Legal Construct Validation: Expanding Empirical Legal Scholarship to Unobservable Concepts

• Author: David S. Goldman
• Pages: 79-127

LEGAL CONSTRUCT VALIDATION:

EXPANDING EMPIRICAL LEGAL SCHOLARSHIP

TO UNOBSERVABLE CONCEPTS

DAVID S. GOLDMAN*

INTRODUCTION

Unobservable concepts such as rights, possessions, incentives, and

deterrence form the foundation of legal thinking. Yet over 200 years ago,

Jeremy Bentham strenuously warned us not to be seduced by the follies of

these types of terms, which he called inefficient legal fictions:1 “What you

have been doing by the fiction⎯could you, or could you not, have done it

without the fiction? If not, your fiction is a wicked lie: if yes, a foolish one.”2

For Bentham, legal fictions were artifacts of the imprecision inherent in

language.3 Such imprecision prevented meaningful distinctions between real

entities, such as “this man, this beast, this bird,” and abstract entities.4 He

claimed that such fictions, rather than expressing lofty ideas, were instead

most often used to subvert the law.5

Why is it then that the use of such terms continued unabated? At the very

least, they constitute shorthand for a list of real entities that would be too

extensive to name in each instance of use; for example, referring to “civil

rights” is simpler than enumerating each constitutional right to personal

liberty.6 But a more critical role can be assumed for these terms that Bentham

could not have yet appreciated. Given techniques that have been developed in

the social sciences in the last 60 years, these “fictions” can be operationalized,

* Copyright © 2007, David Goldman, goldmand@gmail.com. I would like to thank Greg

Smith, Gideon Parchomovsky, Michael O’Donnell, and Spencer Waller for their extremely

helpful comments. The views outlined in this article are entirely my own and do not

necessarily reflect those of any past or present employers.

1 C.K. OGDEN, BENTHAM’S THEORY OF FICTIONS 38 (1932).

2 Id. at 141.

3 Id. at 12.

4 Id. at 38, 60.

5 Id. at 141, 147–50 (listing the ways lawyers use these “forced falsehoods” to make more

business for themselves, such as increasing chance of misstatements, making it easier to allege

that jury verdicts are improper by making the law incomprehensible, confirming arbitrary

power to judges, and corrupting the morals and intellectual faculties of the public).

6 “Civil rights” include the personal liberties guaranteed by the Bill of Rights, the

Thirteenth, Fourteenth, Fifteenth, and Nineteenth Amendments, as well as the Voting Rights

Act, among other sources. BLACK’S LAW DICTIONARY 100 (Pocket ed. 1996).

80 CAPITAL UNIVERSITY LAW REVIEW [36:79

and can actually then enter into empirical relationships that permit their

measurement, as well as the assessment and prediction of the effects of laws

built around these “fictions” on subsequent human actions.7 For instance, in

Ronald Coase’s now fundamental example of the negotiation between a

farmer and a cattle-raiser for use of land, the neighbors were able to broker a

deal because they owned property rights in their land.8 While the rights and

the property may both be unobservable legal fictions, one can easily observe

whether crops or cattle occupy a particular parcel of land. By gaining a

detailed understanding of the operations of property rights, policymakers can

predict how changes in property law will affect future interactions for

efficient use of land. Legal scholars can now, finally, overcome Bentham’s

criticisms and determine empirically the practical usefulness of some of our

most reliable legal fictions.

Empirical observations of legal entities has been gaining particular

popularity recently because it allows debate on fundamental questions that are

informed by the actual impact of law on behavior rather than conjecture and

an appeal to commonsense.9 But now legal scholars can expand the current

horizons of quantitative analysis in legal scholarship by moving beyond

measurement of only observable entities to assessing unobservable legal

models as well through the notion of construct validation—the methods

behavioral scientists use to overcome the difficulty in measuring unobservable

psychological phenomenon, called constructs. Using the metatheories of

construct validation, researchers can infer the effects of unobserved constructs

and can then modify foundational legal theories based on empirical evidence

rather than speculation.10 I propose a standardized procedure, adapted from

7 See Brendan A. Maher & Irving I. Gottesman, Deconstructing, Reconstructing,

Preserving Paul E. Meehl’s Legacy of Construct Validity, 17 PSYCHOL. ASSESSMENT 415, 415

(2005) (describing the evolution of measurement of unobservable entities).

8 Ronald H. Coase, The Problem of Social Cost, 3 J.L. & ECON. 1 (1960), reprinted in THE

FIRM, THE MARKET, AND THE LAW 95, 97 (1988).

9 See, e.g., Tracey E. George, An Empirical Study of Empirical Legal Scholarship: The Top

Law Schools, 81 IND. L.J. 141, 141 (2006) (“Empirical legal scholarship (ELS) is arguably the

next big thing in legal intellectual thought.”); American Association of Law Schools Annual

Meeting: Empirical Sc holarship—What Should We Study and How Should We Study It? (Jan.

3–7, 2006), http://www.aals.org/am2006/index.html; Lee Epstein & Gary King, The Rules of

Inference, 69 U. CHI. L. REV. 1, 2 (2002); Richard H. McAdams & Thomas S. Ulen,

Introduction: Symposium: Empirical and Experimental Methods in Law, 2002 U. ILL. L. REV.

791.

10 Thomas S. Ulen, A Nobel Prize in Legal Science: Theory, Empirical Work, and the

Scientific Method in the Study of Law, 2002 U. ILL. L. REV. 875, 899.

2007] LEGAL CONSTRUCT VALIDATION 81

current understandings of construct validation, to measure intangible legal

constructs: (1) develop generalized legal theories, (2) infer hypotheses from

those theories, (3) design experiments to test the hypotheses, and (4) modify

the general theories based on the research results.

This first step in this process is the development of theories with a

generalized perspective not confined to specific laws that still maintain a

degree of specificity that provides adequate direction for future empirical

testing. Generalized theories allow multiple inferences that can elucidate

finer detail of an unobserved latent variable. For instance, rather than

addressing only the impact of the First Amendment,11 a more generalized

theory will address the influences of all constitutional rights.12 A generalized

theory may state that populations will tend to exhibit more freedom when

specific rights are embedded within those populations’ constitution.

Empirical studies of freedom of speech can then provide particularized

support for the broader constitutional theory, and can also interact with

studies of other rights.

Once a generalized theory has been devised, researchers can then infer

sets of hypotheses to test measurable aspects of those theories. For example,

based on a theory that constitutional rights lead to more freedom, one might

hypothesize that if a nation has a constitutional freedom of speech then the

media may produce more commentary disapproving of the government.

While the actual freedom that the press feels is impossible to measure, the

proportion of articles disapproving of government actions is relatively easy to

calculate. But a useful hypothesis suggests more than just an empirical test; it

should also have the potential to undermine the theory.13 The necessity for

critical hypotheses is borne from the current philosophical understanding that

11 See, e.g., Stephen M. Feldman, The Theory and Politics of First Amendment Protections:

Why Does the Supreme Court Favor Free Expression over Religious Freedom?, 8 U. PA. J.

CONST. REV. 431, 433–51 (2006) (discussing the development of First Amendment theory

since the drafting of the Constitution). While the development of theories for specific laws and

Constitutional provisions is extremely beneficial in its own right, these narrowly-tailored

theories do not lay an appropriate groundwork for empirical study because they do not provide

a broad enough base from which to infer traits of the unobservable variable.

12 See, e.g., Stephen Kanter, The Griswold Diagrams: Toward a Unified Theory of

Constitutional Rights, 28 CARDOZO L. REV. 623 (2006) (attempting to analyze the abortion

debate by using a more generalized theory for Constitutional law when the basis of the right is

unclear, rather than focusing on the specific rights outlined in the Supreme Court’s abortion

cases).

13 Gregory T. Smith, On Construct Validity: Issues of Method and Measurement, 17

PSYCHOL. ASSESSMENT 396, 396–97 (2005).