a multiplicative interaction model as well). For example, because nondemocracy (i.e., a nondemocratic dyad) is at least probabilistically necessary for warfare, this factor must be included as one cause in nearly all of the causal combinations of a QCA model of interstate warfare. INUS causes that are probabilistically sufficient, by contrast, often do not need to be included in causal combinations. This is true because these causes can generate the outcome of interest almost by themselves, needing only a little help from other (nontrivial) factors. For instance, if the presence of a large historically exploited minority group is nearly sufficient for initially low levels of social performance in Spanish America (Mahoney, 2003), then this cause can be analyzed reasonably well by itself without including it as part of a larger causal package. The population-oriented approach can thus help researchers locate important INUS causes. And the INUS causes can then be evaluated subsequently with more complex QCA or interaction models. The importance of finding the right variables and combinations to be included in these models can hardly be overstated. Our theories in comparative politics routinely posit complex interactions (Hall, 2003), but they are almost never precise enough to locate all relevant INUS causes, with potentially devastating implications for testing causal theories in practice (e.g., Kam & Franzese, 2007). Tools for better specifying our combinatorial causal models are of tremendous value. Beyond this, of course, researchers will always be interested in trying to estimate the mean effects of individual variables in their own right. Information about mean effects helps one make informed predictions and estimate the future impact of a given intervention. Individuals and political actors seek this kind of information as they maneuver in the world. For example, it is one thing to know that authoritarianism is a probabilistically necessary INUS cause of social revolution (or that smoking is a probabilistically necessary INUS cause of lung cancer). This tells us that governments

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

429

can ward off social revolutions by promoting formal democratic institutions (or that we can be fairly certain to avoid lung cancer by not being exposed to tobacco smoke). Yet it is another thing to have a good estimate of the risk of social revolution a government faces by having authoritarian institutions (or the risk of lung cancer we face by smoking). The true probability of the outcome occurring at the case level is 1 or 0. However, in the absence of full knowledge about all of the sufficiency combinations that can produce the outcome, we would like to know if the probability is more likely 1 or 0, and exactly how much more likely! We seek insight about what we can most reasonably expect to happen in a context of limited knowledge. The population-oriented tradition can help decision makers formulate better guesses about whether, given a particular course of action and other known conditions in the world, the probability of a specific outcome of interest occurring is really 1 or 0.17 These observations bring us full circle. Just as one might want to translate findings from population-oriented research into the causal language of the case-oriented tradition, so too might one wish to translate findings about probabilistically necessary and probabilistically sufficient INUS causes into the language of mean causal effects.

Conclusion
This article has explored the unification of case-oriented and populationoriented approaches to causality in comparative politics. The method of unification proposed here entailed extending the case-oriented approach to the population level, under the assumption that causal patterns at the population level must be derivative of causation that is occurring at the case level. With this extension, it became clear that INUS causes are really what are indirectly studied in most population-oriented research. In particular, important INUS causes (i.e., probabilistically necessary and/or sufficient INUS causes) manifest themselves as variables that exhibit partial causal effects in well-specified statistical models. A basic implication of this analysis is that more comparative researchers need to try to study INUS causes as directly as possible at the population level, using statistical models with various interaction terms or the tools of QCA. To be sure, these researchers will face daunting methodological challenges in the effort to generate interpretable results given the tremendous data requirements of these models. But this kind of complexity-sensitive analysis may be the only road, over the long run, to valid causal generalization across

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

430

Comparative Political Studies

many cases in comparative politics. Meanwhile, those who work on this agenda will surely benefit from the insights of ongoing research on both mean effects for individual variables and specific causes in particular cases.

Notes
1. Here, I modify Ragins (1987) terminology, which distinguishes between case-oriented and variable-oriented approaches. The distinction between these two approaches also appears in the philosophical literature, sometimes identified as the difference between token causation and type causation, or between single-case causation and property causation (see, e.g., Woodward, 2003). 2. This is known as the singularist view in philosophy. Its advocates include Anscombe (1971), Mellor (1995), and Woodward (1990). The rival position is known as the supervenience view, and it is exemplified by Eells (1991). 3. Gerring (2005) adopts this approach for political science. Philosophers who think about causation at the level of the population have also long embraced this definition. For example, A probabilistic accountessentially the idea that a cause raises the probability of its effect is now commonplace in science and philosophy (Dowe & Noordhof, 2004, p. 1). Also see Cartwright (1989); Eells (1991); Eells and Sober (1983); Good (1961, 1962); Lewis (2000); Menzies (2004); Pearl (2000); Skyrms (1980); Suppes (1970). 4. For instance, Imagine a golf ball rolling towards the cup, such that its probability of dropping in is quite high. Along comes a squirrel, who kicks the ball away from the cup, thus reducing the chance of it going in. Then, through a series of improbable ricochets, the ball drops into the cup. How are the squirrels kick and the balls dropping in the cup related? I want to argue that the squirrels kick reduced the probability of the balls dropping in and the squirrels kick caused the ball to drop in (Sober, 1984, pp. 406-407). A standard objection to this kind of example is that if one characterizes the specific causes (i.e., the squirrels kick and subsequent ricochets) in great enough detail, then it will become clear that this intervention did not really lower the probability of the ball dropping in. I agree; the real probability of the ball dropping in was always equal to 1. Probabilities do not make sense for the explanation of singular events (as I discuss below). Also see Dowe (2004), on chance-lowering causes; and Schaffer (2000), on probability raising without causation. 5. The major attempt to defend single-case probabilities is Poppers (1959) propensity interpretation. But his approach is widely rejected (see Williams, 1999). 6. Seemingly random processes such as dice rolling and coin flipping are not really random (Ekeland 1988; Ford, 1983). It is true that dice rolling and coin flipping exhibit extreme sensitivity to initial conditions and may be good examples of the dynamics of chaos theory. However, chaos theory is fully a deterministic approach to explanation (Waldner, 2002). 7. The problems that redundant causation (also called symmetrical overdetermination) poses for necessity definitions of cause are very well known. See Ehring (1997); Menzies (1996); Scriven (1966). 8. The Humean notion of sufficiency understood in terms of constant conjunction subsequently became the basis for the covering-law model of causation associated with mid-20th century positivists (e.g., Hempel, 1965; Nagel, 1961; Popper, 1959). 9. Examples of comparative works include Amentas (1996) study of New Deal social spending; Berg-Schlosser and DeMeurs (1994) analysis of democracy in interwar Europe; Huber, Ragin, and Stephenss (1993) work on the welfare state; Griffin, Botsko, Wahl, and Isaacs (1991)

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

431

study of trade union growth and decline; Mahoneys (2003) work on long-run development in Latin America; and Wickham-Crowleys (1992) study of guerrillas and revolutions. 10. Determinism is the thesis that there is at any instant exactly one physically possible future (Van Inwagen, 1983, p. 3). For a discussion of common misconceptions of the entailments of determinism, such as the incorrect notion that determinism denies free will and human agency, see Dennett (2003). 11. Angrist, Imbens, and Rubin (1996) assert that causal inference in statistics . . . is fundamentally based on the randomized experiment (p. 444). 12. King, Keohane, and Verba (1994) follow Holland (1986) in referring to mean causal effect. The same idea is varyingly called average treatment effect (Sobel, 1996), average causal response (Angrist & Imbens, 1995), and average causal effect (Dawid, 2000). 13. For some skeptical views, see Clogg and Haritou (1997); Freedman (1991); Lieberson (1985); Shalev (2007). 14. Empirically speaking, a necessary cause is trivial if it is always (or almost always) present, irrespective of the outcome. A sufficient cause is trivial if it is almost never present and thus almost never generates the outcome of interest (see Braumoeller & Goertz, 2000; Goertz, 2006; Ragin, 2006; also see Hart & Honor, 1959). For example, the presence of oxygen is a trivially necessary cause of a revolution, because it is always present in all cases. INUS causes that are close to being trivial in this empirical sense are not important. 15. Apparently, there are relatively few cases (i.e., about 10%) of lung cancer in which smoking is not involved. 16. To be sure, not all INUS causes will exert mean effects in population-oriented research. For example, it is easy to write a Boolean equation in which a variable and its negation are both INUS causes of the same outcome (e.g., Y = A&B v A&D). When this is true, the variable and its negation may end up canceling out one anothers effect, such that the variable overall exerts no mean effect in a population. 17. In the population-oriented tradition, a respected body of work views causation in terms of this kind of decision maker understanding (e.g., Dawid, 2000; Rubin, 1978).

References
Achen, C. H. (2005). Two cheers for Charles Ragin. Studies in Comparative International Development, 40, 27-32. Albert, M. (2005). Should Bayesians bet where frequentists fear to tread? Philosophy of Science, 72, 584-593. Amenta E. (1996). Social politics in context: The institutional politics theory and social spending at the end of the new deal. Social Forces, 75, 33-60. Amsden, A. (1992). A theory of government intervention in late industrialization. In L. Putterman & D. Ruecschemeyer (Eds.), State and market in development: Synergy or rivalry? (pp. 53-84). Boulder, CO: Lynne Reinner. Angrist, J. D., & Imbens, G. W. (1995). Two-stage least squares estimation of average causal effects in models with treatment intensity. Journal of the American Statistical Association, 90, 431-442. Angrist, J. D., Imbens, G. W., & Rubin, D. B. (1996). Identification of causal effects using instrumental variables. Journal of the American Statistical Association, 91, 444-455. Anscombe, G. E. M. (1971). Causality and determination. Cambridge, UK: Cambridge University Press.

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

432

Comparative Political Studies

Appleby, D. M. (2004). Probabilities are single-case or nothing. Optics and Spectroscopy, 99, 447-456. Berg-Schlosser D., & DeMeur, G. (1994). Conditions of democracy in interwar Europe: A Boolean test of major hypotheses. Comparative Politics, 26, 253-279. Bermeo, N. (1990). Rethinking regime change. Comparative Politics, 22, 359-377. Boix, C., & Stokes, S. C. (2003). Endogenous democratization. World Politics, 55, 517-549. Brady, H. E., & Seawright, J. (2004, August). Framing social inquiry: From models of causation to statistically based causal inference. Paper presented at the Annual Meeting of the American Political Science Association, Chicago. Brambor, T., Clark, W. R., & Golder, M. (2005). Understanding interaction models: Improving empirical analyses. Political Analysis, 14, 63-82. Braumoeller, B. F. (2003). Causal complexity and the study of politics. Political Analysis, 11, 209-233. Braumoeller, B. F. (2004). Hypothesis testing and multiplicative interaction terms. International Organization, 58, 807-850. Braumoeller, B. F. (2006). Explaining variance; or, stuck in a moment we cant get out of. Political Analysis, 14, 268-290. Braumoeller, B. F., & Goertz, G. (2000). The methodology of necessary conditions. American Journal of Political Science, 44, 844-858. Cartwright, N. (1989). Natures capacities and their measurement. Oxford, UK: Clarendon. Cheibub, J. A. (2007). Presidentialism, parliamentarism, and democracy. Cambridge, UK: Cambridge University Press. Clark, W. R., Gilligan, M. J., & Golder, M. (2006). A simple multivariate test for asymmetric hypotheses. Political Analysis, 14, 311-331. Clogg, C. C., & Haritou, A. (1997). The regression method of causal inference and a dilemma confronting this method. In V. R. McKim & S. P. Turner (Eds.), Causality in crisis? Statistical methods and the search for causal knowledge in the social sciences (pp. 83-112). Notre Dame, IN: University of Notre Dame Press. Collier, D. (1979). The new authoritarianism in Latin America. Princeton, NJ: Princeton University Press. Collier, R. B. (1999). Paths toward democracy. New York: Cambridge University Press. Dawid, P. (2000). Causal inference without counterfactuals. Journal of the American Statistical Association, 95, 407-450. Dennett, D. C. (2003). Freedom evolves. New York: Viking. Dion, D. (1998). Evidence and inference in the comparative case study. Comparative Politics, 30, 127-146. Dowe, P. (2004). Chance-lowering causes. In P. Dowe & P. Noordhof (Eds.), Cause and chance: Causation in an indeterministic world (pp. 28-38). London: Routledge. Dowe, P., & Noordhof, P. (2004). Introduction. In P. Dowe & P. Noordhof (Eds.), Cause and chance: Causation in an indeterministic world (pp. 1-11). London: Routledge. Eells, E. (1991). Probabilistic causality. Cambridge, UK: Cambridge University Press. Eells, E., & Sober, E. (1983). Probabilistic causality and the question of transitivity. Philosophy of Science, 50, 35-57. Ehring, D. (1997). Causation and persistence. New York: Oxford University Press. Ekeland, I. (1988). Mathematics and the unexpected. Chicago: University of Chicago Press. Eliason, S., & Stryker, R. (in press). Goodness-of-fit tests and descriptive measures in fuzzy-set analysis. Sociological Methods & Research.

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

433

Elman, C. (2005). Explanatory typologies in qualitative studies of international politics. International Organization, 59, 293-326. Ford, J. (1983). How random is a coin toss? Physics Today, 36, 40-47. Franzese, R. J., Jr. (2003). Quantitative empirical methods and the context-conditionality of class and modern comparative politics. CP: Newsletter of the Comparative Politics Organized Section of the American Political Science Association, 14, 20-24. Freedman, D. A. (1991). Statistical models and shoe leather. Sociological Methodology, 21, 291-313. Garrett, G. (1998). Partisan politics in the global economy. Cambridge, UK: Cambridge University Press. Gasiorowski, M. J. (1995). Economic crisis and political regime change: An event history analysis. American Political Science Review, 89, 882-889. George, A. L., & Bennett, A. (2005). Case studies and theory development in the social sciences. Cambridge, MA: MIT Press. Gerring, J. (2005). Causation: A unified framework for the social sciences. Journal of Theoretical Politics, 17, 163-198. Gerring, J., Bond, P., Barndt, W., & Moreno, C. (2005). Democracy and growth: A historical perspective. World Politics, 57, 323-364. Goertz, G. (2003). Cause, correlation, and necessary conditions. In G. Goertz & H. Starr (Eds.), Necessary conditions: Theory, methodology, and applications (pp. 47-64). Lanham, MD: Rowman & Littlefield. Goertz, G. (2006). Assessing the trivialness, relevance, and relative importance of necessary or sufficient conditions in social science. Studies in Comparative International Development, 41, 88-109. Goertz, G., & Levy, J. S. (Eds.). (2007). Explaining war and peace: Case studies and necessary condition counterfactuals. New York: Routledge. Goldhagen, D. J. (1997). Hitlers willing executioners: Ordinary Germans and the holocaust. New York: Vintage. Good, I. J. (1961). A causal calculus I. British Journal for the Philosophy of Science, 11(44), 305-318. Good, I. J. (1962). A causal calculus II. British Journal for the Philosophy of Science, 12(45), 43-51. Griffin, L. J., Botsko, C., Wahl, A., & Isaac, L. W. (1991). Theoretical generality, case particularity: Qualitative comparative analysis of union growth and decline. International Journal of Comparative Sociology, 32, 110-136. Haggard, S., & Kaufman, R. R. (1995). The political economy of democratic transitions. Princeton, NJ: Princeton University Press. Hall, P. (2003). Aligning ontology and methodology in comparative research. In J. Mahoney & D. Rueschemeyer (Eds.), Comparative historical analysis in the social sciences (pp. 373-404). Cambridge, UK: Cambridge University Press. Hart, H. L. A., & Honor, A. M. (1959). Causation in the law. Oxford, UK: Oxford University Press. Hempel, C. (1965). Aspects of scientific explanation and other essays in the logic of explanation. New York: Free Press. Holland, P. W. (1986). Statistics and causal inference (in theory and methods). Journal of the American Statistical Association, 81, 945-960. Huber E., Ragin, C. C., & Stephens, J. D. (1993). Social democracy, Christian democracy, constitutional structure, and the welfare state. American Journal of Sociology, 99, 711-749.

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

434

Comparative Political Studies

Humphreys, P. (1989). The chances of explanation: Causal explanation in the social, medical and physical sciences. Princeton, NJ: Princeton University Press. Huntington, S. P. (1968). Political order in changing societies. New Haven, CT: Yale University Press. Kam, C. D., & Franzese, R. J. (2007). Modeling and interpreting interaction hypotheses in regression analysis: A brief refresher and some practical advice. Ann Arbor: University of Michigan Press. King, G., Keohane, R. O., & Verba, S. (1994). Designing social inquiry: Scientific inference in qualitative research. Princeton, NJ: Princeton University Press. Kohli, A. (Ed.). (1988). Indias democracy: An analysis of changing state-society relations. Princeton, NJ: Princeton University Press. Kohli, A. (Ed.). (2001). The success of Indias democracy. Cambridge, UK: Cambridge University Press. Levitsky, S., & Way, L. A. (1998). Between a shock and a hard place: The dynamics of laborbacked adjustment in Poland and Argentina. Comparative Politics, 30, 171-192. Lewis, D. (1986). Philosophical papers (Vol. 2). Oxford, UK: Oxford University Press. Lewis, D. (2000). Causation as influence. Journal of Philosophy, 97, 182-197. Lieberson, S. (1985). Making it count: The improvement of social research and theory. Berkeley: University of California Press. Londregan, J. B., & Poole, K. T. (1996). Does high income promote democracy? World Politics, 49, 1-30. Luebbert, G. M. (1991). Liberalism, fascism, or social democracy: Social classes and the political origins of regimes in interwar Europe. New York: Oxford University Press. Mackie, J. L. (1965). Causes and conditions. American Philosophical Quarterly, 2, 245-264. Mackie, J. L. (1980). Cement of the universe: A study of causation. Oxford, UK: Oxford University Press. Mahoney, J. (1999). Nominal, ordinal, and narrative appraisal in macrocausal analysis. American Journal of Sociology, 104, 1154-1196. Mahoney J. (2003). Long-run development and the legacy of colonialism in Spanish America. American Journal of Sociology, 109, 50-106. Mahoney, J., & Goertz, G. (2006). A tale of two cultures: Contrasting quantitative and qualitative research. Political Analysis, 14, 227-249. Mahoney, J., Kimball, E., & Koivu, K. (2007, August). The causal logic of historical explanation. Paper presented at the Annual Meeting of the American Political Science Association, Chicago. Mainwaring, S. (1993). Presidentialism, multipartism, and democracy: The difficult combination. Comparative Political Studies, 26, 198-228. Mainwaring, S., & Prez-Lin, A. (2003). Level of development and democracy: Latin American exceptionalism, 1945-1996. Comparative Political Studies, 36, 1031-1067. Marini, M. M., & Singer, B. (1988). Causality in the social sciences. Sociological Methodology, 18, 347-409. Mellor, D. H. (1995). The facts of causation. London: Routledge. Menzies, P. (1996). Probabilistic causation and the pre-emption problem. Mind, 105, 85-117. Menzies, P. (2004). Difference-making in context. In J. Collins, N. Hall, & L. A. Paul (Eds.), Causation and counterfactuals (pp. 139-180). Cambridge, MA: MIT Press. Milne, P. (1986). Can there be a realist single-case interpretation of probability? Erkenntnis, 25, 129-132.

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

Mahoney / Causality

435

Moore, B., Jr. (1966). Social origins of dictatorship and democracy: Lord and peasant in the making of the modern world. Boston: Beacon. Nagel, E. (1961). The structure of science. New York: Harcourt, Brace, and World. ODonnell, G. (1973). Modernization and bureaucratic-authoritarianism: Studies in South American politics. Berkeley: University of California, Institute of International Studies. ODonnell, G., Schmitter, P., & Whitehead, L. (Eds.). (1986). Transitions from authoritarian rule. Baltimore: Johns Hopkins University Press. Pearl, J. (2000). Causality: Models, reasoning, and inference. Cambridge, UK: Cambridge University Press. Popper, K. (1959). The logic of scientific discovery. London: Hutchinson. Przeworski, A., Alvarez, M. E., Cheibub, J. A., & Limongi, F. (2000). Democracy and development: Political institutions and well-being in the world, 1950-1990. New York: Cambridge University Press. Ragin, C. C. (1987). The comparative method: Moving beyond qualitative and quantitative strategies. Berkeley: University of California Press. Ragin, C. C. (2000). Fuzzy-set social science. Chicago: University of Chicago Press. Ragin, C. C. (2006). Set relations in social research: Evaluating their consistency and coverage. Political Analysis, 14, 291-310. Rubin, D. B. (1978). Bayesian inference for causal effects: The role of randomization. Annals of Statistics, 6, 34-58. Rudolph, L. I., & Rudolph, S. H. (1959). Toward political stability in underdeveloped countries: The case of India. Public Policy, 9, 149-178. Schaffer, J. (2000). Overlappings: Probability-raising without causation. Australasian Journal of Philosophy, 78, 40-46. Scriven, M. (1966). Defects of the necessary condition analysis of causation. In W. H. Dray (Ed.), Philosophical analysis and history (pp. 258-262). New York: Harper and Row. Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Boston: Houghton Mifflin. Shalev, M. (2007). Limits and alternatives to multiple regression in comparative research. Comparative Social Research, 24, 261-308. Skocpol, T. (1992). Protecting soldiers and mothers: The political origins of social policy in the United States. Cambridge, MA: Harvard University Press. Skyrms, B. (1980). Causal necessity: A pragmatic investigation of the necessity of laws. New Haven, CT: Yale University Press. Sobel, M. E. (1996). An introduction to causal inference. Sociological Methods and Research, 24, 353-379. Sober, E. (1984). Two concepts of cause. Philosophy of Science Association, 2, 405-424. Stryker, R., Eliason, S., & Tranby, E. (in press). The welfare state, family policies, and womens labor force participation: Combining fuzzy-set and statistical methods to assess causal relations and estimate causal effects. In L. Kenworthy & A. Hicks (Eds.), Method and substance in macro-comparative analysis. Houndmills, Hampshire, UK: Palgrave Macmillan. Suppes, P. (1970). A probabilistic theory of causality. Amsterdam: North-Holland. Tarrow, S. (1989). Democracy and disorder: Protest and politics in Italy, 1965-1975. Oxford, UK: Oxford University Press. Turner, S. (1997). Net effects: A short history. In V. R. McKim & S. P. Turner (Eds.), Causality in crisis? Statistical methods and the search for causal knowledge in the social sciences (pp. 23-45). Notre Dame, IN: University of Notre Dame Press. Van Inwagen, P. (1983). An essay on free will. Oxford, UK: Clarendon.

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

436

Comparative Political Studies

Varshney, A. (1998). Why democracy survives. Journal of Democracy, 9, 36-50. Waldner, D. (1999). State-building and late development. Ithaca, NY: Cornell University Press. Waldner, D. (2002, August). Anti anti-determinism: Or what happens when Schrdingers cat and Lorenzs butterfly meet Laplaces demon in the study of political and economic development. Paper delivered at the Annual Meeting of the American Political Science Association, Boston. Waldner, D. (2006). Indian exceptionalism in Pakistani perspective. Unpublished manuscript, University of Virginia. Weyland, K. (1998). The political fate of market reform in Latin America, Africa, and Eastern Europe. International Studies Quarterly, 42, 645-674. Wickham-Crowley, T. (1992). Guerrillas and revolution in Latin America: A comparative study of insurgents and regimes since 1956. Princeton, NJ: Princeton University Press. Williams, M. (1999). Single case probabilities and the social world: The application of Poppers propensity interpretation. Journal for the Theory of Social Behavior, 29, 187-201. Woodward, J. (1990). Supervenience and singular causal claims. In D. Knowles (Ed.), Explanation and its limits (pp. 211-246). Cambridge, UK: Cambridge University Press. Woodward, J. (2003). Making things happen: A theory of causal explanation. Oxford, UK: Oxford University Press. Yashar, D. J. (1997). Demanding democracy: Reform and reaction in Costa Rica and Guatemala, 1870s-1950s. Stanford, CA: Stanford University Press.

James Mahoney is a professor of political science and sociology at Northwestern University. He is the author of The Legacies of Liberalism: Path Dependence and Political Regimes in Central America and coeditor of Comparative Historical Analysis in the Social Sciences. His work also includes articles on political and socioeconomic development in Latin America, path dependence in historical sociology, and causal inference in small-N analysis.

Downloaded from http://cps.sagepub.com at NORTHWESTERN UNIV LIBRARY on July 29, 2008 2008 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.