Saturday, February 28, 2015

Three concepts of social structure



There is an important ambiguity in the idea of social structure that needs to be addressed. The word is sometimes used to refer to functioning entities or units within society. The state is a structure within society; likewise is the system of public education. A corporation is a structure, and so is the feature of the economic organization of society that defines the role and possibilities of a corporation. This aspect of the concept of social structure as persistent entity with causal and functional properties was the subject of an earlier post (link).

There is a second meaning associated with "structure" as well: the idea that society possesses a structure of interconnected parts and systems, and that the parts influence each other in systematic ways. This aspect of the concept is discussed in an earlier post, focusing on Marx's view that a society possesses a mode of production (link). We might call this "system structure".

But the term can also be used to refer to the structure of society. Here we have statements like "the age structure of Egypt is X" and "the occupational structure of Great Britain is Y". In this usage, we are being directed to a descriptive or statistical feature of society -- the way that various elements hang together. In this case the structural feature is an outcome, not a functional part. Income stratification is a structural feature; the labor market is a structural entity. Other examples might include the rates of racial segregation that exist in different parts of the country; the wage gaps that exist between male and female workers; and gaps in college attendance rates between white and black students. In each case the structural feature is a statistical pattern that exists across socially meaningful groups, and that is presumably the effect of broad social forces and processes -- for example, the sorting mechanisms that influence residence, school, and work by race and gender.

Like the first two interpretations, this third component of the meaning of social structure possesses causal significance, both as effect and as cause. Take wealth stratification. The distribution of wealth in a society is the causal and systemic consequence of other structural features -- the rules determining which individuals and groups gain control of which components of the economic surplus. This is the heart of the definition of the field of structuralist economics in the hands of economists like Lance Taylor (Reconstructing Macroeconomics: Structuralist Proposals and Critiques of the Mainstream). Given how the economic structure works ("structure as system"), the current distribution of income and wealth ensues ("structure as outcome"). Structural characteristics in this sense are outcomes of the workings of the other kinds of structures.

But outcome structures have causal consequences of their own. The concentration of wealth in contemporary market democracies creates a group of people with both the means and the interest to influence political decision making. So the concentration of wealth and income that Thomas Piketty describes has a causal consequence for the politics of the democracies of countries like the United States and South Korea. Or take a different example, the age structure of Egypt. The fact that Egypt's age distribution is substantially skewed to a younger profile than that of the United States has economic and political consequences. There aren't enough job opportunities for the large population of teenagers and young adults in Egypt, which produces a potential for political instability as young disaffected Egyptians demand change. Or consider the economic consequences of China's skewed age distribution in the opposite direction: there aren't enough young workers to support the needs of the elderly population. In each case important social and political consequences are caused by the structural fact of the distribution of demographics or wealth. Consider the dramatically different age structures represented by the demographic pyramids of Egypt, China, and the United States:




These sorts of structural features represent features of society that are analogous to physiological parameters for a biological organism (body temperature, blood pressure, insulin levels). They are outcome parameters that are determined by the workings of other more fundamental forces and processes in society. And, like physiological parameters, there are some clear instances where homeostatic social processes react to correct the outcomes when values of the parameter go too far out of the range of the normal. (For example, extreme wealth inequality may stimulate political forces that lead to higher taxation rates on the rich.) That said, outcome structural features seem to play a less dynamic role in social processes than system and entity structures.


Thursday, February 26, 2015

Sellars and Bhaskar


Robert Brandom has just published a highly interesting book about the coherence of the philosophical thought of Wilfrid Sellars, From Empiricism to Expressivism: Brandom Reads Sellars. In reading the book I am brought back repeatedly to thinking about Roy Bhaskar. There is the affirmation of scientific realism that both endorse; there is the appeal to Kant over Hume; and there is in both philosophers a somewhat surprising and camouflaged pragmatism.

Realism. Sellars is a scientific realist, in the sense that he believes that what exists in a given domain is specified by the concepts of an adequately developed scientific set of theories of that domain. The language of extensively developed science delimits what is "real". Here is Brandom:
I think Sellars is simply and evidently correct in endorsing scientific realism about theoretical entities, as opposed to any sort of instrumentalism. He applies this view in the philosophy of mind to yield important conclusions. For he understands behaviorism as instrumentalism about mental or psychological entities. Among the initially theoretical, postulated entities that later become observable on Sellars's account are thoughts and sense impressions, according to the Myth of Jones at the end of Empiricism and the Philosophy of Mind. It is of the essence of his “philosophical behaviorism,” as opposed to Ryle's “logical behaviorism” (which requires strict definability of mental-theoretical vocabulary in terms of behavioral-observable vocabulary), to be realistic in this sense about theoretical entities postulated to explain observable behavior. (60)
The role of Kant. Sellars' realism is closely intertwined with his appreciation of Kant's metaphysics. Brandom summarizes Sellars' most important effort as being a sustained attempt to move analytic philosophy from Hume to Kant. Here is a version of what that means in the context of Sellars' critique of the brand of empiricism embedded in analytic philosophy of the 1940s and 1950s:
Besides concepts used in empirical description and explanation, there are also concepts whose expressive role it is to make explicit necessary features of the framework that makes empirical description and explanation possible. (2)
This is the "expressivism" of the title: the positivist idea that the vocabulary of science divides between empirical descriptive terms and logical connectives (synthetic and analytic) fails in a very important way. Instead, there are concepts that are substantive but non-empirical, and these concepts play a crucial role in science and philosophy.

Normativeness. Another of Sellars' central insights, according to Brandom, is his conviction that epistemic and semantic vocabulary is inherently normative. To assert that "I know that P" is to make a commitment to the listener: I have credible and compelling justification for believing that P. Epistemic and semantic assertions bring with them normative presuppositions on behalf of the claimant. Bhaskar too believes that philosophy, science, and normative commitments are intertwined. I don't claim that they come to similar conclusions, but Sellars too rejects the strong fact-value distinction that was characteristic of positivism and much analytic philosophy.

Pragmatism. A third important and original theme in From Empiricism to Expressivism is Brandon's discovery that pragmatism plays an important but seldom expressed role in Sellars' thought. The classic idea of language as use, and the symmetry between semantics,syntax, and pragmatics, play important roles in Sellars' reasoning and philosophical positions.
By ‘pragmatism' in this connection I mean that the project of offering a metalinguistic reading of framework-explicating nondescriptive concepts such as modal, normative, and ontological ones is conducted in terms of pragmatic metavocabularies: vocabularies for talking about the use of expressions, about discursive social practices. (5)
Non-extensional properties and modalism. Another important part of Sellars' work, according to Brandom, is his effort to make a new kind of sense of modal philosophy and counterfactual semantics. This leads him to what Brandom refers to as a "broadly aristotelian metaphysical understanding of objects and properties" (70). This creates another potential dimension of relationship between Sellars and Bhaskar -- the metaphysics of powers and causal capacities. Sellars wants to create space for a kind of necessity that is distinct from logical necessity but stronger than Humean constant conjunction. Brandom refers to Sellars' concept here as "causal" modality (184). (This seems compatible with the argument made earlier for causal necessity; link.)

Brandom extends these ideas further by constructing s view that he calls "modal realism":
This sketch of a program for extending the Kant-Sellars tradition of modal expressivism raises a myriad of questions, some of detail, others more substantial. Rather than beginning to fi ll in that sketch by addressing some of those questions, I want to confront the ideas that motivate it with a different set of intuitions: those that motivate a robust modal realism. By “modal realism” I mean the conjunction of the claims that

MR1) Some modally qualified claims are true.
MR2) Those that are state facts.
MR3) Some of those facts are objective, in the sense that they are independent of the activities of concept users: they would be facts even if there never were or never had been concept users.

There are strong reasons to endorse all three of these claims. (194-195)
Here again, it seems that critical realism theorists are well advised to consider this intellectual relationship to Sellars. And Brandom's book is an excellent place to start.

So I would be tempted to argue that Sellars should be added to the mix of arguments and positive efforts to extend the theories of critical realism and causal powers. What do you think, Ruth Groff, Stephen Mumford, and Rani Lill Anjum?

Saturday, February 21, 2015

Social structures as entities



Image: first organizational chart (link)


One mark of modernity is the fact that we swim in a sea of structures -- states, markets, militaries, employment systems, social networks, taxation systems, and systems of racial and gender disadvantage, to name several. In place of a perhaps mythical pre-modern society in which social life was mediated by personal relationships, the modern world is mediated by institutions, structures, and systems of impersonal rules. Relations of patron and client and the influence of the village priest illustrate the model of pre-modern personalism and particularism. The bureaucratic state that Hegel describes and the impersonal markets that Weber describes illustrate powerful examples of modern impersonal institutions. Essentially this is the story of traditionalists versus modernists -- Thomas Carlyle, Sartor Resartus, against GWF Hegel, Hegel's Philosophy of Right and Max Weber, The Theory of Social and Economic Organization.

The images above illustrate different aspects of the idea of a social structure as an entity. The first represents the ties of authority and information flow that exist within the organization, allowing executives to control the behavior and properties of the organization. The second illustrates the structuring of the labor force that results from the changing environment within which individuals confront the labor market.

In ordinary thinking, structures have enduring properties that are largely independent of the individuals whom they encompass; they affect the behavior of individuals within them, and they affect the outcomes that individuals achieve through their efforts; and they are causally influential in large processes of social change and social stability. Workers enter a labor market that is largely independent of their own choices and actions. Structures are things with sufficient fixity over time and sufficient firmness to permit them to be regarded as social entities.

These are the ideas of ordinary commonsense about structures -- the contents of folk sociology. But how have gifted sociologists theorized this crucial element of the social world? It turns out that things are less simple than commonsense would have it.

First, many sociologists agree on the basic point that social structures exist with reasonably stable properties and that they influence the behavior of individuals. Examples include the government of the state of Michigan, the system of labor organization and representation in the United States, and the vertically organized corporation. In each case the social entity consists of a set of roles, regulations, and actors, along with a set of physical and financial resources, that confront the citizens and consumers who interact with them as objective realities. Developers go to the appropriate agencies for permits and conform to the requirements of state law and state and local inspectors. Criminals conform to the requirements of their organizations as well. Institutions and organizations regulate and constrain the lives of individuals. (For a brief treatment of the Black Panthers as a bureaucratic organization see this post on Mayer Zald's theories of organizational behavior; link.)

But second, it is evident that structures are populated and constituted by individuals. So whatever causal powers the structures may have and whatever persistent features they possess must somehow be embodied in the actions and states of mind of those individuals. The effects of the structure on individuals and other social entities and processes depend on the coordinated actions of individuals within the structure. Individual actions within the structure are constrained by the meanings, rules, meanings, and enforcement mechanisms that exist within the structure. And external individual actions and social arrangements are affected by the knowledge and representations that these external players have of the workings of the structure. Structures rest upon microfoundations (link). And there must be specific processes of inculcation, coordination, and enforcement through which the required forms of action transpire at the level of the actors. (Research on "institutional logics" helps to fill out this story; link.) Anthony Giddens puts the point this way in Central Problems in Social Theory: Action, Structure, and Contradiction in Social Analysis:
The term 'social structure' thus tends to include two elements, not clearly distinguished from one another: the patterning of interaction, as implying relations between actors or groups; and the continuity of interaction in time. (62)
Third, structures are plastic and frangible (link). Zygmunt Bauman makes this point in a particularly provocative way by insisting the social world is "liquid" (link). The metaphor overstates the case, since many structures have significantly more stability than the image of liquidity implies; but the fundamental notion that social structures and other entities are subject to change is certainly a valid one. We can observe the drift in the functioning and roles of institutions in every aspect of life -- political, economic, educational, and ideological.

Several fundamental kinds of questions must be answered when we contemplate the idea of structures and organizations possessing persistent properties over time. First is the question of the conformant behavior of role players within the organization at various points of time. What are the microfoundations of compliance? Individuals do not conform simply because the employee handbook specifies that they should. Rather, they need to be incentivized, trained, motivated, supervised, and disciplined, in order to bring about the forms of orderly behavior that the organization requires. And this means that an organization depends on the existence of roles presenting incentives and constraints for supervisors and trainers as well.

Second is a larger question of accounting for the stability of the rules and roles themselves. What factors work against entropy and the conflicting interests of various internal actors, each of which tends to undermine the stability of the rules and roles? Are there forms of weak homeostasis that work to restore a social structure in face of minor deviations? The "entropic" forces that should be expected to push organizations towards incessant change are fairly obvious; most evidently, individuals in strategic positions within an organization often have interests that are well served by adapting, reconceptualizing, or disregarding the rules. (Michel Crozier makes this point in his analysis of organizations as strategic sites; link.) So a large question for the theory of organizations and institutions is that of stability. What are the features of an organization in an environment that lends stability to its current constitution? Kathleen Thelen considers both aspects of this question in much of her work, including especially How Institutions Evolve: The Political Economy of Skills in Germany, Britain, the United States, and Japan and (with James Mahoney) Explaining Institutional Change: Ambiguity, Agency, and Power (link).


Sunday, February 15, 2015

Fissioning community


What is a community, and what forces exist to either preserve the strands of community or to erode those strands?

To start, a community is a population within a defined geographical space. It is a group of people who have their own interests, identities, and affinities. It is a group of people who act in a variety of ways — work, garden, vote in elections, raise children, join protest movements, support the government, steal cars, and attend public events. It is a group of people who are located within a space of social relationships — a set of social network graphs detailing their relations to family, friends, co-workers, fellow worshippers, fellow members of sports teams and civic associations. It is a group of people who have strong ties and weak ties to a range of other people and organizations. And it is a group of individuals who have ideas about how society should work, what is fair, who benefits from current social and political arrangements.

What makes a collection of people into a community rather than a jumble of unrelated individuals? There needs to be something in common, something that ties individuals together so they are to some degree willing to act in consideration of others in the group—a set of values, a set of shared purposes, a set of shared commitments to rules and laws of action in society. A strong community demonstrates a high degree of common respect for these values, rules, and laws. A weak community is the society that Hobbes feared—an agglomeration of self-interested actors competing for power and wealth, in which life is nasty, brutish, and short.

We might specify that sometimes a population embodies a plurality of sub-communities, each of which has a set of shared values and purposes—the Catholics and the Protestants, the Hindus and the Muslims, the bourgeois and the proles. Sub-groups may have very different cultural and religious beliefs, and they may also have very different positions in the social-property system and the political system. These sub-groups may have strong links among themselves but only weak links with individuals in the other sub-group. They may have strong adherence to a set of values and norms shared by their group but not across groups, and their norms may specify different behavior towards in-group and out-group individuals.

Crucially, a population with nested sub-communities may also possess an overarching unifying identity. That is, a mixed society may contain sub-groups that have their own norms and affinities, but that also share higher-level norms and commitments with other sub-groups. "We X’s live our lives differently from those Y’s; but we all accept the legitimacy of law and the freedom of worship." Those higher-level commitments might include items along these lines:
  • recognition of the legitimacy of government
  • commitment to the rule of law
  • commitment to rights of association, speech, religion, movement
  • commitment to the equal rights and worth of all members of the community
  • a compassionate regard for the lives and flourishing of members of other communities
It is very appealing in our multi-cultural world to conceive of a mixed community — a city or a region, say — as a harmonious and stable set of relationships among groups who differ in many important ways but who recognize the basic legitimacy and rights of the other groups, and who strive towards better inter-group understanding and cooperation. And we do in fact have good examples of such communities.

But all too often this ideal of a tolerant and stable community embracing multiple racial and ethnic groups is broken on the rocks of inter-group mistrust and violence. Groups polarize, individuals in one group develop mistrust or animosity against members of other groups, isolated instances of disrespect or violence escalate into conflict or pogroms. The civility that existed in one period breaks down into hostility and violence in another period.

It is important to understand the social mechanisms and strategies that exist in this arena in both directions: the strategies that exist for promoting mutual understanding and peaceful cooperation, and the mechanisms that exist that have the potential for undermining peaceful relations among groups in a population.

There are a number of familiar strategies and mechanisms that work in support of civility within a multi-ethnic or multi-religious community: organizations devoted to promoting inter-group understanding, institutions that encourage the formation of ties across group lines (like universities), leaders who promote the values of civility and mutual respect, effective law enforcement that quickly addresses violent actions by individuals in one group against those of another. The ways that these strategies work are also fairly plain. Equitable law enforcement enhances confidence that everyone will be treated fairly, no matter what group they identify with. Good leadership is exerted to make salient the unifying collective commitments that extend across all groups, and these commitments become motivationally effective. Cross-group organizations create new ties among individuals from different groups which gives them resilience in resisting the impulse of hostility when conflicts arise.

What are some of the mechanisms that come into play in multi-group populations that lead to instability and conflict? Here are some common factors that lead to the breakdown of community: mistrust, failures of assurance, political entrepreneurship, media incitement, rumor, and conflicts of interest over property and political arrangements. Some groups and leaders have a political interest in elevating tensions between groups. Rumors of bad actions by members of the other group lessen the resolve of group members to maintain their stance of solidarity. (Atul Kohli considers some of these mechanisms in India in Democracy and Discontent: India's Growing Crisis of Governability.)

It seems apparent that the mechanisms and impulses towards fission are perennial and need to be contained if a multi-ethnic or multi-religious community is to be sustained over time. Work by scholars like Donald Horowitz (Ethnic Groups in Conflict) and Ashu Varshney (Ethnic Conflict and Civic Life: Hindus and Muslims in India) illustrate the historical processes through which mixed communities have broken down into ethnic violence. It would be very interesting to attempt to use agent-based modeling to model the workings of pro- and anti-solidarity mechanisms in a mixed population, along the lines of a public health simulation of disease, to evaluate the conditions that favor or disfavor the survival of community in a population incorporating a number of groups.

Monday, February 9, 2015

Expectations and performance


There is a growing accumulation of evidence suggesting that individuals' performance in a wide range of activities is powerfully and insidiously affected by the expectations that are subtly conveyed to them by the people around them. If the people around a child or adult believe the individual will have particular difficulty with an upcoming task, this doubt is conveyed through cues the individual is able to sense. And, through cognitive mechanisms not fully understood, the subject's performance is affected, with results that are lower than otherwise possible.

The social psychologist Claude Steele was one of the first to study this phenomenon in the context of the SAT gap between white and black students. He introduced the concept of stereotype threat and documented its workings in a range of contexts (link; 1997). A crucial element of the research is the discovery that the negative cues do not need to be particularly overt or explicit; small elements of body language, verbal phrasing, or facial expression seem to be enough to significantly lower performance.
Negative stereotypes about women and African Americans bear on important academic abilities. Thus, for members of these groups who are identified with domains in which these stereotypes apply, the threat of these stereotypes can be sharply felt and, in several ways, hampers their achievement. First, if the threat is experienced in the midst of a domain performance -- classroom presentation or test-taking, for example -- the emotional reaction it causes could directly interfere with performance.... Second, when this threat becomes chronic in a situation, as for the woman who spends considerable time in a competitive, male-oriented math environment, it can pressure disidentification, a reconceptualization of the self and onf one's values so as to remove the domain as a self-identity, as a basis of self-evaluation. (614)
Key in Steele's work is the role that "self-identification" plays in learning and performance: the way that the child or young person thinks about his or her capabilities and talents is itself a key element in determining the level of performance that he or she will achieve.

Social psychologist Carol Dweck has contributed a great deal to this field through her research on the social and emotional influences on cognitive development. She and colleagues argue for a new hybrid field of research, "social cognitive development". Her paper with Kristina Olson, "A Blueprint for Social Cognitive Development" (2008), lays out the central features of this approach. Particularly important in this approach is the importance these researchers give to the role of the mental frameworks through which people navigate the world and solve the problems they encounter.
At its core, SCD is a field concerned with how mental representations and mental processes relevant to social development change across development. It also involves the study of how these mental representations and processes may mediate or moderate the impact of particular antecedents (e.g., parental input) on children's outcomes (e.g., well-being). (193)
Olson and Dweck highlight four key goals for research in social cognitive development studies:
  1. Identify and measure a social cognitive mental representation or process that is believed to be important in development.
  2. Manipulate the mental representation, and observe its impact on outcomes of interest over development.
  3. Investigate the antecedents of the mental representation or process of interest.
  4. Compare how the mental representation operates in the laboratory and in the real world. (195-196)
This is a crucial theoretical innovation, because it insists on what we have previously called an "actor-centered" approach to behavior. What is the mental framework on the basis of which the individual acts? Olson and Dweck consider examples of aggression and achievement motivation and demonstrate how beliefs and mental frameworks influence the child's dispositions to behavior in various circumstances.

Another important body of research on this topic comes from the field of career and educational choices. An earlier post highlighted career choice -- the idea that certain professions or careers are "typed" for particular profiles of gender, race, and political beliefs (link). This typing is conveyed through broad cultural motifs, and it gets inside the heads of young people at a level that is almost below the level of conscious awareness. Elizabeth Armstrong and Laura Hamilton's work aimed at decoding of the educational choices made by women of different socioeconomic segments illustrates a related phenomenon (link). Armstrong and Hamilton demonstrate that women at MU are tracked into different pathways of education and career opportunity, depending on the socioeconomic status of their families.

A third relevant area of research on this topic comes from a different field, the study of the social components of blindness. Robert Scott, a sociologist who focused on the institutions supporting the needs of blind people in The Making of Blind Men, argued that a large part of the mobility limitations experienced by visually impaired people derives from the script of disability that sighted society insists upon. A fascinating case study is Daniel Kish, the blind bicyclist. (This story was highlighted in a recent episode of the podcast Invisibilia; link.) The key finding is this: if most of us, including the parents and teachers of sight-impaired children, believe that blindness means dependency and highly limited mobility, then this will be the reality for blind children and adults. But if the assumption is that sight-impaired people can develop alternative means of navigating their environments ( like the tongue-clicking system that Kish developed as a child) then a very different form of life will result for the sight-impaired.



In each of these areas of research, it is striking to find how sensitive capabilities and performance are to social expectations. Low or negative expectations cause poor performance. These findings underline how sensitive human beings are to the expectations other people have of us when it comes to development and performance. Second, these findings provide a basis for understanding some important differences in performance that exist along gender, racial, and ethnic lines. And they pose very tough challenges to teachers of young children; how can preschool and early childhood teachers assure themselves that their own hidden stereotypes about the children they teach do not wind up hampering the child's development? In a racialized and genderized society that is laden with stereotypes, how can we ensure that learning environments will bring out the best in every child?

But universities too need to worry about this phenomenon. Do engineering and math faculty give off subtle cues about their expectations of low performance by young women? Do programs aimed at supporting the academic success of educationally disadvantaged students -- often a racialized group -- give rise to diffuse assumptions among faculty that these students have lower potential for excellence? Does a visibly blue-collar student receive subtle cues that graduate studies are not a good choice for him or her? And, given the emerging research showing how stereotype and expectation dramatically influence performance and choice, how can we work to offset these mechanisms in order to provide a genuinely equal quality of education and career advice to all students?

Tuesday, February 3, 2015

Self-selection and "liberal" professions

image: Professor Charles Hill provides a conservative voice on Yale’s faculty

Neil Gross stirred up a quite a storm a few years ago when he released a body of research findings on the political complexion of university professors. Conservative organizations and pundits have made hay by denouncing the supposed liberal bias of universities. Gross opens his most recent book, Why Are Professors Liberal and Why Do Conservatives Care?, by confirming that multiple measures demonstrate that faculty members are substantially more likely to be liberal than the general population. But he believes this is both indisputable and uninteresting. What is most interesting for a sociologist is the "why" -- how can we explain the skewed distribution of political identities across this professional group?

Gross positions his work as falling in a tradition of research that included two major survey-based studies in the past 75 years, Lazarsfield and Thielens (The Academic Mind: Social Scientists in a Time of Crisis) and Everett Karll Ladd and Seymour Martin Lipset (The Divided Academy: Professors and Politics). He also finds creative ways of incorporating the GSS surveys. In addition, Gross and Solon Simmons carried out their own substantial survey, the Politics of the American Professoriate survey (PAP).

It will be noted that this is a problem that calls out for a social mechanisms explanation, and a fairly simple one at that. Suppose marbles of two colors in equal numbers are raining down in a thoroughly mixed stream on a pair of urns. Occasionally a marble falls into one urn or the other. When we count the marbles in both urns we find that 65% of the marbles in the urn on the left are green, whereas the larger urn on the right has 50% of each color. How is it that there are a higher percentage of green marbles in the urn on the left? There are only a few possibilities, each corresponding to a different mechanism. (i) The marbles have a degree of choice about which urn they enter, and green marbles have a preference for the left urn. (Or a variant: red marbles have a preference for avoiding the left urn.) We could call this "selection bias by chooser". This is different from two other possible mechanisms: (ii) there is a filter on the left urn, bumping green marbles in and red marbles out ("selection bias by receiver"); or (iii) marbles have a slight tendency to shift color from red to green when they enter the left urn ("environment transformation").

Fundamentally the question is analogous to the "nature-nurture" conundrum in the study of personality. Are universities more liberal than average occupations because they cultivate liberal thinking (nurture)? Or are they more liberal because of some sort of selection mechanism, drawing liberal members more frequently than chance (nature), and liberal-tending new faculty members bring their political values with them? Gross makes a powerful empirical case for the latter possibility.
I develop an alternative account: for historical reasons the professoriate has developed such a strong reputation for liberalism that smart young liberals today are apt to think of academic work as something that might be appropriate and suitable for them to pursue, whereas smart young conservatives see academe as foreign territory and embark on other career paths. (p. 105)
It might be speculated that this distribution exists because the faculty selection process is biased -- conservative candidates are turned away. Gross's explanation is different. He draws on a strong literature studying gendered occupations that finds that the reputation of the profession has a powerful influence on girls and women as they make educational and career choices. Gross extends this reasoning to liberals and conservatives contemplating an academic career. Essentially he explains the political composition of university faculties as a consequence of a powerful public reputation for being a liberal workplace and a distinctly skewed process of self-selection towards this career. The profession is "typed" as being a particularly good career for more liberal young people, and young people make their career choices in consideration of that assumption. Essentially he argues that universities are publicly perceived as being hospitable to people on the left, and liberal-leaning young people are drawn to the career because of this reputation.

The question of political bias within universities is treated using an interesting experiment that Gross and colleagues Joseph Ma and Ethan Fosse conducted to test whether conservative students have a harder time gaining entrance to graduate programs (164). The project involved sending fictitious letters of inquiry to directors of graduate studies in leading departments in a wide range of disciplines. The letters indicated the same level of preparation for the field. One batch indicated no political information about the student, while the other two batches included the phrases "Worked for the McCain campaign" or "Worked for the Kerry campaign." Responses were rated according to the degree of encouragement or discouragement they expressed. The experiment is ingenious but it indicates "no result". There is no statistically significant evidence of bias against applicants who self-identify as conservative. (Gross does report a strong negative response from some of the academics whose potentially discriminatory behavior was tested.)
The study should count as reasonably strong evidence that most social scientists and humanists in leading departments work hard to keep their political feelings and opinions from interfering with their evaluations of academic personnel. (165)
I find Gross's treatment of this topic to be an exemplary use of quantitative survey data in theoretically informed ways. The PAP survey that Gross initiated (along with colleagues and research assistants) provides substantial new information about the political attitudes and social backgrounds of faculty in the United States. Gross makes deft use of this data source (as well as several others) to evaluate hypotheses about what causes the distribution of political profiles among faculty. Gross's question is about both groups and individuals, and the survey data helps to evaluate answers to both. And, incidentally, this appears to be a sterling example of the kind of theoretical work that John Levi Martin calls for (link): careful stipulation of various explanatory theories, accompanied by a rigorous effort to evaluate them using appropriate empirical data.

For anyone who cares about universities as places of learning for undergraduate students, Gross's book is an encouraging one. He provides a clear and convincing explanation of the mechanisms through which a non-random distribution of political attitudes wind up in the population of university and college professors, and he provides strong evidence against the idea that universities and professors exercise discriminatory bias against newcomers who have different political identities. And finally, Gross's analysis and my own experience suggests that professors generally conform to Weber's ethic when it comes to proselytizing for one's own convictions in the classroom: the function and duty of the professor is to help students think for themselves (Max Weber, "The Meaning of Ethical Neutrality,", Methodology of Social Sciences.)

There is an interesting set of replies to Gross in Society here.