Infant Cognition: Predicting Later Intellectual Functioning

I am pleased to welcome John Colombo's book to the Sage Series on Individual Differences and Development. I confidently predict that this book will become a classic because it caps a new wave of interest in individual differences in infant cognition. This is the best example of what can happen when a field that traditionally focused on normative themes of development begins to consider individual variations on those themes, the very standard deviation.

Personally, this book puts to rest an irritation that has bothered me for a decade. Ten years ago, the current edition of the Handbook of Child Psychology was published. I was on sabbatical and went through the thousands of pages, page by page, keeping myself alert by recording the number of pages that referred to individual differences. Day after day my irritation rose because more than 70% of the pages are normative, not even mentioning individual differences. The worst offenders were the chapters from the heartland of experimental research: perception. A total of 189 large pages of text with small print consists of chapters on visual perception; auditory and speech perception; and attention, learning, [Page xiv]and memory. Not a single page in these three chapters refers to individual differences.

The relative disregard of individual differences is unfortunate because, as Dr. Colombo's book attests, experimental psychologists have designed especially sensitive, process-oriented measures that can be very profitably applied to the investigation of individual differences. Other domains are also likely to profit by following the example of infant cognition by moving in the direction of individual differences. Two of the other chapters in the handbook that do not address individual differences are prime candidates: language and brain development.

Colombo's book also exemplifies some of the stresses that emerge as a field turns to consider individual differences. Compared to the study of means, the study of variance is more demanding in three ways. First, it is more demanding psychometrically. Dr. Colombo emphasizes that tests of infant cognition must be made more reliable. Indeed, consistent predictions of childhood cognitive abilities from these infant tests are all the more remarkable given the modest reliability of the infant tests. Second, individual differences research requires larger samples, but this strains against the intensive testing requirements of most measures of infant cognition. Third, individual differences research employs different statistics, the statistics of individual differences. These statistics focus on variability rather than treating it as “error valiance” in analyses of mean differences between groups. A subtle and somewhat paradoxical strain is that the statistics of individual differences are readily translated into the amount of variance explained. In contrast, analyses of group differences focus on statistical significance. The focus on variance explained is often humbling because it rudely reminds us, for example, that correlations of about .45 between tests of infant cognition and childhood cognitive ability only explain about 20% of the variance.

However, despite its bitter taste, this is important medicine. The preoccupation of the social and behavioral sciences with statistical significance has left a mountain of statistically significant results that are insignificant by any definition of societal relevance. The emphasis on variance explained is a virtue of [Page xv]individual differences research. Any mean group difference can also be converted to a statement of effect size. Rarely do such differences explain as much as 20% of the variance.

Another impediment to studying individual differences is that individual differences research often seems atheoretical. Without a theory, data gathering can lead to a collection of inconsequential facts. The issue, however, is not a dichotomy between theory and data but rather the balance between them. Researchers interested in individual differences tend to start with data and stay close to their data; their theories often remain implicit. As Dr. Colombo indicates in his preface, he was reluctant to be theoretical and speculative and was persuaded to do so only by persistent prodding by friends. The last chapter of this volume proves that his friends were right. Dr. Colombo develops a theory, although he modestly insists on calling it only a “speculative, tentative, working hypothesis,” that focuses on the processes underlying individual differences in infant cognition. This chapter sets the agenda for research in this field, an agenda that will maintain the field's momentum far into the next century.

The proliferation of research on individual differences in development has been especially dramatic in the study of information processing in the preverbal human infant. Much recent work has sought measures that might reflect aspects of information processing during infancy, with the hope that such measures might offer better prediction of later intellectual functioning than that afforded by more traditional standardized infant tests. This recent work has, to a degree, confirmed and realized this hope; four different measures that were developed as part of basic laboratory work on attention and learning during infancy have been shown to predict to later childhood intellectual performance.

The purpose of this volume is to present the historical and procedural contexts of these four measures of infant cognition; this is accomplished in four distinct steps. First, the history of the development of the measure is reviewed, followed by a relatively nontechnical account of how the measures are collected. Next, the psychometric properties of the measure involved are reviewed, leading to the final step, a review of the evidence on the predictive validity of the measure. On this last point, along with presenting evidence on how well the infant measure predicts to subsequent childhood measures (lagged prediction), the degree to which the [Page xviii]infant measure in question relates to other characteristics evident during the infancy period (concurrent prediction) is also addressed.

Two other issues are addressed in this book with respect to the predictive validity of these measures from infancy. The first deals with criticisms directed toward this research concerning the strength of the predictions involved (which have been declaimed as only modest in nature). The second concerns the number and nature of the infant information processing components thought to underlie this prediction.

There is both too much and not enough in this volume. Even the most casual reader will notice that I have spent more time on some measures than on others; the disproportionate length of Chapter 2 has caught the attention of several readers of earlier drafts of the book. Chapter 2 is a little longer because the measures discussed there have a longer history than do measures discussed in Chapter 3. During this more protracted history, there has been considerable variance with which the measures have been collected and applied. As a result, there is much confusion over which measures are best, and what they might reflect. Therefore, I have expended some effort in trying to clear some of this up in Chapter 2, particularly with respect to visual habituation. I hope that the reader will allow me this indulgence. Furthermore, some things are missing. For example, although I had been urged to review the research on the behavior genetics of infant information processing in Chapter 4, I have not done so because there are others who are more facile in the concepts and subtleties that guide that work. I have instead referred the reader to much more competent summaries of that area than I could have provided in the space allowed. This should not connote to the reader discomfort with either the topic or this research; behavior genetics is an exciting and fruitful field, particularly for the student of development.

In the course of writing this volume, I have accrued many debts. First, I thank Dr. Robert Plomin for asking me to contribute to the Sage Series, and C. Deborah Laughton at Sage for her unwarranted patience and tact. Dr. Plomin, Dr. Lee A. Thompson at Case Western Reserve University, and a third (anonymous) reviewer offered cogent and helpful comments on the first draft of the [Page xix]manuscript. Dr. Jeffrey W. Fagen at St. John's University and Dr. Marshall M. Haith at the University of Denver also provided critical suggestions on Chapter 3; I am especially grateful to Dr. Haith, who most graciously provided me with a copy of the complete file of unpublished materials on the visual anticipation paradigm and a cogent account of his theoretical interpretation of what occurs within it.

Here at Kansas, earlier drafts of either all or some of the manuscript were read and reviewed by Dr. Aletha Huston, Janet Frick, Jennifer Ryther, and Charlie Cleanthous (who helped me in securing permissions for tables and figures from other sources). Sara Coleman assisted in compiling the index.

In addition to the credits acknowledged above, I owe debts to a number of other individuals into whose stimulating intellectual environments I have stumbled at one time or another during my career. Dr. Frances Degen Horowitz facilitated and counseled my work during most of the past decade; whatever level of maturity and sensibility that I may be judged to possess at this point in my life may be reasonably attributed to her. J. Ken Nishita, Jack Meacham, Jim Pomerantz, and my adviser, Bob Bundy, were pivotal influences during my years at the State University of New York at Buffalo.

This book in many ways represents the influence of my longtime friend and collaborator, D. Wayne Mitchell, now at Southwest Missouri State University. Wayne and I worked together in conceptualizing the book, and his contribution is evident throughout Chapter 1. Furthermore, credit for the program of research on fixation duration, which is described in somewhat excruciating detail in Chapter 2, is equally shared with him. Unfortunately, Wayne had to withdraw from the project, because of other commitments, and would not allow himself to be listed as a co-author without a higher degree of specific contribution to it. Although he is not listed on the cover or title pages of this book, without his professional and personal support and counsel to me over the past 12 years, it would most certainly not exist.

Finally, I am most grateful for the support and encouragement of a relatively recent collaborator in my life, Dr. Dale Walker.

These acknowledgments aside, any errors of commission, omission, or misinterpretation that appear on the pages that follow are mine alone. In the past few years, I have been criticized on more than one occasion (and by more than one individual) for not being theoretical enough. Indeed, during a recent promotion review at the University of Kansas, a letter from an outside evaluator contained the criticism that I rarely speculated “beyond the data” at hand. (Luckily, the subcommittee viewed this as a positive attribute, and I got the promotion.) In any case, I have been encouraged to be a little more speculative in this volume, and I have tried to comply, especially in the concluding and admittedly speculative final chapter. Whatever the ultimate disposition of the model proposed there, I hope that students of the field find this bit of work helpful, and that it serves to generate further interest and inquiry into this important area of developmental psychology.