How do sociologists use theory? And what do you think about Lieberson’s ideas about the use of evidence in testing sociological theory?

P1: FRK

June 10, 2002 12:5 Annual Reviews AR163-FM

A nn

u. R

ev . S

oc io

l. 20

02 .2

8: 1-

19 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

26 01

:4 c1

:4 00

0: 4f

c0 :2

94 9:

e6 2a

:3 bf

1: 98

55 o

n 03

/1 9/

23 . F

or p

er so

na l u

se o

nl y.

10 Jun 2002 20:0 AR AR163-01.tex AR163-01.SGM LaTeX2e(2002/01/18)P1: IKH 10.1146/annurev.soc.28.110601.141122

Annu. Rev. Sociol. 2002. 28:1–19 doi: 10.1146/annurev.soc.28.110601.141122

Copyright c© 2002 by Annual Reviews. All rights reserved

BARKING UP THE WRONG BRANCH: Scientific Alternatives to the Current Model of Sociological Science

Stanley Lieberson and Freda B. Lynn Department of Sociology, Harvard University, Cambridge, Massachusetts 02138; e-mail: [email protected]

Key Words experimentation, evidence, evolution, Darwin, physics

A fortiori it is unlikely that a mere repetition of the tricks which served us so well in physics will do for the social phenomena too.

(John von Neumann and Oskar Morgenstern 1944, p. 6)

I do not think the relationship between theory and measurement in the social sciences is much like what Kuhn describes for physics. Talcott Parsons was right about the lack of interaction between the two in sociology. If Kuhn is right about the preconditions for such interaction in physics, and if physics is the model for sociology, then it will be a long time before measurement makes an important contribution to sociology as a basic science. . . .

But sociology is not like physics. Nothing but physics is like physics. . . . (Otis Dudley Duncan 1984, p. 169)

■ Abstract The standard for what passes as scientific sociology is derived from classical physics, a model of natural science that is totally inappropriate for sociology. As a consequence, we pursue goals and use criteria for success that are harmful and counterproductive. Even those dismissing such efforts use the standards of physics as grounds for their objection. Although recognizing that no natural science can serve as an automatic template for our work, we suggest that Darwin’s work on evolution provides a far more applicable model for linking theory and research since he dealt with obstacles far more similar to our own. This includes drawing rigorous conclusions based on observational data rather than true experiments; an ability to absorb enormous amounts of diverse data into a relatively simple system that did not include a large number of what we think of as independent variables; the absence of prediction as a standard for evaluating the adequacy of a theory; and the ability to use a theory that is incomplete in both the evidence that supports it and in its development. Other sciences are briefly cited as well, but the main emphasis is on the lessons that Darwin provides for social sciences such as sociology that obtain their evidence primarily from non-experimental sources.

0360-0572/02/0811-0001$14.00 1

A nn

u. R

ev . S

oc io

l. 20

02 .2

8: 1-

19 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

26 01

:4 c1

:4 00

0: 4f

c0 :2

94 9:

e6 2a

:3 bf

1: 98

55 o

n 03

/1 9/

23 . F

or p

er so

na l u

se o

nl y.

10 Jun 2002 20:0 AR AR163-01.tex AR163-01.SGM LaTeX2e(2002/01/18)P1: IKH

2 LIEBERSON ¥ LYNN

INTRODUCTION

Ever since the establishment of sociology, various polemical and philosophical issues have surrounded the attempt to model a science of society after the hard sciences. Since the natural sciences differ in theoretical concerns, nature of evi- dence, and the special obstacles encountered in combining evidence with theory, these sciences also vary in their potential for generating ideas that could be useful for sociology. (Of course, the natural sciences at best are simply sources of ideas that might be helpful to us; under no circumstances are they templates that can be automatically adopted for usage in social science.) Our thesis is that deeply ingrained in sociology and other social sciences is a special model of natural sci- ence that is exceptionally inappropriate. It is derived from physics, particularly the classical physics that existed before the beginning of the twentieth century. Because of the impressive results obtained in physics and also its precision, it is understandable why this science would be favored as a model for the social sci- ences to follow. The use of physics as the ideal model for sociology is so embedded in our thinking that the influence and appropriateness of this particular model is rarely questioned. There are those, of course, who do criticize attempts to create a socialsciencemodeled after the natural sciences as pretentious, bound to fail, and intellectually sterile. Ironically, they too use physics as the model of hard science in order to demonstrate the error of such an attempt (see, for example, Flyvbjerg 2001). The appeal of classical physics is easy to understand; it offers very straight- forward simple cause-effect connections that are very powerful and operate in an incredible array of situations. The actual path of a cannonball shot into the air, for example, is very close to the theoretical prediction that takes into account the angle of the cannon, the thrust of the explosion, the weight of the ball, and the altitude from which it is fired. For human societies, we have great difficulties making such simple cause-effect connections with comparable precision.

Our thesis is that other natural sciences actually offer epistemological and pro- cedural models that are more relevant for the obstacles encountered in sociology and other social sciences: Their ways of interacting between data and theory are more suitable, and their notions about theory, causality, and explanation are also more pertinent for us. In particular, we focus on the early development of evo- lutionary theory. This highly successful natural science has much to teach us. In many dimensions it encounters obstacles similar to those in sociology, but evo- lution employs solutions very different from the ones we use. At the very least, we need to consider these approaches as an alternative to present-day practices. In doing so, we can actually revise our notions of the difficulties that occur when we pursue a very mechanical notion of science. Evolution and other natural sciences help us realize that what appears to be problematic and disappointing (when so- ciology is approached with physics as the model of natural science) is simply a consequence of utilizing an inappropriate model for studying social phenomena. That is, unnecessary difficulties or obstacles are created because an incorrect stan- dard is in use. The absence of predictive power and high correlations, for example, are not a weakness of our theory (as Rein & Winship 1999 would have you believe)

A nn

u. R

ev . S

oc io

l. 20

02 .2

8: 1-

19 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

26 01

:4 c1

:4 00

0: 4f

c0 :2

94 9:

e6 2a

:3 bf

1: 98

55 o

n 03

/1 9/

23 . F

or p

er so

na l u

se o

nl y.

10 Jun 2002 20:0 AR AR163-01.tex AR163-01.SGM LaTeX2e(2002/01/18)P1: IKH

SOCIOLOGICAL SCIENCE MODELS 3

but should be seen as an inherent reflection of the way the social order operates and can best be understood.

Finally, the reader should note that we are not interested in applying biological models to social life (as in the notions of social evolution or social Darwinism) or promoting them. Nor are we concerned with biological questions about the hard- wiring of human behavior and the social order. Rather, our interest in evolution is strictly an epistemological one, looking at the very different ways theory and evidence are formulated and linked together in a natural science that confronts epistemological obstacles similar to those occurring in the social sciences.

DARWIN: A BRIEF REVIEW

The early developments of evolutionary theory provide an epistemological model of how a discipline, although facing many of the same obstacles commonly en- countered in social science, is able successfully to combine sound evidence with a viable and useful theory. Arguably the early evolutionists encountered more obstacles similar to sociology than virtually any other hard science. The form of theorizing in evolution, the form of gathering evidence, and the form of interacting between theory and evidence have considerable relevance to issues encountered in sociology.1

Jones (2000, pp. xviii–xix) helps us understand the magnitude and importance of evolutionary theory: “Before Darwin, the great majority of naturalists believed that species were immutable productions, and had been separately created. Today, his theory that they undergo modification and are the descendants of pre-existing forms is accepted by everyone (or by everyone not determined to disbelieve it). . . . The struggle for existence, the survival of the fittest and the origin of species are wisdom of the most conventional kind. . . . Although the notion is as simple as that of the solar system, Darwinism is not the obvious explanation of how the world works. Common sense tells us that life—like the Sun—revolves around ourselves. The idea has but one fault: it is wrong. . . . The Origin of Speciesis, without doubt, the book of the millennium.”

The evolutionary model can be characterized as a two-stages process: the pro- duction of variation followed by the sorting of this variability by natural selection, the critical mechanism or motor in evolution. In more detail, the stepwise process is as follows: First, among individual organisms, there is variation, both natural (i.e., random) and domesticated, some of which is heritable. Second, organisms in nature generally produce offspring in excess of the number that can reach the reproductive age because of competition. Assuming a stable population, a “strug- gle for existence” ensues among the offspring for survival. Organisms with certain variations or characteristics are favored during this struggle. Those who are favored or naturally selected will thus have a significantly better opportunity to reproduce

1The substantive ideas also influenced social scientists such as Marx, Engels, Comte, Morgan, Spencer, and Weber (Antonio, 2000, p. 1782; Richter, 2000, p. 877; Cohen, 1994).

A nn

u. R

ev . S

oc io

l. 20

02 .2

8: 1-

19 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

26 01

:4 c1

:4 00

0: 4f

c0 :2

94 9:

e6 2a

:3 bf

1: 98

55 o

n 03

/1 9/

23 . F

or p

er so

na l u

se o

nl y.

10 Jun 2002 20:0 AR AR163-01.tex AR163-01.SGM LaTeX2e(2002/01/18)P1: IKH

4 LIEBERSON ¥ LYNN

offspring and thus later descendants who possess some of their traits and variations. Over time, beneficial (i.e., well-adapted) characteristics will be accumulated in the population through the process of natural selection. Finally, over great lengths of time (i.e., geological time), new species may evolve in a manner congruent with this process of variation and natural selection (the discussion in this paragraph and the following one draws in part from: Jones 2000, Gould 1977, 1982, Bonner & May 1981, Mayr 1967, Simpson 1950, Lasker & Tyzzer 1982).

Evolution employs a set of powerful mechanisms to help one understand finer and finer parts of the story. In doing so, it helps the theory deal with the variety of specific types of causes that operate in various circumstances but are not always factors. Genetic drift or migration, for example, do not always occur, yet they are critical in some instances. These mechanisms serve to deal more successfully with a complex system in which there are often different crucial conditions present. In sociology, some of us try to control for all of these through statistical means; others try to treat each as an historical event that shares a common major condition. Here we have a solution that is far more promising. A variety of mechanisms may be called into play. They have something in common: They all fit into the overriding theory and support its goals; in this case they are part of the theory of evolution. One is not bothered if not all of the mechanisms are relevant to any specific case. However, the mechanisms are stated in generalized form as principles or theories. Evolution operates with neither unique historical events reflecting unique historical causesnora crass mechanical approach that evaluates a mechanism by how often it operates or how much of the variance it accounts for. A mechanism is still pertinent regardless of how often it operates to generate changes in the survival of a species or a species’ growth or decline in numbers. The mechanism is not evaluated in terms of the variance it explains. This bears greater consideration in our own enterprise (see Richter 2000, p. 876 for a distinction between individual history and the unique history of evolution). Sequential analysis is central to evolution. Namely a process of events is traced, as say leading from an environmental shift to a genetic shift as the living entity adapts to the new condition. A massive shift such that a new species appears would involve many steps along the way. Tracing these changes would provide evidence to support the claim.

LESSONS FOR US

Neither Right Nor Wrong

In reviewing developments from Darwin to more recent periods, we are struck by how the theory has a tolerance for problems and incompleteness that gives it a certain durability and that enables one to better cope with errors. The evolution- ists do not confuse fatal errors, on the one hand, with problems stemming from incompleteness, information that is still insufficient or not yet determined, or even unresolved. The latter cases are worrisome and certainly not to be glossed over. Yet, it does not necessarily mean that the theory is to be abandoned or that Darwin waswrong. In evolution,incompleteis not the same aserroneous.

A nn

u. R

ev . S

oc io

l. 20

02 .2

8: 1-

19 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

26 01

:4 c1

:4 00

0: 4f

c0 :2

94 9:

e6 2a

:3 bf

1: 98

55 o

n 03

/1 9/

23 . F

or p

er so

na l u

se o

nl y.

10 Jun 2002 20:0 AR AR163-01.tex AR163-01.SGM LaTeX2e(2002/01/18)P1: IKH

SOCIOLOGICAL SCIENCE MODELS 5

The excerpt below from Darwin’sOn the Origin of the Speciesis valuable to us because it is hard to imagine a sociological theorist being able to admit to such difficulties in a newly proposed theory. Read this not for the content of the theory or the missing information, but as an example of a way of dealing with evidence that rarely occurs in sociology:

There is another. . . difficulty, which is much more serious. I allude to the manner in which species belonging to several of the main divisions of the animal kingdom suddenly appear in the lowest known [Cambrian-age] fossil- iferous rocks. . . . If the theory be true, it is indisputable that before the lowest Cambrian stratum was deposited, long periods elapsed. . . and that during these vast periods, the world swarmed with living creatures. . . . [But] to the question why we do not find rich fossiliferous deposits belonging to these as- sumed earliest periods before the Cambrian system, I can give no satisfactory answer. The case at present must remain inexplicable; and may be truly urged as a valid argument against the views here entertained.

(Schopf 2000, p. 7, quoted from Darwin 1859, Ch. X).

To be sure, through the century there were debates about Darwin’s conjectures, but the theory survived uncertainty and incompleteness during that long span. It was only one hundred years later that fossils were uncovered that fully supported Darwin’s conjectures (Schopf 2000, p. 18). One might argue that this is different from the incompleteness and uncertainty existing in sociology because in Dar- win’s case there was reason to hope that eventually data (i.e., fossils) would be uncovered to provide missing information about the past that could furnish in- formation about the period in a definitive way. However, there is every reason to assume that additional information and understanding can be obtained about vir- tually any topic that sociology concerns itself with as well. It just takes work and patience.

Like sociology, evolution deals with complex evidence. Both are exposed to two enormous problems: (a) Given the lesser certainty that we have in working with the data that are available to us, we must use different standards from the conventional ones, and (b) in order to be sure that what we say is valid, we cannot rely simply on variance explained by models of evidence or act as if we are dealing with a sample from the universe. This is also a problem for qualitative small-N studies where a “full explanation” is the standard for knowing that the investigator is on track. On the other hand, if we are not using simple rules for accepting and rejecting, how do we avoid taking bad ideas and using them to expand onward and onward, further and further from the true path. In other words, how do you set standards that enable one to separate the wheat from the chaff? We have no immediate answer here, but a conviction that part will be intuitive, as it always is, and part will require us to develop clearer thinking about when the evidence truly is consistent with the theory, when it is irrefutably inconsistent, and when we have to say that “the evidence is not all in.” We observed earlier how Darwin recognized huge holes in what he knew and what he could account for. How would we evaluate a theory with such gaps? Under present-day standards, how often could/would a

A nn

u. R

ev . S

oc io

l. 20

02 .2

8: 1-

19 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

26 01

:4 c1

:4 00

0: 4f

c0 :2

94 9:

e6 2a

:3 bf

1: 98

55 o

n 03

/1 9/

23 . F

or p

er so

na l u

se o

nl y.

10 Jun 2002 20:0 AR AR163-01.tex AR163-01.SGM LaTeX2e(2002/01/18)P1: IKH

6 LIEBERSON ¥ LYNN

theorist admit lack of full knowledge or understanding? How do we implement an epistemology that permits softer and yet harder standards?

A Successful Theory

Note also that Darwin’s theory is a successful theory, but it is not a static theory. While there is every reason to read Darwin for historical reasons or to gain an admiration of his thinking qualities, his theory of evolution is not the same as the theory that is taught now. It is not necessary to read Darwin in order to work on evolution since the discipline has gone long past him. This is of interest because in social science we talk about theories being either right or wrong. It is not that simple, and moreover, it is a gross distortion of how a theory should change. Keep in mind the view of a theory as active process involving an interaction with known facts and then in turn providing generalizations that go beyond these known facts. So, as we know more through new observations and also as we reach new problems based on what we believe we already know, it is inevitable that the theory should change or at least be modified. It is a matter not of simply rejecting a theory, but rather of evaluating a theory, knowing at some point that it will have to be modified or even superseded. Put another way, if Darwin waswrong, then we should all be lucky enough to be as wrong as he was. Since Darwin was writing at a time when nothing was known about mutations and genetics, how could he possibly have a complete theory? In social science, we are more likely to want to destroy a theory or, at the other extreme, worship a theory and therefore resist its change or modification. This is very different from asking “What can I get from this idea?”, “Where is this helpful in dealing with observed events?”, “How can I modify the idea such that I keep the parts that are useful and yet expand its application?” Of course, one has to know when to give up and when something is genuinely useful and merits modification and further exploration.2 In any case, as one goes on in the development of a theory of evolution, more and more detailed principles and mechanisms develop to explain finer and finer parts of the theory.

Way of Working

The formal way of writing up one’s work is misleading because it is typically presented, particularly in articles, as an orderly product starting with a problem, usually derived from a theory and/or a social problem, and ending with conclu- sions based on the ensuing investigation. This is unfortunate. The formal way is indeed practiced in the sciences, but it is hardly the only pathway. Often scientific pursuits and sociological pursuits are much more intuitive, scattered, nonlinear, informal, and trial-and-error than the published literature suggests. Darwin is an interesting model on that score; his notebooks show that his process was for the most part nonlinear with unordered sequences of “theorizing, experimenting,

2Of course, always relevant is the internal logic of the theory, its application to specific empirical evidence, and matters of that nature.

A nn

u. R

ev . S

oc io

l. 20

02 .2

8: 1-

19 . D

ow nl

oa de

d fr

om w

w w

.a nn

ua lr

ev ie

w s.

or g

A cc

es s

pr ov

id ed

b y

26 01

:4 c1

:4 00

0: 4f

c0 :2

94 9:

e6 2a

:3 bf

1: 98

55 o

n 03

/1 9/

23 . F

or p

er so

na l u

se o

nl y.

10 Jun 2002 20:0 AR AR163-01.tex AR163-01.SGM LaTeX2e(2002/01/18)P1: IKH

SOCIOLOGICAL SCIENCE MODELS 7

casual observing, cagey questions, reading, etc [that] would never have passed muster in a methodological court of inquiry” (cited from Gruber in Bonner & May, 1981, p. xiii). Darwin himself characterizes the pursuit as the struggle to- ward “seeing the bearing of scattered facts.” We fail to recognize a distinction between how many natural scientists and social scientists actually do operate as opposed to their formal style of presenting outcomes. We do ourselves a disser- vice when teaching methodology to our students as if the highly stylized form of presentation in journals and books is the same as the intellectual processes that lead to these conclusions. This actually encourages a more mechanical model of how knowledge progresses. This, along with the tendency to base our conclusion on only one data set, actually fits very well with a mechanical model of what our theories should be like. However, as was the case in evolution, it does not help us master the incredibly complex universe in which the social sciences operate.

JIGSAW PUZZLES, DETECTIVE STORIES, AND WHITE LAB COATS

Like sociology, evolution deals with a huge number of conditions that can influence the topic of interest. And, like sociology, much of the basic knowledge, at least certainly what Darwin had to work with, was drawn from observing naturally occurring events, that is, not from true experiments. So we have two issues here: first, how does the study of evolution manage to survive so well in dealing with the same type of complex and diverse sets of causal conditions that appear in a very large set of combinations and permutations, whereas sociology has trouble managing such data other than through very complex statistical models or through small studies that depend on deterministic and mechanical outcomes driven by a small number of conditions? Second, how can the students of evolution get along learning so much more from observational data than we do, even though sociology is saturated with procedures designed to estimate true experimental results from the nonrandom observational data that we usually have?

We have several speculations about this. One of the great accomplishments of Darwin was that he narrowed the problem down to a specific question. Certainly, he worked on a problem that was huge by any and all standards, the origin and changes in species, but it was still a restricted problem. It did not encompass all possible questions that you could ask about evolutionary biology. Indeed, one of Darwin’s great leaps was that he never attempted to account for how variations arise, how variations are maintained, and how variations are inherited. His solu- tion was simple: He took these as given and did not try explain these facts (it was through genetics that they were later resolved). Sociology does not seem to be as focused on a central big question in society. As a consequence, it is less likely to generate cumulative knowledge. But probably the biggest difficulty is that sociol- ogy attempts to simulate true experiments, hence the use of a variety of statistical techniques designed to

Collepals.com Plagiarism Free Papers

Are you looking for custom essay writing service or even dissertation writing services? Just request for our write my paper service, and we'll match you with the best essay writer in your subject! With an exceptional team of professional academic experts in a wide range of subjects, we can guarantee you an unrivaled quality of custom-written papers.

Get ZERO PLAGIARISM, HUMAN WRITTEN ESSAYS

Why Hire Collepals.com writers to do your paper?

Quality- We are experienced and have access to ample research materials.

We write plagiarism Free Content

Confidential- We never share or sell your personal information to third parties.

Support-Chat with us today! We are always waiting to answer all your questions.