The Trouble with Twin Studies

A Reassessment of Twin Research in the Social and Behavioral Sciences

Jay Joseph

• Publisher: Routledge
• Year: 2014
• Topic: Psychology
• eBook ISBN:
  9781317605904
• Subtopic: History & Theory in Psychology
• Index: Psychology

The Trouble with Twin Studies questions popular genetic explanations of human behavioral differences based upon the existing body of twin research. Psychologist Jay Joseph outlines the fallacies of twin studies in the context of the ongoing decades-long failure to discover genes for human behavioral differences, including IQ, personality, and the major psychiatric disorders. This volume critically examines twin research, with a special emphasis on reared-apart twin studies, and incorporates new and updated perspectives, analyses, arguments, and evidence.

1 Introduction

DOI: 10.4324/9781315748382-1

Because of their utility as a tool or method, twins are often referred to informally as the “Rosetta Stone” of behavior genetics.

— Twin researcher Thomas Bouchard, 1999 (Bouchard, 1999, p. ix)

Twin studies . . . provide the bulk of the evidence for the widespread influence of genetics on behavioral traits.

— Behavioral genetic researchers Robert Plomin, John DeFries, Valerie Knopik, & Jenae Neiderhiser, 2013 (Plomin, DeFries, Knopik, & Neiderhiser, 2013, p. 82)

Decades of attempts to find genes for the normal range of IQ, personality, socially disapproved behavior, and psychiatric disorders have been tried, and they apparently have failed.¹ The search for genes that researchers believe underlie IQ and personality goes back to the early 1990s, while the search for the genes presumed to underlie psychiatric disorders goes back to the 1960s and earlier. Psychiatric genetic researchers of the 1980s were preparing for the discovery of genes in that decade, but the discoveries never came (American Psychiatric Association, 2013a; Faraone, 2013; see Chapters 8–10). Although researchers argue that they need better methods and larger samples to uncover these long-lost genes, an alternative explanation holds that the methods that led them to believe that genes exist are faulty. First among these methods is the study of twins.

Despite the stunning failures of molecular genetic research, many observers have concluded that the “nature–nurture” debate is now over because “everyone knows” that human behavior is the result of both genetic (nature) and environmental (nurture) influences. The nature–nurture (heredity–environment) issue refers to the question of whether genes or environments are the most important factors influencing differences in human behavior and ability, even when it is assumed that both play a role. Human intelligence (cognitive ability), which IQ tests are purported to measure, has been a major area of focus in the debate. Although claims of the obsolescence of the nature–nurture debate are sometimes made by people suggesting or implying that “nature” has won, the debate continues in full force and has major implications for social policy decisions. It also shapes how we view our fellow human beings and the human condition in general.

While it is obvious that all organisms are the product of both their genes and their environments, the controversy has centered on whether genetic or environmental factors have a predominant influence on human behavior, or more properly, on human behavioral differences. “The nature–nurture debate,” as one author put it, “is essentially a question of the determinants of individual differences in the expression of specific traits among members of the same species” (Meaney, 2010, p. 41). When a behavior is called “genetic” it is often seen by the general public as unchangeable, or as largely resistant to change. When it is called “environmental” (non-genetic), attention may be drawn to the necessity of making changes in the environment. In both cases, the position that society takes on the issue often provides an impetus to take action of some kind. The developmental psychologist Richard Lerner pointed to words that are associated with the nature and nurture conceptions of development. Nature (genetic) terms include “genetic,” “heredity,” “inborn,” “innate,” “instinct,” “intrinsic,” “maturation,” “nativism,” and “preformed.” Nurture (environmental) terms include “acquired,” “education,” “environment,” “learning,” and “socialization” (Lerner, 2002, p. 19). Another nurture term is “malleability,” which refers to the ability to shape or change behavior.

Criminal behavior provides an example of the differing approaches society can take on the basis of whether a behavior is seen mainly as the result of heredity, or mainly as the result of the environment. For people taking an “environmentalist” position, strategies to reduce crime rates might include greatly increasing the funding of public education, paid maternity leave, increasing the minimum wage, passing laws and adjusting tax rates to greatly increase income equality, promoting full employment and job creation programs, fighting racism and other forms of oppression, restoring the “rehabilitation” aspect of incarceration, and so on. For people taking the “hereditarian” position that criminal behavior is largely the result of an individual’s genetic predisposition to commit crimes or other “antisocial” acts, possible strategies include early intervention programs for people seen as being predisposed to criminal behavior, increasing the size and funding of the police force, lengthening jail terms, building new prisons, instituting (or increasing) capital punishment, increasing the availability of contraception and abortion in the inner cities, promoting genetic counseling programs, and even the promotion of “parental licensure” laws (Lykken, 1995, 2000). A more extreme strategy, based on eugenics, is captured in the words of the German author of the first twin study of criminal behavior, who concluded in 1931, “We must try to make it impossible for human beings with positive criminal tendencies to be born” (Lange, 1931, p. 198).

Even if differences in human behavior are accepted as having an important genetic component, society might still choose to focus on improving the environment. Responding to a 1977 comment by hereditarian psychologist Hans Eysenck (1916–1997) that genetic interpretations of a twin study on “earning capacity” suggested that the British Royal Commission on the Distribution of Income and Wealth should “pack up,” the American economist Arthur Goldberger (1930–2009) wittily responded,

If it were shown that a large proportion of the variance in eyesight were due to genetic causes, then the Royal Commission on the Distribution of Eyeglasses might as well pack up. And if it were shown that most of the variation in rainfall is due to natural causes, then the Royal Commission on the Distribution of Umbrellas could pack up too.

(Goldberger, 1979, p. 337)

Research funding is also impacted by the approach society takes in dealing with the heredity–environment question. If it is directed in the genetic direction, other approaches suffer. As the historian of science Robert Proctor observed, “Scientific attention always comes at a certain cost: the decision to investigate one area is simultaneously a decision to ignore another” (Proctor, 1995, p. 243).

Since the late 1920s, the main technique used by supporters of genetic theories of human development and human behavioral differences has been twin research, which has been put forward as a scientifically validated research method that provides an ideal “natural experiment” for assessing the relative importance of heredity and environment. According to one estimate, by 2009 about 800,000 twin pairs had been studied (Johnson, Turkheimer, Gottesman, & Bouchard, 2009). Twin research has grown considerably over the past few decades, with studies now being conducted in 28 countries based on over 70 twin registries, and involving roughly 1.5 million participants (Hur & Craig, 2013). In almost all cases these studies are based on twin pairs reared together in the same family, while in an extremely small yet influential handful of studies, twin pairs were said to have been reared apart in different families.

Some critics, on the other hand, view most twin research as pseudo-science that is based on clearly unsupported theoretical assumptions and other biases, and is used to promote false hereditarian or “genetic determinist” ideas. A central aspect of this book, then, is an attempt to answer the crucial question of whether twin research is good science, uncertain science, or pseudoscience.

In this book I examine the use of twin research in the social and behavioral sciences. Behavioral science fields include psychology, psychiatry, and cognitive science. Social science fields include political science, economics, sociology, and anthropology. Although twin studies have been carried out by researchers in many fields, in the past few decades they have been promoted and defended by the (overlapping) fields of behavioral genetics and psychiatric genetics. The latter is a subfield of psychiatry, and will be examined more closely in Chapter 8. Twin research is also widely used in the study of non-psychiatric medical conditions.

Social and behavioral science research using human participants (subjects) is often referred to as “soft science,” as opposed to “hard” sciences such as physics, biology, and chemistry. In the hard sciences researchers are often able to rigorously control and identify the environments and variables used in their experiments, whereas social and behavioral science researchers often must gather correlational data based on people growing up in environments that the researchers did not design, control, or observe. Twin studies are an example of “non-experimental” research of this type.

Researchers using such correlational data based on uncontrolled and unobserved environments are forced to make assumptions about these environments, and the conclusions they reach usually depend on the validity (truth) of these assumptions. An assumption is something taken for granted or accepted as true without proof. The project or investigation then treats it, and researchers arrive at conclusions, as if it were true. Whether an assumption is true or false can completely change the findings of a study—for example, a finding that the behavior is caused or influenced by genetic factors, or a finding that it is caused by non-genetic factors.

The sociologist Howard Taylor once asked, “What assumptions does the researcher make? Are the assumptions explicit or implicit? Would changing an assumption alter the researcher’s conclusion a great deal or only slightly?” (Taylor, 1980, p. 9). These are very important questions, yet they are often overlooked. A major theme of this book relates to the question of whether the underlying assumptions of twin research, both stated and unstated, are true. If they are not true, a massive reevaluation of both twin studies and the theories based on them becomes the order of the day.

This book is divided into three parts. Part I focuses on investigations claiming to have studied “separated” twins reared apart in different family environments. The Minnesota Study of Twins Reared Apart is the most well-known and highly publicized study of this type, and is examined in detail. Leading up to that discussion, I will examine the three earlier “classical” reared-apart twin studies. I also look at some basic yet controversial concepts used in twin research and behavioral genetics in general, which include IQ testing, personality, and heritability. Part II looks at problem areas in the much more common studies of reared-together twins in the social and behavioral sciences. Here, I assess the validity of the most important theoretical assumptions underlying studies of reared-together twins, using the fields of political science and psychiatry as examples. I also examine other basic assumptions in psychiatry that have been the subject of criticism for many years, such as the reliability and validity of its diagnoses. In Part III, I examine the ongoing failure to uncover genes for behavioral characteristics and psychiatric disorders in the context of leading genetic researchers’ unfulfilled gene discovery claims and predictions. Part III includes a story that shows the potentially harmful consequences of emphasizing genetic explanations of human behavior and many common medical conditions, with the final chapter devoted to an evaluation of twin research, and the conclusions that follow.

I focus on twin research used to assess the role of genetic influences on behavioral differences. I will not cover other areas of research where twins or twin data are used, including the much less frequent use of twins to investigate the role of environmental influences. An example is the “co-twin control method,” where researchers assess the impact of environmental interventions or factors on monozygotic (MZ) twins, compared with their MZ co-twins who did not experience the interventions or factors. Another type of study using twins to assess environmental factors is the study of discordant MZ pairs, where researchers attempt to identify environmental differences between the twins that may have led to their differing psychiatric or medical diagnoses (see Mosher, Pollin, & Stabenau, 1971; the term discordant refers to one twin being diagnosed with a disorder, while the other is not).

The Historical Background

Chapter 2 of my 2004 book The Gene Illusion contained a detailed critical history of twin research, which should be consulted by those interested in this history (Joseph, 2004). Here I provide a brief summary. Twin research was initiated in the nineteenth century by the British statistician and founder of the eugenics movement, Francis Galton. The eugenics movement held that the human race can be improved by policies that promote selective breeding for “desirable” hereditary traits, and that prevent the reproduction of people and groups seen as harboring “undesirable” hereditary traits. Galton proposed the study of twins in an attempt to assess the relative roles of hereditary and environmental influences on intelligence and other psychological characteristics. In his article on twin research, he “was seeking some new method by which it would be possible to weigh in just scales the respective effects of nature and nurture” on the “intellectual ability of men.” Galton concluded, “...

Table of contents

• Cover Page
• Half Title Page
• Title Page
• Copyright Page
• Dedication
• Contents
• List of Illustrations
• Preface
• Acknowledgements
• List of Abbreviations
• I Studies of Reared-Apart Twins Scientific Proof or Scientific Illusion?
• 1 Introduction
• The Historical Background
• Human Behavioral Genetics
• Primary Methods of Behavioral Genetics and Psychiatric Genetics
• Family Studies
• The Twin Method (Twins Reared Together)
• Adoption Studies
• Reared-Apart Twin Studies
• Molecular Genetic Research
• Notes
• 2 Studies of Reared-Apart Twins Origins, publications, and scandal
• The First Studied Reared-Apart Monozygotic Twin Pair
• The Classical Reared-Apart Twin Studies
• Newman, Freeman, and Holzinger, 1937
• Sampling bias
• Were these really “reared-apart” MZ twins?
• Conclusion
• Shields, 1962
• Sample, goals, and biases
• Results and Shields' interpretations
• Degree of separation
• Conclusion
• Juel-Nielsen, 1965
• Overview
• Assumptions
• Lack of a control group
• Degree of contact and environmental similarity
• Other sources of bias
• Conclusion
• The Cyril Burt Scandal
• The Controversy Surrounding Arthur Jensen
• The Minnesota Study of Twins Reared Apart
• Selectively Reported MZA Pairs
• Other TRA Studies
• Conclusions
• Notes
• 3 Studies of Reared-Apart Twins The Critics Respond
• Leon Kamin, 1974
• Howard Taylor, 1980
• Susan Farber, 1981
• Leon Kamin, in Eysenck vs. Kamin, 1981
• Richard Rose, 1982
• Richard Lewontin, Steven Rose, and Leon Kamin, 1984
• Ken Richardson, 1998
• Leon Kamin and Arthur Goldberger, 2002
• My Previous Analyses
• Conclusions
• Notes
• 4 Studies of Reared-Apart Twins Basic Assumptions and Potential Fallacies
• Psychometrics
• Heritability
• Heritability ≠ Inherited
• Variation ≠ Cause
• Heritability and Psychiatric Disorders
• Some Leading Behavioral Geneticists Now Recognize that Heritability Estimates “Are Not Very Important”
• Model Fitting
• An “Important First Step”
• Questionable Assumptions
• Circular Reasoning
• Random Assignment
• IQ Tests
• Criticism of IQ Testing
• IQ Score Differences Based on Accepted and Rejected Assumptions
• “Personality” and Personality Tests
• Personality Tests Rejected by Leading Personnel Psychologists
• Cultural Influences
• Twins' Answers on Personality Test Questions
• Do MZA Pairs Grow Up in Different (Uncorrelated) Environments?
• Similar Appearance
• Other Cohort Effects
• Range Restriction
• Epigenetics
• Conclusions
• Notes
• 5 The Minnesota Study of Twins Reared Apart I Biases, Assumptions, and Other Problem Areas
• The MISTRA Twin Sample
• DZA Pairs: The MISTRA Designated Control Group
• MZA versus DZA Correlations
• How the Researchers Arrived at Conclusions in Favor of Genetics
• Finding # 1: The MZA Correlation is Significantly Higher than the DZA Correlation
• Finding # 2: The MZA correlation is Assumed to Directly Estimate Heritability
• Finding # 3: Model-Fitting Procedures Produce Substantial Heritability Estimates
• Assumptions That Are “Likely Not to Hold”
• Did the MISTRA MZA Pairs Experience Less Similar Environments than Pairs in the Previous Studies?
• Recruitment Bias
• Age and Sex Confounds
• Reliance on Twins' Accounts of Separation and Behavioral Similarity
• Sample Size
• Confirmation Bias
• A Gene for Pinkie Finger Curling?
• Failure to Make the Data Available for Inspection and Analysis
• A 1991 Exchange in Science
• The MISTRA Researchers and Leon Kamin
• Use of the MISTRA Findings by Disreputable Groups
• Access to the Data Must Become a Requirement
• Conclusion: The MISTRA was Based on a Large Number of Questionable Assumptions and Concepts
• Notes
• 6 The Minnesota Study of Twins Reared Apart II IQ and Personality Studies
• MISTRA IQ Studies
• Early Reports
• Two 1990 Publications
• The 1990 Science article
• The 1990 Acta Geneticae Medicae et Gemellologiae article
• Subsequent Reporting (and Non-Reporting) of the MISTRA DZA IQ correlations
• Bouchard, 1991
• Bouchard, 1993
• McGue, Bouchard, Iacono, and Lykken, 1993
• Bouchard, 1994
• Bouchard, Lykken, Tellegen, and McGue, 1996
• Bouchard, 1996
• Bouchard, 1997
• Bouchard, 1997
• McGue and Bouchard, 1998
• Newman, Tellegen, and Bouchard, 1998
• Bouchard, 1998
• Segal, 1999
• Bouchard and Pedersen, 1999
• McCourt, Bouchard, Lykken, Tellegen, and Keyes, 1999
• Bouchard and McGue, 2003
• Segal, 2003
• Johnson, Bouchard, McGue, Segal, Tellegen, Keyes, and Gottesman, 2007
• Segal and Johnson, 2009
• Segal, 2012
• Bouchard, 2013
• Bouchard, 2014
• Evaluation of the MISTRA Cognitive Ability Studies
• MISTRA Personality Studies
• Tellegen and Colleagues, 1988
• Two 1990 Publications
• Subsequent MISTRA Personality Studies
• Evaluation of the MISTRA Personality Studies
• Conclusions
• Notes
• II Studies of Reared-Together Twins
• 7 The Mzt–Dzt Equal Environment Assumption The Achilles Heel of the Classical Twin Method
• The Traditional Definition of the EEA
• The “Fallacy” of the Twin Method
• The Two Main Arguments Put Forward in Defense of the Equal Environment Assumption
• Argument A
• Argument B
• Argument A Potentially Renders Argument B Irrelevant
• Identity Confusion and Attachment
• Family Studies and the Twin Method
• A 2012 Defense of the Equal Environment Assumption
• Points # 1 and # 2: MZT Pairs Have More Similar Experiences, and Influence Each Other More, Than DZT pairs
• From Argument B . . . to Argument A
• Point # 3: Is More Similar Treatment (Based on Similar Appearance) a Genetic Effect?
• Point # 4: Do Prenatal (Intrauterine) Environmental Differences Support Genetic Interpretations?
• “An Obviously Confounded, Unreliable Methodology”
• Conclusions
• Notes
• 8 Twin Research in Psychiatry
• The Equal Environment Assumption in Psychiatric Twin Research
• Psychiatric Genetic Defenses of the EEA
• Kendler
• Flint, Greenspan, and Kendler
• Problems with Psychiatric Diagnoses
• Reliability and Validity
• The DSM-5 and Genetics
• Schizophrenia: The Classic Psychiatric Diagnosis
• Environmental Factors that Contribute to Schizophrenia and Psychosis
• An Implicit Assumption of Genetic Theories in Psychiatry
• An Article Emblematic of the Failures of Psychiatric Genetics
• Is Psychiatric Genetics Moving Toward the Status of a “Null Field”?
• Conclusions
• Notes
• III Approaching a Post-Behavioral- Genetics Era?
• 9 Molecular Genetic Research The Ultimate Test of Genetic Interpretations of Twin Studies
• Current Status
• Cognitive Ability (IQ)
• Personality and Behavior
• Psychiatric Disorders
• The “Missing Heritability” Explanation
• Challenging the Missing Heritability Position
• Missing Heritability and the Human Genome Project
• Four Key Problems with the Missing Heritability Argument
• Missing Heritability and the MISTRA
• A New Approach
• Negative Results, Once Again
• A Faltering Paradigm?
• Conclusions
• Notes
• 10 The Crumbling Pillars of Behavioral Genetics
• Three and a Half Decades of Claims and Predictions
• 1978–1990
• 1991–1995
• 1996–1999
• 2000–2002
• 2003–2004
• 2005–2011
• 2012–2014
• Conclusions
• Notes
• 11 A Human Genetics Parable
• Introduction
• A Human Genetics Parable
• Notes
• 12 Summary and Conclusions
• A Final Evaluation of Twin Research
• Appendix A: The Funding of MISTRA
• The Pioneer Fund
• Segal on the Pioneer Fund
• Rushton and Lynn
• Notes
• Appendix B: A Little-Known Behavioral Genetic Adoption Study Whose Results Contrast with the MISTRA Personality Findings
• Overview of the Study
• Alternative Explanations of the Results
• The Converging Evidence Argument
• Heritability and Model Fitting
• The 1998 CAP Study and the Twin Method
• The 1998 CAP Study and the MISTRA
• The 1998 CAP Study's Lack of Impact
• Conclusions
• Notes
• Appendix C: List of Quotations from Twin Researchers and Others Invoking the “Twins Create Their Own Environment” Argument A in Defense of the MZT–DZT Equal Environment Assumption of the Twin Method: 1954–2014
• (All quotations refer to twins reared together: MZTs and DZTs)
• Glossary
• References
• Index