الرئيسية The Sourcebook of Nonverbal Measures: Going Beyond Words
The Sourcebook of Nonverbal Measures: Going Beyond WordsValerie Lynn Manusov
This book will serve both to introduce new methods and to familiarize readers with the variety of choices available in the study of nonverbal behavior. Highly Recommended.
You may be interested in
Most frequently terms
TeAM YYeP G Digitally signed by TeAM YYePG DN: cn=TeAM YYePG, c=US, o=TeAM YYePG, ou=TeAM YYePG, firstname.lastname@example.org Reason: I attest to the accuracy and integrity of this document Date: 2005.06.15 06:49:04 +08'00' The Sourcebook of Nonverbal Measures Going Beyond Words This page intentionally left blank The Sourcebook of Nonverbal Measures Going Beyond Words Edited by VALERIE MANUSOV University of Washington 2005 LAWRENCE ERLBAUM ASSOCIATES, PUBLISHERS Mahwah, New Jersey London Copyright © 2005 by Lawrence Erlbaum Associates, Inc. All rights reserved. No part of this book may be reproduced in any form, by photostat, microform, retrieval system, or any other means, without prior written permission of the publisher. Lawrence Erlbaum Associates, Inc., Publishers 10 Industrial Avenue Mahwah, New Jersey 07430 Cover design by Kathryn Houghtaling Lacey Library of Congress Cataloging-in-Publication Data The sourcebook of nonverbal measures: going beyond words / edited by Valerie Manusov. p. cm. Includes bibliographical references and index. ISBN 0-8058-4746-4 (cloth : alk. paper) ISBN 0-8058-4747-2 (pbk.: alk. paper) 1. Nonverbal communication. 2. Interpersonal relations. I. Manusov, Valerie. P99.5.S58 2004 302.2'22—dc22 2003049545 CIP Books published by Lawrence Erlbaum Associates are printed on acidfree paper, and their bindings are chosen for strength and durability. Printed in the United States of America 10 9 8 7 6 5 4 3 2 1 Contents Preface ix PartI INTRODUCTION TO NONVERBAL COMMUNICATION RESEARCH Researcher Choices and Practices in the Study of Nonverbal Communication Cindy H. White and Jack Sargent 3 Part II NONVERBAL MEASURES Participant/Observer Reports The Social Skills Inventory (SSI): Measuring Nonverbal and Social Skills Ronald E. Riggio 25 The Emory Dyssemia Index Marshall Duke and Stephen Nowicki, Jr. 35 The Affectionate Communication Index Kory Floyd and Alan C. Mikkelson 47 The Touch Avoidance Measure Peter A. Andersen 57 The Touch Log Record: A Behavioral Communication Measure Stanley E. Jones 67 V vi CONTENTS Measuring Live Tactile Interaction: The Body Chart Coding Approach Peter A. Andersen and Laura K. Guerrero 83 The Nonverbal Perception Scale Maureen P. Keeley 93 Self- Report Measures of Emotional and Nonverbal Expressiveness Ronald E. Riggio and Heidi R. Riggio 105 Measurements of Perceived Nonverbal Immediacy Peter A. Andersen andjanis F.Andersen 113 The Relational Communication Scale Jerold L Hale, Judee K. Burgoon,and Brian Householder 127 Coding and Rating Coding: Behavioral Coding of Visual Affect Behavior Patricia Noller 141 Assessing Display Rules in Relationships Krystyna S. Aune 151 Specific Affect Coding System Stephanie Jones, Sybil Carrere, and John M. Gottman 163 Measuring Conversational Facial Displays Nicole Chovil 173 The Nature and Functions of Tie-signs Walid A. Afifi and Michelle L. Johnson 189 A Procedure to Measure Interactional Synchrony in the Context of Satisfied and Dissatisfied Couples' Communication Danielle Julien 199 The Oral History Coding System: A Measure of Marital Cognition Kim Buehlman, Sybil Carrere, and Chelsea Siler 209 Rating: Observer Ratings of Nonverbal Involvement and Immediacy Laura K. Guerrero 221 Measuring Nonverbal Indicators of Deceit Judee K. Burgoon 237 CONTENTS vii Interactional Sensitivity: Rating Social Support, Attachment, and Interaction in Adult Relationships April R. Trees 251 The Affect Measures of the Couple Communication Scales Nigel Roberts and Patricia Noller 267 Rating Interactional Sense-Making in the Process of Joint Storytelling Jody Koenig Kellas and April R. Trees 281 Measuring Conversational Equality at the Relational Level Leanne K. Knobloch and Denise Haunani Solomon 295 Measures of Judged Adaptation Joseph N. Cappella 309 Methods for Measuring Speech Rate David B. Buller 317 Measuring the Relevance of Relational Frames: A Relational Framing Theory Perspective James Price Dillard and Denise Haunani Solomon 325 Assessing Attributions Given to Nonverbal Cues Valerie Manusov 335 Combining Reports and Ratings: The Expression of Rapport Frank J.Bernieri 347 Measuring Nonverbal Dominance Norah E. Dunbar and Judee K. Burgoon 361 Commentaries on Coding and Rating Choices: Analysis of Coded Nonverbal Behavior Roger Bakeman 375 Coding Mutual Adaptation in Dyadic Nonverbal Interaction Joseph N. Cappella 383 Physiological Measures Objective Measurement of Vocal Signals Kyle James Tusing 393 viii CONTENTS The Role of Physiological Measures in Understanding Social Support Communication Terry A. Kinney 403 Part III PARADIGMS AND PRACTICES Standard Content Methodology: Controlling the Verbal Channel Patricia Noller 417 The Passing Encounters Paradigm: Monitoring Microinteractions Between Pedestrians Miles L. Patterson 431 The Meta-Emotion Interview and Coding System Eve-Anne M. Doohan, Sybil Carrère, and Marianne G. Taylor 441 Measuring Emotional Experience, Expression, and Communication: The Slide-Viewing Technique Ross Buck 457 Conflict, Real Life, and Videotape: Procedures for Eliciting Naturalistic Couple Interactions Linda J. Roberts 471 Meta- Analysis of Nonverbal Behavior Judith A. Hall 483 Considering the Social and Material Surround: Toward Microethnographic Understandings of Nonverbal Communication Curtis D. LeBaron 493 Nonverbal Research Involving Experimental Manipulations by Confederates Laura K. Guerrero and Beth A. Le Poire 507 Author Index 523 Subject Index 537 Preface Valerie Manusov University of Washington A few years ago, a bold graduate student in communication, Jack Sargent (now a faculty member and contributor to this book) asked me, and several other experi enced researchers who study nonverbal behavior, to be part of a conference panel. At that time, Jack was interested in doing a study that used coding and rating tech niques to measure nonverbal cues, and he could not find a current published source to help him. So he gathered us together in Denver, Colorado, and asked us a series of questions. It was useful for Jack but also for those of us on the panel—and hopefully for the audience as well. All of us realized that much of the work we had been doing involved creating and re-creating measures to assess the aspects of non verbal behavior in which we were most interested. We all concurred that having a more cohesive discussion of research choices specifically for the investigation of nonverbal phenomena seemed important. That is the primary aim of this book. All of us who study—or wish to include—nonverbal cues in their scholarship know the myriad ways that nonverbal behavior can be conceptualized. Depending on how we think about the nature, functions, and meanings of nonverbal cues, we are likely to go about investigating them in a certain way. The complexity and ambi guity of nonverbal processes and products allows for a plethora of research oppor tunities. It also provides for so much variety that it may be difficult sometimes to compare research findings. An additional aim of this volume, then, is to encapsu late some of the primary means by which researchers assess nonverbal cues with the hope that these may be used by others as well. The particular selection of chapters to include in this volume was strategic. My hope was to include some well-known and repeatedly validated assessment mea sures alongside some novel approaches just beginning their research "journeys." I also wanted specifically to include research and researchers from a number of disci plines, as nonverbal behavior is one of those research topics that spans an array of scholarly interests. Additionally, and although the term "measurement" entails, typically, an orientation toward quantitative assessment, I hoped that this book ix X PREFACE would show qualitative means through which nonverbal cues can be illustrated and understood. It is my good fortune that most of the potential contributors I con tacted said "yes," and this volume has taken much of the shape I hoped it would. It has also been expanded farther than my original intentions, and I am glad of this. Specifically, in addition to measures that can be used by other researchers with an interest in nonverbal cues, much of the book also orients around broader issues involved in doing this research. Several of the chapters, for example, provide discus sion of larger research "paradigms" into which a particular measure may be placed. Other chapters are devoted more to strategies that researchers have available to them to help answer their questions. The latter includes discussions meant to help researchers think through the array of choices available to them and to show the strong connection between the type of answers one seeks and the means through which she or he may seek them. The book is limited necessarily to those measures that, for the most part, can "fit" into these pages. Important, but "larger," well-known assessments, such as Rosenthal's Profile of Nonverbal Sensitivity, and Paul Ekman's Facial Action Coding System, were too big for this book but are available elsewhere (Judith Hall, a contributor to this book, can be contacted for PNS, and Ekman can be contacted for FACs and several other very useful assessments). We hope, however, that having most of the measures actually printed in this volume will make it a particularly use ful sourcebook. The authors of these chapters and I hope that this book will be useful to its audi ence. We encourage each reader to read it in the manner that works best for his or her individual needs. Some people may read only those chapters that have specific relevance to a particular study they are planning. Others, especially those who wish to become researchers of nonverbal processes and/or those taking classes in non verbal behavior, may benefit most from reading all parts of the book to look at how these authors made the choices they did. Whatever form of help this book takes, I am glad of it. I am also glad to have this chance to thank a few people for this book's creation. In addition to Jack Sargent, whom I mentioned previously, and another colleague, Kimo Ah Yun, who also lamented on the difficulty of finding a range of methods eas ily for studying nonverbal cues, I thank the other authors for their willingness to write these excellent chapters. Most specifically, thanks to Kory Floyd for allowing me to use his chapter as an exemplar to entice others to join the project. I also thank Lane Akers at Lawrence Erlbaum Associates for calling me back excitedly the first morning he received my proposal, Linda Bathgate, also at LEA, who lived up to the extremely positive rumors about her character and competence, and Sara Scudder, a senior book production editor at LEA who was the consumate copy editor. Thanks also to my colleagues in the Department of Communication at the University of Washing ton, who always provide encouragement for the work that I do. Finally, my love and thanks to my son, Cameron, and my husband, Chuck, who provide ample opportu nities to see the importance of nonverbal cues in the everyday of our lives together. I. INTRODUCTION TO NONVERBAL COMMUNICATION RESEARCH This page intentionally left blank Researcher Choices and Practices in the Study of Nonverbal Communication Cindy H. White University of Colorado Jack Sargent Kean University INTRODUCTION There is sometimes a temptation to believe that a book like this provides research ers with ready-made approaches for examining interaction and answering re search questions. So, we begin this chapter with a caution that fulfilling such a belief is unlikely because ready-made research solutions are rare. A book like this one can, however, be a great asset to researchers, because such volumes allow re searchers to see dearly what others are doing and to make thoughtful decisions about how to examine nonverbal behavior. Decisions about measurement, whether the measures involve self- (or other-) reports, coding, rating, or physiological mea sures are theoretical acts (Bakeman & Gottman, 1997). They are the researchers' assessments of what is important in the interaction and what will have relevance to the theory being tested or the context being examined. It makes sense, then, to be gin this book with a discussion of some of the decisions that researchers face as they begin to examine nonverbal cues. Once a researcher makes that first decision to study nonverbal behavior, he or she is put on a path that requires many other decisions. These decisions include who is going to be studied, what type of interaction will be studied, where the study will take place, how the study will be designed, which behavior will be examined, and how long the behavior will be observed (Scherer & Ekman, 1982, see Cappella, this volume and Dunbar & Burgoon, this volume). Some of these decisions are easy be cause they are indicated clearly by the behavior or situation being studied (for in 3 4 WHITE AND SARGENT stance, if a researcher wants to study affection displays during airport departures, he or she has to go to the airport as Heslin & Boss, 1980, did). But many of these choices are decisions the researcher must make along the way as he or she considers how to best capture the elements of interaction that are important and meaningful (e.g., should he or she study touch during departure or facial expressions or both, across what time frame, and among which passengers?). Moreover, each of these choices has implications for the others, so that the researcher is faced not with discrete choices but with a set of interrelated judgments. In the sections that follow, we examine the advantages and disadvantages of some selected research choices related to the study of nonverbal communication. We first discuss the potential uses and limitations of self-reports for the study of nonverbal behavior. Then we explore in depth the decisions related to categoriza tion of observations, because oftentimes the nature and the practical enactment of coding or rating schemes are somewhat undetectable once a study comes to publi cation. Specifically, we examine choices researchers make about the type of obser vational system to employ (type of observation scheme, coding versus rating), decisions regarding where to collect data (field or laboratory settings), and issues re lated to the nature of the interaction being examined (coding individuals versus dyads/groups, examining interaction events or coding slices of time). We then dis cuss briefly when a researcher may choose physiological measures. Our goal is to provide readers with an introduction to some of the issues nonverbal researchers grapple with and to highlight chapters in this volume where readers can see that choice in action. Self-Reports of Nonverbal Behavior One choice researchers have is whether to use self-report measures in their studies. Self-report tools rely on participants' own perspectives (or, occasionally, the per spectives of others who are connected to the one observed), rather than the views of independent observers, to not only determine what specific behaviors were en acted but also delineate their meanings and significance. These measures are often used to gather information about microbehaviors, naturally occurring social in teractions, meanings assigned by observers to behaviors, or to assess numerous communication-related skills and abilities across a variety of contexts (see Riggio, this volume; Riggio & Riggio, this volume). For instance, in a number of studies, Stanley Jones has used the self-report Touch LogRecord to record the touching be havior of individuals. In his chapter in this book, he describes work by Jones and Yarbrough (1985), who were interested in the meanings associated with specific touching behaviors. Their study identified 12 mutually exclusive meanings associ ated with specific touching behaviors, and the self-report coding system generated in their work is one available framework. In another example, Palmer and Simmons (1995) used a self-report method that asked research confederates to re cord in an open-ended questionnaire the specific behaviors they used to convey RESEARCHER CHOICES AND PRACTICES 5 liking to their interaction partner. Later these responses were content analyzed. This allowed the authors to compare participants' understanding of the behaviors they had used with observer assessments of that behavior. The advantages of nonverbal self-reports are many. They are easy and inexpen sive to administer (see Duke & Nowicki, this volume), the costs may be lower than for typical behavioral observation, and they provide access to a wide variety of inter actions that may not be available any other way (see Keeley, this volume, and Floyd & Mikkelson, this volume). In addition, because participants are not required to re port to a research site, they may be more willing to participate. Finally, many self-report measures have strong reliability (see Andersen & Andersen, this volume; Riggio & Riggio, 2001) and are correlated positively with independent observer rat ing tests. Despite the benefits of using self-report measures to examine nonverbal com munication, a number of limitations to their use also exist. These types of measures can be less precise than coders' observations and more prone to response biases, such as social desirability. Additionally, participants may recall their own behaviors more accurately than they recall the behavior of the person with whom they are in teracting (Metts, Sprecher, & Cupach, 1991; Riggio & Riggio, 2001). Some re searchers even question if self-report measures record only "impressionistic" rather than the "actual" nonverbal behaviors (Metts et al., 1991). Furthermore, partici pants may be overwhelmed by the amount of recording that is required from them. Jones (this volume) points out that participants need to be "motivated to record events conscientiously" but that this can be remedied through training. Partici pants, also, need to understand what specific behaviors or proximal cues the re searcher is looking for, as well as the differences in the various constructs that the participant is expected to measure. Numerous self-report measures of nonverbal behavior exist—particularly mea sures of emotional expressiveness—a number of which are discussed in this book. For a more thorough review of a number of these measures, as well as their internal and construct validity, see the section on self reports in PART II. Observer Assessments of Nonverbal Behavior Although we can learn much about human interaction by asking communicators to reflect on their own behavior and the behavior of others, such reflections cannot always provide detailed information about the myriad behaviors that are actually enacted as people communicate. If a researcher wants to learn how the behavior of one person influences another or observe differences in how interaction in man aged (see Cappella's discussion of coding, this volume), for example, he or she may need to examine interaction as it occurs. Observational studies of nonverbal behavior provide insight into a number of communication and relational processes. Researchers have demonstrated, among other things, the ways in which deceptive interactions are shaped jointly by both 6 WHITE AND SARGENT participants in a deceptive conversation (e.g., Burgoon, Buller, White, Afifi, & Buslig, 1999; see Burgoon, this volume), the way interaction between relational partners is impacted by their respective attachment styles (e.g., Guerrero, 1996), the differences in nonverbal behavior that can distinguish couples whose relationships are successful from those who are not (e.g., Gottman & Levenson, 1992; Manusov, 1995), and the importance of nonverbal behavior in adult-infant bonding and in teraction (Bernieri & Rosenthal, 1991; Cappella, 1981). Observational studies focus the attention of researchers on the actual behavior of participants and provide an opportunity for researchers to explore how meanings and feelings are manifest in interaction. Although it is clear that observational studies can reveal many interest ing aspects of interaction, the decision to observe behavior is not a simple one. A researcher who decides to observe behavior is faced with a number of choices about what to observe and how to catalog observations in an observational system. Although it might seem that what to observe would be clear once a researcher has decided on a research question or identified a focus of the study, the decision about what to code actually entails many judgments about what constitutes communica tive action and how social interaction is organized (see Bakeman, this volume). Bakeman and Gottman (1997) noted that coding schemes could be thought to exist along a continuum, with one end anchored by physically based schemes that reflect the organism's physiology, and the other end anchored by socially based schemes "that deal with behavior whose very classification depends far more on the mind of the investigator (and others) than on the mechanisms of the body" (p. 18). In the discussion that follows, we focus on what we consider to be socially based schemes; that is, observational systems that examine behaviors or messages that have more to do with social categories of interaction (such as smiling or involve ment) than with physiological elements of behavior (such as amplitude; for more on physiological measures, however, see Kinney, this volume; Tusing, this volume). Socially Based Measurement Schemes One very important decision that researchers make when examining nonverbal communication via a socially based coding scheme is the level of measurement. By this term we mean that researchers make choices about the amount of behavior they will examine within an interaction and the extent to which the assessment in volves more concrete indicators of the behavior's occurrence or more abstract as sessments of the social meaning of the behavior. Burgoon and Baesler (1991) discussed this choice as one of micro versus macro levels of measurement. Spe cifically, they argue that micro level measurement "involves single, concrete be haviors," whereas macro level measurement involves "larger samples of a given behavior or collection of behaviors" (p. 59). They also noted that macro level mea surement typically entails more abstract terms and larger time intervals than does micro level measurement (see Cappella, this volume, on time choices), but they distinguish the level of measurement from the abstractness of what is measured. RESEARCHER CHOICES AND PRACTICES 7 In nonverbal research, it is typical for coding of larger amounts of behavior to also involve perceptual judgments that entail interpretation by coders. Likewise, micro level coding tends to involve more concrete assessments of single behaviors. For instance, assessment of involvement typically involves coders' assessments of the extent to which the relational message is displayed based on examination of a number of behaviors that are used to convey involvement (see Guerrero, 1996, this volume); this choice can be contrasted with coding that assesses the occurrence of a specific behavior such as smiling (Julien, this volume; Segrin, 1992). Thus, we dis cuss this distinction as one of rating versus coding. Rating. Rating entails having coders attend to a set of behaviors that comprise the meaning or message conveyed within an interaction. For instance, a number of deception researchers have examined the involvement displayed in interaction (Burgoon et al., 1999; White & Burgoon, 2001). Ratings of involvement reflect ob servers' assessments of the degree of involvement displayed (i.e., the meaning), but such assessments are based on careful observation of a set of behaviors that are re lated to involvement (e.g., eye contact, forward lean, body orientation). Similarly, Knobloch and Solomon (this volume) trained coders to rate the level of conversa tional equality displayed in interaction based on verbal contributions and nonver bal elements of interaction that contribute to equality of the conversation (such as eye contact, volume/rate of speaking, and gestures). One advantage of rating is that it tends to make effective use of raters' time by asking them to make more comprehensive judgments of larger chunks of behav ior. Another advantage of this type of coding is that it reflects what Burgoon and Baesler (1991) refer to as isomorphism between the "coder's [or rater's] task and the phenomenological experience of communicators" (pp. 60-61). In others words, these types of judgments are similar to the types of judgments made by communicators during interaction, and they take advantage of the ability of rat ers to understand how behaviors work together to convey social messages. Finally, rating is often sufficient for research projects that seek to understand how social meaning relates to aspects of the relationship or interaction task (see Koenig Kellas & Trees, this volume; Manusov, this volume) or how it influences interac tion outcomes (see Buller, this volume; Roberts & Noller, this volume; Trees, this volume). Rating does, however, limit researchers' understanding of interaction in impor tant ways. First, ratings require considerable inference on the part of raters; this means that the perceptual judgment being made by the raters must be elaborated clearly, the observers must be well trained, and the reliability of their perceptions must be checked and recalibrated throughout the entire rating process. Addi tionally, rating does not provide insight into the extent to which specific behaviors contribute to interaction processes (Burgoon & Baesler, 1991). Finally, when be haviors are changing frequently across an interaction, rating may be ineffective be cause raters cannot capture the dynamic nature of the interaction effectively. 8 WHITE AND SARGENT Coding. Coding, on the other hand, provides a level of precision and accuracy that may be very useful when researchers wish to understand how micromomen tary actions contribute to interaction dynamics (see Aune, this volume; Chovil, this volume). This type of assessment has been used successfully in research on marital dyads (Buehlman, Carrere, & Siler, this volume; Doohan, Carrere, & Taylor, this volume; Jones, Carrere, & Gottman, this volume; Noller, this volume), mother-infant interaction (Cappella, 1981), initial interaction (Palmer & Simons, 1995), and couples' communication (seeAfifi & Johnson, this volume; Dillard & Solomon, this volume). One advantage of coding is that it often yields high levels of reliability because single behaviors are examined, and their presence or absence is relatively easy to identify. When a number of separate behaviors are coded, researchers also have the opportunity to combine behaviors in their analyses so that they can examine the in fluence of individual behaviors and sets of behavior as they co-occur as well as the individual variance contributed by each behavior vis-a-vis the other cues. Finally, when assessments of a behavior are made across time, coding is a good way to cap ture if and how behaviors change across an interaction. The disadvantages of coding relate to the fact that typically it does not address the social meaning of a behavior (i.e., the focus is on the behavior's occurrence). As a result, coded behavior may not predict interaction outcomes as well as rated be havior does. Additionally, coding of behavior is often more time and cost intensive than is rating (Burgoon & Baesler, 1991). Also, although coding specific behaviors allows researchers to combine behaviors for analysis, it is unclear how this should be accomplished, as it seems certain that the impact of different behaviors is not di rectly "additive." As a result, the combinations of behaviors that researchers exam ine may not have relevance to the experience of communicators in interaction. Choosing Rating or Coding. Cappella (1991) argued that coding and rat ing (as well as participant judgments and untrained observer assessments) pro vide "different frames of reference" (p. 1 l l ) from which to view what is happening in interaction. He asserted that although each approach yields data that are somewhat different, information from each approach can be reconciled if researchers take into account the differences in point of view that influence what is captured by each approach (see Bernieri, this volume; Dunbar & Burgoon, this volume). Such a view of rating and coding is useful because it re minds us that choosing to rate or code is really a decision about the point of view from which we will see things. Coding provides a tighter focus, fixing our view on specific behaviors, whereas rating shifts our gaze to a wider angle, encourag ing us to see the social features of interaction. Researchers' decisions about which form of observational assessment to use should reflect the aspects of in teraction they wish to describe. Of course, the type of communication situations we are studying and the nature of the data we have collected can affect the point of view we select for our analyses. RESEARCHER CHOICES AND PRACTICES 9 Selecting and Training RatersICoders. The issue of coder/rater selection and training has received little attention in writings about nonverbal research. Re searchers involved in observational research, however, are aware that coder selec tion and training is an important aspect of the research process. Although there is not space here to provide a complete overview of issues related to rater/coder selection/training, we mention some of the decisions that researchers must make as they begin to work with coders/raters. In terms of selection, coders/raters need to be skilled observers of interaction. That is, they need to be able to discern what behavior is being displayed and, in some cases, to make a judgment of the message conveyed by the behavior. The substantial research on nonverbal sensitivity and decoding skill (see Riggio, this volume) sug gests that, in many cases, female coders/raters are likely to be more effective than male coders/raters, although we know of no research that explores this specifically. Additionally, the "effectiveness" of coders might depend on the nature of the cod ing system being applied and the nature of the interaction being observed. Coders/raters who have experience with the type of interaction being examined may make different judgments than coders who know little about the interaction setting or the experience of the interaction situation (Woolfolk, 1981). Coders/raters can provide an insider or outsider perspective on the interaction. Which is more valuable depends on whether the researcher assumes that the coding/rating system is largely objective and requires only careful observation of the behavior displayed or if the researcher assumes that the meaning of the behavior displayed is informed by an understanding of interaction dynamics and the situa tion. In short, selecting coders requires researchers to think carefully about their as sumptions regarding the way communication is conducted. Some researchers employ a number of coders/raters to watch an interaction and provide a holistic assessment of the interaction. This type of assessment, although useful, relies on observers' intuitive assessments of interaction, and trained coders or raters receive instruction in how to understand and apply an observational sys tem. Additionally, it is worthwhile to consider issues of coder/rater training and the management of coders/raters throughout the project. Guerrero (this volume) pro vides a number of helpful suggestions for such training. She also provides sugges tions for ways to make observers' tasks more manageable. We simply add to her discussion the idea that when researchers train coders/raters, they are, in a sense, provided with an opportunity to see how their concepts and observational systems intersect with people's understandings of interaction. The observational system that coders/raters use to examine interaction is a way of mak ing sense of the interaction. Training provides a way to learn if the system is coher ent enough for coders to apply it. Additionally, the demonstration of coder/rater reliability initially, after training, and throughout the coding process provides a way for researchers to determine if their observational system can be applied in a consis tent way (see Bakeman, this volume, for discussion of assessment of interrater reli ability). In sum, coder/rater selection and training can be treated by researchers as 10 WHITE AND SARGENT an opportunity to reflect on the assumptions inherent in the observational system they have developed. Such reflection may provide researchers great conceptual in sight into the work that they are doing. Choices Regarding Where to Collect Data Imagine that a researcher decides he or she would like to know how nonverbal cues of politeness influence the way others respond to a request. Because requests occur in many settings, the researcher has numerous options for studying this process. He or she can monitor requests for help at a local library, watch what happens when people ask for directions at a gas station, or spend time observing mishaps at a local ice skating rink (see LeBaron, this volume). Alternately, he or she could train confederates to ask for assistance in any of these settings and watch what oc curs, or bring participants into a lab setting, creating a situation where confeder ates ask participants for help (see Guerrero & Le Poire, this volume). The decision of where and how to study nonverbal behavior is important be cause it has implications for the specific aspects of behavior that can be observed, the extent to which the behavior occurs in a context of ongoing interaction, and the way that behavior can be captured for analysis. In some cases, the nature of the be havior being studied dictates the context in which it must be observed. For instance, one cannot study wedding behavior in the laboratory (we hope!). But, in other cases, researchers can choose to observe behavior in the field or transport it into the laboratory. Both settings have advantages and disadvantages that the researcher needs to consider in the design of the study. Field Study Environments. When nonverbal researchers discuss conducting their studies in field environments, they are referring to the physical setting in which the communication event or behavior occurs, such as in bowling alleys, living rooms, classrooms, and shopping malls. Coding in naturalistic settings is advantageous when little is known about a specific behavior because these venues provide a better under standing of how behavior is exhibited (Scherer & Ekman, 1982; see Roberts, this vol ume, for a discussion about making "naturalistic" laboratory observations). In fact, some researchers have argued that an understanding of nonverbal behavior can only really be garnered by studying interaction that occurs within natural contexts because communication is part of the social and physical context in which it is used (Jones & LeBaron, 2002). Although we recognize that behavior is situated, for most research ers, the decision of whether or not to study nonverbal behavior in the field is tied to the extent to which the behavior can be examined and understood adequately in the field. In this section, we consider the advantages of studies that involve observation of naturally occurring behavior and then examine situations where a manipulation of behavior is introduced in a field setting. The decision to study nonverbal communication in a field setting is often made because a researcher wants to retain the spontaneity and "situatedness" of the be RESEARCHER CHOICES AND PRACTICES 11 havior. If a researcher chooses simply to observe behavior in the field (which Hecht & Guerrero, 1999, called naturalistic observation), he or she does not attempt to in fluence the behavior in any way. For example, a researcher may be interested in studying the immediacy behaviors of romantic partners departing on international flights at airports. In this situation, the researcher codes the couple's behaviors as they occur naturally. The advantages of observing behavior in this way are many (see Buck, this vol ume, and Patterson, this volume, for discussion of useful observation methods). The behavior is enacted and motivated by the needs of the communicators; it is re sponsive to a real situation and is likely to reveal how social norms and contexts in fluence interaction. Additionally, as Scherer and Ekman (1982) note, some events such as weddings or political rallies mustbe observed in the field in order to under stand the nature of communication within the event. Of course, naturalistic observation also poses a number of challenges. It may be difficult to observe a large number of naturally occurring interactions that are simi lar in nature, and such coding typically offers the researcher only limited access to the thoughts and reactions of participants, things that are often of interest to re searchers. Moreover, until recently this type of research required that researchers rely on the observations of trained coders, because no record of the interaction could be obtained easily via videotape. Advances in the size and nature of video equipment have made it more feasible to record field situations relatively unobtru sively, which may provide an opportunity for researchers to capture and review communication events observed in the field. This additional ability does not, how ever, provide researchers the chance to examine the effects of particular communi cation behaviors on interaction events as well possible ethical concerns arise in unobtrusive recordings. There are situations, however, in which the behavior being examined in the field allows researchers to introduce some type of manipulation into the study environ ment. These studies are often called field experiments or field studies (Hecht & Guerrero, 1999). This type of design not only has the advantage of allowing the re searcher to observe behaviors in a naturalistic setting, but it also allows researchers to observe the behaviors repeatedly and to have greater control over the specific manner and setting of the behavior. Field experiments also allow researchers to ex amine how behaviors occur differently under varying natural environments. When introducing a manipulation, researchers often ask participants to engage in certain behaviors and then observe the behavioral reactions of the study partici pant or another person within the interaction. Field experiments have the advan tage of situating behaviors in a context and reducing the likelihood that participants are stylizing their communication in reaction to the laboratory setting. When studying communication in natural settings, scholars face a number of challenges that exist simply as a result of the venue. One of the most difficult issues researchers encounter in examining communication behaviors within this envi ronment is seeing the behavior occur with enough frequency and strength to study. 12 WHITE AND SARGENT In other words, observing similar, repeated occurrences of the desired behavior can be difficult. Researchers may have to wait long periods of time to observe specific behaviors, and then these behaviors may occur only fleetingly (Patterson, Webb, & Schwartz, 2002). The lack of control of over the frequency of behavior and the na ture of individuals engaging in the communication behavior must be weighed early on in the research process in deciding where to observe the communication. Choices such as the ones described earlier will have significant effects on the system atic study of the behavior. Laboratory Settings. Researchers also have the option of conducting their studies in a laboratory. Laboratory-based studies provide the greatest amount of control over a behavior and can lead to increases in internal validity: Various set tings can be created and behaviors observed with greater detail (see Roberts, this volume). Scherer and Ekman (1982) suggest that when the specific event under study is "problematic," laboratory-type experiments may yield the greatest reliabil ity due to the difficulty of observing behaviors under similar conditions in more naturalistic environments. Furthermore, researchers may wish to use a laboratory if they are interested in videotaping the behaviors of their participants. By reviewing videotapes in a controlled setting, researchers may be able to increase the accuracy of aggregating data across individuals and situations. Videotaping is advantageous when studying the microbehaviors of individuals, such as eye blinks or participants' eyebrows raises, and it provides the researcher the opportunity to reexamine the data, asking different questions. Laboratory studies allow researchers to manipulate the behavior or intentions of the participants, the situation/context or task, or some combination of these. For instance, the researcher can ask the participant to engage in a specific role or behav ior, either emphasizing or deemphasizing it during the interaction. Burgoon, Olney, and Coker (1987) used this approach when they had one participant in crease or decrease nonverbal involvement during mock employment interviews. Alternately, Palmer and Simmons (1995) gave participants the goal (intention) of conveying liking, and then examined what behaviors participants used in a conver sation with a stranger. Other studies have explored how deception is enacted (Burgoon et al., 1999) or how nonverbal cues are related to accounts of failure events (Manusov & Trees, 2002). As we have noted, not all environments or interactions can be recreated in the laboratory. Despite this, many behavioral interactions between individuals in a lab oratory are a good diagnostic of behavior as it occurs in the natural settings, because the interactional resources that communicators have at their disposal are present in the laboratory just as they are in everyday interaction (Bavelas, 1995). Nonetheless, when researchers use the laboratory as a site for interaction, there are a number of issues to consider. Researchers need to have participants engage in familiar behaviors that they feel comfortable enacting. If participants are asked to engage in a behavior or action that they believe is irrelevant for the RESEARCHER CHOICES AND PRACTICES 13 situation, then it is likely their behaviors will be contrived and unnatural (Scherer & Ekman, 1982). Additionally, participants may not have much of a stake or investment when conversing with individuals with whom they share no past or any possible future (such as confederates or strangers). As a result, their behavior may be different from what it would be with those whom they know or with whom they have some relational investment; this can be modified if the re searcher chooses to use relational partners or friends in laboratory studies, but needs to be considered. Finally, the environment or situation chosen needs to have some relevance to participants' frame of reference. If a researcher creates a situation for participants that is unfamiliar or that they feel is inappropriate, then unless the researcher is looking purposely for some specific effect, partici pants' resultant behaviors may be very different from what would happen in a naturally occurring interaction. In sum, whether a researcher chooses to conduct his or her study in either the field or laboratory, there are a number of decisions that go hand-in-hand with this choice, such as whether to manipulate a behavior, employ a confederate, or video tape the interaction. Ultimately, the study's research questions or hypotheses should drive which context the researcher selects. Once the researcher has se lected a venue for the study, he or she must consider if examination of the data in volves focusing on individuals or exploring the aspects of interaction that are jointly produced by all participants. The chapters in Part III of this book, on "Par adigms and Practices" provide additional discussion of these choices. The Nature of Data: Coding the Individual Versus the Interaction, Relationship, or the Group One important decision about the nature of data concerns how the behavior that is produced in the interaction. Although some research, such as work on person per ception, focuses on how individuals assess another's nonverbal display, most re search on nonverbal communication involves situations where two or more individuals are interacting. Scherer and Ekman (1982) argue that in such situa tions, the behavior of all participants should be sampled and measured. But deci sions about sampling and measurement are fairly complicated. We see two decisions as important in this area. First, researchers must decide if they will code/rate the behavior of each participant separately or if they will code/ rate a feature of the interaction that reflects the joint behavior of both participants. Second, researchers must decide how to treat behavior in analyses. The advantage of assessing the behavior of individuals is that such coding allows researchers to de termine, at least to some degree, how the behavior of one person affects the behav ior of another. For instance, Jones, Carrere, and Gottman (this volume) use the Specific Affect Coding System to examine the emotional displays of each marital partner and then to consider how the emotion of one partner influences the reac tions of the other. 14 WHITE AND SARGENT Likewise, a researcher might be interested in determining if a decrease in in volvement on the part of one participant has an effect on the involvement level of another interaction participant (e.g., White & Burgoon, 2001). In order to assess this, the researcher must code/rate the involvement of each participant separately and then must examine change in involvement of each participant across the inter action (see Julien, this volume). This type of coding also has the advantage of allow ing researchers to examine how individual characteristics of participants (such as goals or personality characteristics) are related to the behavior of an individual in the interaction. Researchers may also choose to examine a feature of the interaction that is jointly produced. In this case, the assumption is that the feature of interaction being examined can only be understood by examining the combined behavior of partici pants; such work typically seeks to reveal how features of interaction distinguish particular types of dyads or groups from one another. For instance, Knobloch and Solomon (this volume) describe a system for rating conversational equality: Con versational equality is a feature of the interaction that is revealed by examining communicators conversing together. It cannot be assessed for an individual. Cappella (this volume) describes ratings systems for assessing adaptation between partners, which reflects the fit of partners' behaviors with one another. Similarly, Koenig Kellas and Trees (this volume) describe a system that assesses the engage ment and coherence of a jointly produced family story. The aspects of storytelling they code reflect the nature of the storytelling endeavor as a whole rather than the contributions of specific participants. The advantage of this type of coding is that it provides an assessment of features of the interaction that could not be determined by coding the individual behavior of participants. Coding individual behavior does not, of course, preclude coding be havior related to the interaction (and vice versa). The decision about coding indi viduals or an interaction reminds researchers that they need to think carefully about their assumptions regarding the impact of participants' behavior on one another and the way aspects of interaction may emerge from the contributions of multiple participants. Coding Practices We have been discussing rather broad choices that reflect decisions about different ways to design a study or different ways to look at the interaction. We now turn to issues that are more narrowly focused and that reflect specific techniques used in examining nonverbal communication. We discuss these to provide a sense of the reasoning behind each coding practice. The use of each practice is more fully ex plained in other chapters in this volume. Selecting a Coding Unit. Even after researchers have made many important decisions about how to collect and treat their data, they are faced with an important RESEARCHER CHOICES AND PRACTICES 15 choice regarding what Bakeman and Gottman (1997) call the coding unit. The cod ing unit refers to the decision researchers make about when to code within the inter action and the length of time observation lasts. This decision is related to assessing either a behavioral event or assessing intervals of time within the interaction (Bakeman & Gottman, 1997; also see Bakeman, this volume). In attempting to de termine whether to use event or interval coding, researchers need to consider a number of issues, such as the level of accuracy of the data they desire, the complexity of their coding scheme, and how frequently the behavior occurs (Bakeman & Gottman, 1997). Event Coding. A researcher who decides to code events within an interaction (i.e., a temper tantrum, an instance of parallel play, or a self-disclosure), may be in terested in examining several things, including how frequently a behavioral event occurs, the order in which different events happen, or the duration of an event (Bakeman & Gottman, 1997; Cappella, this volume). These assessments of events reflect different coding strategies. The first type of coding strategy measures how frequently an enacted behavior occurs. In this type of event coding, the researcher looks for the behavioral event and simply records whether it occurred or not. For instance, if a researcher is interested in the number of hugs enacted between young siblings, any time one sibling hugged the other, it would be recorded. A second type of event coding records the enactment of multiple behavioral events and the sequence in which they occurred. For instance, a researcher may be interested in not only hugging between siblings, but also other expressions of car ing, such as kissing, patting, or holding hands. This type of event coding would not only record the specific expression of care, but in what order the behaviors oc curred. This is similar to Bakeman and Gottman's (1997) sequential analysis. Finally, a third type of event coding records the length of time the behavioral event occurred—in other words, the duration of time participants devote to enact ing a specific behavioral event. Again, using the previous example, a researcher may be interested in the length of time toddler siblings hold hands or how long they hug each other (Bakeman& Gottman, 1997). The length of time a behavior is enacted is what is most important here. Interval Coding. When researchers use the term interval coding, they are de scribing typically those portions or segments of an interaction from which they will sample and code nonverbal behaviors. In this research strategy, the entire in teraction is divided into predetermined brief units of time or intervals in which coders/raters observe and code/rate behaviors (Bakeman & Gottman, 1997). In other words, researchers restrict the sampling of nonverbal behaviors to segments or slices within the communication interaction that are of specific interest (Ambady & Conner, 1999; Ambady & Gray, 2002). For instance, a researcher may be interested in coding the immediacy behaviors of close relational partners dur ing indirect self-disclosures to one another. The researcher will specifically select 16 WHITE AND SARGENT intervals of disclosure within the interaction and code only those slices of behav ior (e.g., Sargent, 2000). Once researchers elect to code intervals within the interaction, depending on the purpose of the study, they must decide on the number of intervals to use within the interaction and the length of each interval. In making these decisions, researchers need to consider the frequency with which the behavior is likely to oc cur, how long it takes participants to enact a behavior, and the number of behav iors to be coded/rated in each interval. If the particular behavior of interest occurs only occasionally during the interaction, the researcher may consider coding all of the intervals in which the behavior is found, but if the behavior is seen repeatedly during the interaction, the researcher may elect to code fewer intervals (Scherer & Ekman, 1982). Once this choice is made, the researcher decides whether the placement of the intervals should be defined by the communication event or positioned in the in teraction by fixed time intervals, such as 1 minute of coding for every 5 minutes of interaction. (Bakeman & Gottman, 1997, called this systematic observation). For example, if the focus of a study is on the affection cues displayed as parents drop off children at preschool, the event itself will define the coding interval nat urally. By making this choice, the researcher allows the event to determine the placement of the coding interval. In contrast, when the particular behavior of interest occurs regularly within an interaction, there are more opportunities to place coding intervals throughout the interaction. For instance, if a researcher is interested in examining politeness cues exhibited during a baby shower, the re searcher can more easily select fixed points in the interaction to place the inter vals. Depending on the frequency of the politeness cues, a researcher could decide to code 2 minute intervals within every 5 minutes of interaction. An ad vantage of fixed intervals in coding sequence is that the behavior is observed consistently throughout the interaction, allowing for a more accurate reflection of how it is represented in the data. Researchers can also be somewhat less "systematic" in their placement of cod ing intervals. For example, behaviors could be coded for 2 minutes at the begin ning of the first 10 minutes of an interaction, 2 minutes in the middle of the second 10 minutes of an interaction, and 2 minutes at the end of a third 10 min utes of interaction, with this system repeated until the interaction has ended. For example, Guerrero and Andersen (1994) coded the initiation of touching between individuals waiting in movie theater and zoo lines. They defined their coding in terval as the first 2 minutes of interaction between couples as soon as they started standing at the end of the line. There are clear and important differences between event and interval coding strategies. One distinction between the two is that event coding may be more "ob jective" than interval coding. According to Bakeman and Gottman (1997), event coding can result in "more accurate data" (p. 38) by requiring less observer infer ence. In event coding, observers are not required to determine the degree to which a RESEARCHER CHOICES AND PRACTICES 17 behavior is enacted or the intention or function of a specific behavior. Also, event coding may allow researchers to observe the enactment of more complete patterns of behaviors, providing a clearer relationship between a behavior and its function (Scherer & Ekman, 1982). Despite the strengths of event coding, interval coding has advantages as well. In terval coding allows researchers to examine the occurrence of microbehaviors, such as eye movements, smiles, or fluctuations in eyebrows. This strategy also provides researchers the ability to isolate and examine extremely small, but meaningful, changes in a single behavior as well as precise differences between different behav iors. Additionally, if a researcher has videotaped participants, it is often easier at a later time to reexamine and code for more global impressions or meanings. Choosing to event or interval code is an important decision and one that will signifi cantly impact the outcome of the study. As such, researchers need to carefully con sider the purpose of their study as well as the strengths and weaknesses of both approaches before settling on one. Combined or Separate Assessments of Verbal/Nonverbal Channels Another choice that researchers face concerns whether to examine verbal and non verbal behavior together or whether to attempt to determine their separate effects by modifying one of these channels during the research process or coding process. Given that Noller (this volume) provides a nice discussion of the standard content method, which seeks to control the verbal channel so that the effects of nonverbal behavior can be more clearly understood, we focus here on the choices researchers make when they code communication behavior that includes both verbal and non verbal interaction. Early research (e.g., Mehrabian & Ferris, 1967) sought to reveal the power and importance of nonverbal behavior; typically, it involved coding that restricted ac cess to the verbal content (by having coders/raters view interaction without audio) or used data that had been altered (such as speech that had been content filtered) to reveal the impact of paralinguistic cues (Krauss, 1981). More recently, however, re searchers have acknowledged that the meaning of nonverbal behavior is often tied to the verbal content of interaction, and so coders/raters have been asked to use both verbal content and nonverbal cues to make assessments of the interaction (e.g., Ebesu Hubbard, 2000; Guerrero, 1996). Additionally, both sets of cues are used in some self-reports (e.g., Floyd & Mikkleson, this volume). The key advantages of coding based on access to both verbal communication and nonverbal behavior are related to (a) the isomorphism of the coding task with real interaction, and (b) contextualization of meaning. Observers who make ratings based on access to all aspects of the interaction are likely making judgments that are similar to the types of assessments made in day-to-day interaction. Moreover, as sessments of any specific behavior will reflect the meaning of the behavior within the stream of other behaviors (verbal and nonverbal) that are occurring. 18 WHITE AND SARGENT The decision to combine verbal and nonverbal behavior in assessments of an in teraction is particularly useful when researchers want to assess aspects of interac tion, such as involvement, that are affected by many behaviors or when researchers want to examine aspects of the interaction that are produced jointly by participants. The cost of coding in this way is that researchers cannot make claims about the in fluence of a particular behavioral channel on the interaction, and researchers are not able to determine how interpretation of the interaction would vary if certain channels of communication are modified. The decision to limit coder access, or to modify a channel of interaction (e.g., standard content method, Noller, this volume), is usually made with one of the fol lowing goals in mind. Modifying a channel of communication (for instance, mak ing the verbal content of interaction ambiguous) allows researchers to assess communicator competence and to determine the impact of a particular channel of communication on interaction outcomes. Coding that limits access to a particular channel (for instance, asking coders to make ratings without access to the audio portion of an interaction) may produce more fine-tuned assessments of particular behaviors. Additionally, when different channels are coded separately, it may be possible to identify inconsistencies between channels that would not be evident to coders who have access to all channels and to find a way to reconcile them or access the combined effects of different channels. Having noted these opportunities, however, this type of coding also has the po tential to produce highly artificial results because the actual or full meaning of be havior in interaction is likely to be the result of combined verbal and nonverbal channels (Burgoon, Buller, Hale, & deTurck, 1984). Additionally, when channels are coded separately and researchers seek to combine them, it is unclear how this should be done: As we mentioned in our discussion of level of measurement, it seems unlikely that simply adding these codes together is an appropriate way to sum their effects. It should be clear that the practice of combining or separating is tied to differ ences in the types of issues researchers want to examine. One approach seeks to ex amine how different channels of behavior influence meaning; the other asks questions about what meaning is conveyed when both verbal and nonverbal aspects of interaction are examined. It is important to notice that these practices are related to the researcher's view of behavior; it is easy to believe that coding practices such are just that: practices. But, practices have at their roots assumptions about how in teraction is enacted, understood, and managed. These assumptions are also evident when researchers choose use physiological measures to assess nonverbal behavior. Physiological Measures At times, researchers may consider measuring the physiological responses of par ticipants. By examining physical responses such as heart rates, skin conductivity, body temperature, and hormonal fluctuations, researchers may be able to tap RESEARCHER CHOICES AND PRACTICES 19 into the intensity levels of some emotional and social states that cannot necessar ily be detected by observer coders or by participants themselves with self-report measures. Because these emotional or social states may not be manifested overtly, physiological measures could yield additional insights into the nature of certain behaviors. Often, researchers correlate physiological responses with communication be haviors or perceptions. For instance, in a study by Gottman and Levenson (2002), couples were connected to devices that measured heartbeat, skin conductance level, body movement, and finger pulse. Once connected, couples were asked to discuss three different topics specifically designed to elicit physiological reac tions. While conversing, couple members' physiological responses to the conver sation they were having with their partner were measured, allowing researchers to examine how the different conversations influenced each partner's physiological states. In another example, Tusing and Dillard (2000) tested the effects of varying vocal cues on perceptions of dominance and influence. They found that vocal am plitude was positively associated with dominance judgments, but speech rate was negatively correlated with judgments. In later chapters, both Kinney (this vol ume) and Tusing (this volume) discuss further the merits of physiological coding and measures and argue that physiological assessments have been greatly underutilized in nonverbal research. CONCLUSION The study of nonverbal communication can be simultaneously one of the most re warding and most challenging of research endeavors. One needs only to examine a small portion of the nonverbal research literature—including what is included in this volume—to see the large number of decisions that go into the design of a study. It cannot be emphasized enough, though, that the choices researchers make will influence the study—for good or ill. Despite the numerous decisions that are required, we hope that researchers will see these choices as opportunities to con sider how nonverbal behavior can best be studied and understood. REFERENCES Ambady, N., & Conner, B. (1999). Accuracy of judgments of sexual orientation from thin slices of be havior. Journal of Personality and Social Psychology, 77, 538—548. Ambady, N., & Gray, H. M. (2002). On being sad and mistaken: Mood effects on the accuracy of thinslice judgments. Journal of Personality and Social Psychology, 83, 947—962. Bakeman, R., & Gottman, }. M. (1997). Observing interaction: An introduction to sequential analysis. Cambridge, UK: Cambridge University Press. Bavelas, J. B. (1995). Quantitative versus qualitative? In W. Leeds-Hurwitz (Ed.), Social approaches to communication (pp. 49-62). New York: Guilford. Bernieri, F. J., & Rosenthal, R. (1991). Interpersonal coordination: Behavioral matching and interac tional synchrony. In R. S. Feldman & B. Rime (Eds.), Fundamentals of nonverbal behavior (pp. 401-432). Cambridge: Cambridge University Press. 20 WHITE AND SARGENT Burgoon, J. K., & Baesler, E. J. (1991). Choosing between micro and macro nonverbal measurement: Application to selected vocalic and kinesic nonverbal indices. Journal of Nonverbal Behavior, 15, 57-78. Burgoon, J. K., Buller, D. B., Hale, J., & deTurck, M. (1984). Relational messages associated with nonver bal behavior. Human Communication Research, 10, 351—378. Burgoon, J. K., Buller, D. B., White, C. H., Afifi, W., & Buslig, A. (1999). The role of conversational involve ment in deceptive interpersonal interactions. Personality and Social Psychology Bulletin, 25,669-706. Burgoon, J. K., Olney, C. A., & Coker, R. A. (1987). The effects of communicator characteristics on pat terns of reciprocity and compensation. Journal of Nonverbal Behavior, 11, 146-165. Cappella, J. N. (1981). Mutual influence in expressive behavior: Adult-adult and infant-adult dyadic in teraction. Psychological Bulletin, 89,101-132. Cappella, J. N. (1991). Mutual adaptation and relativity of measurement. In B. M. Montgomery & S. Duck (Eds.), Studying interpersonal interaction (pp. 103-117). New York: Guilford. Ebesu Hubbard, A. S. (2000). Interpersonal coordination in interactions: Evaluations and social skills. Communication Research Reports, 17, 95—104. Gottman, J. M., & Levenson, R. W. (1992). Marital processes predictive of later dissolution: Behavior, physiology, and health. Journal of Personality and Social Psychology, 63, 221—233. Gottman, J. M., & Levenson, R. W. (2002). A two-factor model for predicting when a couple will di vorce: Exploratory analyses using 14-year longitudinal data. Family Process, 41, 83-96. Guerrero, L. K. (1996). Attachment style difference in intimacy and involvement: A test of the four-category model. Communication Monographs, 63, 269-292. Guerrero, L. K., & Andersen, P. A. (1994). Patterns of matching and initiation: Touch behavior and touch avoidance across romantic relationship stages. Journal of Nonverbal Behavior, 18, 137-153. Harvey, J. H., Hendrick, S. S., & Tucker, K. (1988). Self-report methods in studying personal relation ships. In S. W. Duck (Ed.), Handbook of personal relationships (pp. 99-113). New York: John Wiley & Sons. Hecht, M. L., & Guerrero, L. K. (1999). Perspectives on nonverbal research methods. In L. K. Guerrero, J. A. DeVito, & M. L. Hecht (Eds.), The nonverbal communication reader: Classic and contemporary readings (2nd ed., pp. 24-41). Prospect Heights, IL: Waveland Press. Heslin, R., & Boss, D. (1980). Nonverbal intimacy in airport arrival and departure. Personality and Social Psychology Bulletin, 6, 248-252. Jones, S. E., & LeBaron, C. D. (2002). Research on the relationship between verbal and nonverbal com munication: Emerging interactions. Journal of Communication, 52, 499-521. Jones, S. E., & Yarbrough, A. E. (1985). A naturalistic study of the meanings of touch. Communication Monographs, 52, 19-56. Krauss, R. M. (1981). Verbal, vocal, and visible factors in judgments of another's affect. Journal of Per sonality and Social Psychology, 40, 312-320. Manusov, V. (1995). Reacting to changes in nonverbal behaviors: Relational satisfaction and adaptation patterns in romantic dyads. Human Communication Research, 21, 456-477. Manusov, V., & Trees, A. R. (2002). "Are you kidding me?": The role of nonverbal cues in the verbal ac counting process. Journal of Communication, 52, 640-656. Mehrabian, A., & Ferris, S. R. (1967). Inferences of studies from nonverbal communication in two chan nels. Journal of Consulting Psychology, 31, 248-252. Metts, S., Sprecher, S., & Cupach, W. R. (1991). Retrospective self-reports. In B. M. Montgomery & S. Duck (Eds.), Studying interpersonal interaction (pp. 162-178). New York: Guilford. Palmer, M. T., & Simmons, K. B. (1995). Communicating intentions through nonverbal behaviors: Conscious and nonconscious encoding of liking. Human Communication Research, 22, 128-160. Patterson, M. L., Webb, A., & Schwartz, W. (2002). Passing encounters: Patterns of recognition and avoidance in pedestrians. Basic and Applied Social Psychology, 24, 57-66. Riggio, R., E., & Riggio H. R. (2001, July). Using self-report methods to measurenonverbal communication skill. Paper presented to the International Network on Personal Relationships, Prescott, AZ. Sargent, J. (2000). Nonverbal boundaries: An examination of the immediacy behaviors in explicit and im plicit disclosures. Unpublished doctoral dissertation, The University of Nebraska, Lincoln. Scherer, K. R., & Ekman, P. (1982). Methodological issues in studying nonverbal behavior. In K. R. Scherer & P. Ekman (Eds.), Handbook of methods in nonverbal behavior research (pp. 1-44). Cam bridge, UK: Cambridge University Press. RESEARCHER CHOICES AND PRACTICES 21 Segrin, C. (1992). Specifying the nature of social skill deficits associated with depression. Human Com munication Research, 19, 89-123. Tusing, K. J., & Dillard, J. P. (2000). The sounds of dominance: Vocal precursors of perceived domi nance during interpersonal influence. Human Communication Research, 26, 148-171. White, C. H., & Burgoon, J. K. (2001). Adaptation and communicative design: Patterns of interaction in truthful and deceptive conversations. Human Communication Research, 25, 9-37. Woolfolk, A. E. (1981). The eye of the beholder: Methodological considerations when observers assess nonverbal communication. Journal of Nonverbal Behavior, 5, 199-204. This page intentionally left blank II. NONVERBAL MEASURES This page intentionally left blank The Social Skills Inventory (SSI): Measuring Nonverbal and Social Skills Ronald E. Riggio Claremont McKenna College INTRODUCTION The measurement of individual differences in nonverbal and social skills is, in many ways, rooted in research on intelligence, particularly the early work by Thorndike (1920) and others (Moss, Hunt, Omwake, & Ronning, 1927) in mea suring social intelligence (i.e., the ability to understand and manage people and to act wisely in human relations). Researchers of social intelligence (e.g., Chapin, 1942; O'Sullivan, 1983; O'Sullivan & Guilford, 1975) realized that the ability to read or decode the feelings and intentions of others, and to decode and understand social interactions and social settings, were critical components of social intelli gence. Guilford (1967), in his structure of intellect model and in the development of measures of multiple intelligences, included several nonverbal tests of social intel ligence (O'Sullivan & Guilford, 1976). Whereas the "intelligence" line of research represents attempts to define and measure individual differences in social interaction skills, it was the pioneering research of Robert Rosenthal and his colleagues in measuring nonverbal decoding skills (see Rosenthal, 1979) that led to the first performance-based measures of nonverbal ability. One such measure, the Profile of Nonverbal Sensitivity (PONS; Rosenthal, Hall, DiMatteo, Rogers, & Archer, 1979), was used widely in research on individual differences in nonverbal skill. Another measure of both verbal and nonverbal decoding skill—one that is closely linked to notions of social intelligence—is Archer and Akert's (1977) work with the Social Interpretations Task (SIT), a performance-based measure of ability to decode and interpret social situ ations. The SIT evolved into the updated Interpersonal Perception Task (IPT; Costanzo & Archer, 1993). 25 26 RIGGIO This chapter, however, presents The Social Skills Inventory (SSI; Riggio, 1986, 1989; Riggio & Carney, 2003). The SSI is a 90-item, self-report instrument that measures basic skill in nonverbal/emotional communication as well as verbal/social skills that are related to social competence. The inventory was derived from multidisciplinary research focusing on the development of nonverbal, emotional, and interpersonal skills. This research included groundbreaking work for measur ing nonverbal encoding and decoding skills (Archer & Akert, 1977; Friedman, 1979; Rosenthal et al, 1979; Zuckerman & Larrance, 1979), Snyder's (1974, 1987) re search on assessing skill in impression management/self-monitoring, and Guilford's and O'Sullivan's (O'Sullivan, 1983; O'Sullivan & Guilford, 1975) schol arship for measuring social intelligence. THE SOCIAL SKILLS INVENTORY Drawing on a basic communication model, the SSI framework breaks down basic communication skills into three types: skill in encoding, or expressivity, skill in de coding, or sensitivity, and skill in regulation, or control, over communication. These three basic communication skills operate in two domains: the nonverbal, or emo tional, and the verbal, or social. Table 1 provides an illustration of the SSI model and brief description of each of the six SSI subscales along with sample items from the SSI. Information on obtaining the SSI is provided at the end of the chapter. The three SSI emotional skill subscales are tied to nonverbal communication skill most directly. Emotional Expressivity (EE) is a measure of emotional expres siveness, related closely to other self-report measures of the same construct (see Riggio & Riggio, this volume). The Emotional Sensitivity (ES) scale is a self-report measure of nonverbal decoding skill. The SSI Emotional Control (EC) scale is an as sessment of ability to monitor and control one's own emotional expressions. It is also linked theoretically to posed nonverbal/emotional encoding, because emo tionally controlled individuals need to regulate their facial expressions in order to mask felt emotions, either by appearing stoic, or by using a conflicting emotional state as a mask (e.g., putting on a happy face to cover felt sadness or anger). Research suggests, however, that it is a combination of Emotional Control and Emotional Expressivity that contributes to posed emotional sending/encoding ability (Tucker & Riggio, 1988). Summed, these three emotional skills can form an index of general emotional/ nonverbal skill competence. Indeed, the recent surge of interest in the construct of emotional intelligence has much in common with the SSI emotional skill framework. In fact, Mayer and Salovey's (1997) ability model of emotional intel ligence includes abilities to identify/decode others' emotions, express/encode one's own emotions accurately, and monitor and regulate felt emotional states. The Social Skills Inventory assesses these three core dimensions of emotional in telligence. Moreover, in psychometric terms, the SSI compares favorably to some of the existing self-report measures of emotional intelligence (see Ciarrochi, TABLE 1 The Social Skills Inventory (SSI) Framework, Scale Definitions, and Sample Items from the SSI Nonverbal/Emotional Domain (Emotional Skills) [related to Emotional Intelligence] Verbal/Social Domain (Social Skills) [related to Social Intelligence] Emotional Expressivity (EE) Social Expressivity (SE) Skill in nonverbal encoding. Ability to accurately express felt emotional states. Skill in verbal encoding and ability to engage others in social interaction. Associated with verbal fluency. I am able to liven up a dull party. I have been told that I have expressive eyes. Emotional Sensitivity (ES) Skill in nonverbal decoding. Being attentive to subtle emotional cues; being empathic. I sometimes cry at sad movies. I am often told that I am a sensitive, understanding person. Emotional Control (EC) Skill in regulating and controlling emotional expressions. Hiding felt emotions behind an emotional "mask." • When telling a story, I usually use a lot of gestures to help get the point across. • I usually take the initiative to introduce myself to strangers. Social Sensitivity (SS) Skill in verbal decoding. Sensitivity to and understanding of norms governing appropriate social behavior. Ability to decode social situations. • I'm generally concerned about the impression I'm making on others. • Sometimes I think that I take things other people say to me too personally. Social Control (SC) Skill in social role-playing and social self-presentation. Social adeptness and tact."Savoir-faire." I am easily able to make myself look happy one minute and sad the next. • I am usually very good at leading group discussions. I am very good at maintaining a calm exterior even if I am upset. • I can easily adjust to being in just about any social situation. 28 RIGGIO Chan, Caputi, & Roberts, 2001), such as those developed by Bar-On (1997) and Schutteetal. (1998). Although there may appear to be little direct connection between research in nonverbal communication and the social skill subscales of the SSI, the skill of So cial Expressivity (SE) involves verbal speaking skill and the ability to engage oth ers in social interaction. This is, in many ways, a complement to Emotional Expressivity and suggests skill in initiating interpersonal interactions. The SSI di mension of Social Sensitivity (SS), although primarily related to verbal decoding skill (i.e., listening), also involves ability to decode and interpret social situations, a skill that is very important in decoding tasks as represented in measures such as the Interpersonal Perception Task (IPT; Costanzo & Archer, 1993). In fact, the SSI-SS scale is significantly and positively correlated with scores on the IPT (Riggio & Carney, 2003). The skill of Social Control (SC) is linked to ability to manage impressions—similar to the construct of self-monitoring—but more recently equated with the con struct of savoir-faire: the ability to know how to act in social situations (Eaton, Funder, & Riggio, 2002). Taken together, these three social skill dimensions—SE, SS, and SC—can be considered an index of social intelligence (Riggio, Messamer, & Throckmorton, 1991). Thus, researchers of nonverbal behavior and interpersonal processes may be interested not only in the nonverbal skill scales of the SSI but also in the entire scale. Psychometric Properties of the SSI Scale Reliability. The SSI scales have shown good test-retest reliability, rang ing from .81 to .96 for a 2-two week interval (N= 40). Cronbach's alpha coefficients ranging from .65 to .88 were obtained from a sample of 549 employed adults re cruited from a number of work organizations across the United States. Alpha coeffi cients from a group of 389 undergraduate students ranged from .64 to .89. These findings suggest that the various SSI scales have acceptable to good internal consis tency (see Riggio & Carney, 2003). Scale Intercorrelotions and Sex Differences. The SSI subscales are posi tively correlated, with some notable exceptions. For instance, Emotional Expres sivity and Emotional Control are somewhat negatively correlated, as are Social Sensitivity and Social Control. The actual relationships among the various SSI di mensions are quite complex (see Riggio & Carney, 2003). Consistent with expectations, however, women tend to be more expressive and sensitive than are men, with women obtaining typically higher scores on Emotional Expressivity, Social Expressivity, Emotional Sensitivity, and Social Sensitivity. Men tend to score significantly higher on Emotional Control, with no significant differ ences on Social Control. These sex differences are consistent for both samples of adults and college students (Riggio & Carney, 2003). SOCIAL SKILLS INVENTORY 29 Factor Structure. Confirmatory factor analyses have supported the basic subscale structure of the SSI. The factor structure has held up both in tests of the SSI in English and in other cultures/languages, such as Italian (Galeazzi, Franceschina, & Holmes, 2002; Riggio, 1986). Convergent and Discriminant Validity. The SSI scales have demonstrated very good convergent validity with other self-report measures of nonverbal skill-related constructs, such as measures of emotional expressiveness, emo tional reactivity, and emotional empathy. There is also evidence from a study using undergraduate students that SSI Emotional Sensitivity correlates signifi cantly with performance-based measures of emotional decoding skill (see Riggio & Carney, 2003) such as the Profile of Nonverbal Sensitivity (PONS; Rosenthal et al., 1979), the Diagnostic Analysis of Nonverbal Accuracy (DANVA; Nowicki & Duke, 1994; Duke & Nowicki, this volume), and an emo tional decoding subscale of the Multifactor Emotional Intelligence Scale (MEIS-Pictures; Mayer, Salovey, & Caruso, 1997). A study of undergraduate students conducted by Tucker and Riggio (1988) ex plored the relationships between the SSI scales and both posed and spontaneous emotional encoding, the latter using the Buck slide-viewing paradigm (see Buck's chapter, this volume). As predicted, SSI-Emotional Expressivity was related to both posed and spontaneous emotional encoding. Emotional Control, as one might sus pect, was significantly negatively correlated with spontaneous emotional expres sion while viewing emotion-eliciting slides, but was, contrary to prediction, unrelated to posed emotional encoding (although a combination of EE and EC was significantly positively correlated to posed sending). SSI-Social Control was, how ever, significantly positively related to posed sending, further suggesting that SC is an important social acting skill. An important concern to many researchers is the use of self-report measures to assess nonverbal skill, with critics suggesting that individuals do not have the insight or the unbiased perspective to make accurate assessments of their non verbal communication skills. However, validity evidence demonstrating signifi cant relationships between self-report measures of nonverbal skill and performance-based assessments of skill suggest that the self-report methodol ogy is valuable, useful, and a cost-effective alternative to more time-consuming and costly performance measures (see Riggio& Riggio, 2001). In addition to ev idence of convergent validity, there is good evidence demonstrating the dis criminant validity of the SSI scales. For example, although Emotional and Social Expressivity are theoretically and empirically linked to personality constructs such as extraversion, evidence suggests that they are distinct constructs (see Friedman, 1983; Riggio & Riggio, 2002). Additionally, there has been little con cern about socially desirable responding for the SSI scales (see Riggio, 1986; Riggio & Carney, 2003). 30 RIGGIO The SSI and Research in Nonverbal Communication The SSI has been used widely in research on nonverbal behavior. One line of re search examined the impact of emotional expressiveness and global nonverbal/social skills on impressions made in initial encounters and in initial attractiveness. Consistent with past research (e.g., Friedman, Riggio, & Casella, 1988; Riggio & Friedman, 1986; Riggio & Woll, 1984; Sabatelli & Rubin, 1986), emotionally and socially expressive persons (high scores on EE and SE), and persons with high over all SSI scores, were rated as more likable and attractive than persons scoring low on these SSI dimensions in initial encounters, even after controlling for static cues of physical attractiveness (Riggio, 1986; Riggio, Widaman, Tucker, & Salinas, 1991). These studies included undergraduate students as well as adult members of a videodating organization (Riggio & Woll, 1984). Nonverbal social skills have also been investigated in the context of deception by undergraduate students. An important finding is that nonverbally skilled commu nicators, as measured by the SSI, are more successful deceivers primarily because they have a more "honest" overall demeanor-emitting cues that are stereotypically associated with truthfulness-than are persons lacking these important nonverbal skills (Riggio, Tucker, & Throckmorton, 1987; Riggio, Tucker, & Widaman, 1987). Likewise, nonverbal social skills, as represented by the SSI framework, have been shown to be important in relationship formation and maintenance and in the abil ity to garner social support from these relationships to cope effectively with every day stress (Riggio, 1992; Riggio & Zimmerman, 1991). Moreover, the SSI has important implications for evaluating the quality of communication in marriages (both young adult marriages and marriages of more than 50 years) and interper sonal relationships (see Riggio & Carney, 2003). Most recently, the SSI has been used to study the nonverbal and social behavior of leaders in small groups and persons in managerial or business leadership posi tions. For instance, the SSI was found to correlate positively and significantly with observer ratings of participants' communication skills in a managerial assessment center conducted for students in a university school of business (Riggio, Aguirre, Mayes, Belloli, & Kubiak, 1997). In addition, scores on the SSI predicted group members' satisfaction with their leaders in simulated work groups, and were related to followers' ratings of leader effectiveness in the fire service (Riggio, Riggio, Salinas, & Cole, 2003). An exciting line of research explores the role that nonverbal social skills play in contributing to a leader's charisma and extends some of the ideas pre sented in Friedman et al. (1988) and Riggio (1987). In an additional study, the SSI predicted performance evaluations of hospice workers (Riggio & Taylor, 2000), and SSI scores correlated negatively with indices of loneliness, depression, and social maladjustment in student populations (Riggio, Watring, & Throckmorton, 1993; Segrin& Flora, 2000). An important line of research investigates the role of nonverbal skills and non verbal skill imbalances (e.g., wide variations such as high scores on Emotional Con SOCIAL SKILLS INVENTORY 31 trol coupled with very low scores on Emotional Expressivity) in predicting psycho pathology in outpatients from mental hospitals (Perez & Riggio, 2003; Perez, Riggio, & Kopelowicz, 2003). Although wide discrepancies in scores on the various subscales of the Social Skills Inventory are hypothesized to be indicative of social skill "imbalances" that may suggest an overall social skill deficit, more work needs to be done looking at how different combinations of high and low scores on the SSI subscales relate to social performance and psychosocial adjustment. FUTURE DIRECTIONS Future research on the SSI can investigate the role of nonverbal behavior skills in leadership and in contributing to effective management in work organizations, as well as additional research on how specific nonverbal and social skills impact rela tionship formation and relationship quality. Besides its use as a research tool for assessing nonverbal and social skills, the SSI can also be used to get baseline assess ments of possession of communication skills or to measure the development of skills over time. This can be important for individual development or for larger scale nonverbal/social skill development and training programs. The Social Skills Inventory is available to researchers for a nominal fee at www.mindgarden.com REFERENCES Archer, D., & Akert, R. M. (1977). Words and everything else: Verbal and nonverbal cues in social inter pretation. Journal of Personality and Social Psychology, 35, 443-449. Bar-On, R. (1997). Bar-On Emotional Quotient Inventory: Technical manual. Toronto, ON: Multi-Health Systems. Chapin, F. S. (1942). Preliminary standardization of a social insight scale. American Sociological Review, 7,214-225. Ciarrochi, J., Chan, A., Caputi, P., & Roberts, R. (2001). Measuring emotional intelligence. In J. Ciarrochi, J. P. Forgas, & J. D. Mayer (Eds.), Emotional intelligence in everyday life: A scientific inquiry (pp. 25-45). Philadelphia, PA: Psychology Press. Costanzo, M. A., & Archer, D. (1993). Interpreting the expressive behavior of others: The Interpersonal Perception Task (IPT). Journal of Nonverbal Behavior, 13, 225-245. Eaton, L. G., Funder, D. C, & Riggio, R. E. (2003). Skill in social role-playing: The essence of savior faire. Manuscript submitted for publication. Friedman, H. S. (1979). The concept of skill in nonverbal communication: Implications for understand ing social interaction. In R. Rosenthal (Ed.), Skill in nonverbal communication (pp. 2-27). Cam bridge, MA: Oelgeschlager, Gunn, & Hain. Friedman, H. S. (1983). On shutting one's eyes to face validity. Psychological Bulletin, 94, 185-187. Friedman, H. S., Riggio, R. E., & Casella, D. F. (1988). Nonverbal skill, personal charisma, and initial at traction. Personality and Social Psychology Bulletin, 14, 203-211. Galeazzi, A., Franceschina, E., & Holmes, G. R. (2002). Factor analysis of Social Skills Inventory re sponses of Italians and Americans. Psychological Reports, 90, 1115-1121. Guilford, J. P. (1967). The nature of human intelligence. New York: McGraw-Hill. Mayer, J. D., & Salovey, P. (1997). What is emotional intelligence? In P. Salovey & J. D. Sluyter (Eds.), Emotional development and emotional intelligence (pp. 3-34). New York: Basic Books. Mayer, J. D., Salovey, P., & Caruso, D. (1997).Multifactor Emotional Intelligence Test (MEIS). Needham, MA. Moss, F. A., Hunt, T., Omwake, K. T., & Ronning, M. M. (1927).Social intelligence test. Washington, DC: George Washington University. 32 RIGGIO Nowicki, S., & Duke, M. P. (1994). Individual differences in the nonverbal communication of affect: The Diagnostic Analysis of Nonverbal Accuracy Scale. Journal of Nonverbal Behavior, 18, 9-35. O'Sullivan, M., & Guilford, J. P. (1976). Four factor tests of social intelligence (behavioral cognition): Manual of instructions and interpretations. Orange, CA: Sheridan Psychological Services. O'Sullivan, M. (1983). Measuring individual differences. In J. M. Wiemann & R. P. Harrison (Eds.), Nonverbal interaction (pp. 243-270). Beverly Hills, CA: Sage. O'Sullivan, M., & Guilford, J. P. (1975). Six factors of behavioral cognition: Understanding other peo ple. Journal of Educational Measurement, 12, 255-271. Perez, J. E., & Riggio, R. E. (2003). Nonverbal social skills and psychopathology. In P. Philippot, E. J. Coats, & R. S. Feldman (Eds.), Nonverbal behavior in clinical settings (pp. 17-44). New York: Oxford University Press. Perez, J. E., Riggio, R. E., & Kopelowicz, A. (2003). Social skill imbalances as indicators of psychopathology: An exploratory investigation. Manuscript submitted for publication. Riggio, H. R., & Riggio, R. E. (2002). Emotional expressiveness, extraversion, and neuroticism: A meta-analysis. Journal of Nonverbal Behavior, 26, Riggio, R. E. (1986). Assessment of basic social skills. Journal of Personality and Social Psychology, 51, 649-660. Riggio, R. E. (1987). The charisma quotient. New York: Dodd, Mead. Riggio, R. E. (1989). Manual for the Social Skills Inventory: Research edition. Palo Alto, CA: Consulting Psychologists Press. Riggio, R. E. (1992). Social interaction skills and nonverbal behavior. In R. S. Feldman (Ed.), Appli cations of nonverbal behavioral theories and research (pp. 3-30). Hillsdale, NJ: Lawrence Erlbaum Associates. Riggio, R. E., Aguirre, M., Mayes, B. T., Belloli, C, & Kubiak, C. (1997). The use of assessment center methods for student outcome assessment. Journal of Social Behavior and Personality, 12, 273-288. Riggio, R. E., & Carney, D. R. (2003). Social Skills Inventory manual: Second edition. Redwood City, CA: MindGarden. Riggio, R. E., & Friedman, H. S. (1986). Impression formation: The role of expressive behavior. Journal of Personality and Social Psychology, 50, 421-427. Riggio, R. E., Messamer, J., & Throckmorton, B. (1991). Social and academic intelligence: Conceptually distinct but overlapping constructs. Personality and Individual Differences, 12, 695—702. Riggio, R. E., & Riggio, H. R. (2001). Self-report measurement of interpersonal sensitivity. In J. A. Hall & F. J. Bernieri (Eds.), Interpersonal sensitivity: Theory and measurement (pp. 127-142). Mahwah, NJ: Lawrence Erlbaum Associates. Riggio, R. E., Riggio, H. R., Salinas, C., & Cole, E. J. (2003). The role of social and emotional communication skills in leader emergence and effectiveness. Group Dynamics: Theory, Research, and Practice, 7,83-103. Riggio, R. E., & Taylor, S. J. (2000). Personality and communication skills as predictors of hospice nurse performance. Journal of Business and Psychology, 15, 347-355. Riggio, R. E., Tucker, J., & Throckmorton, B. (1987). Social skills and deception ability. Personality and Social Psychology Bulletin, 13, 568-577. Riggio, R. E., Tucker, J., & Widaman, K. F. (1987). Verbal and nonverbal cues as mediators of deception ability. Journal of Nonverbal Behavior, 11, 126—145. Riggio, R. E., Watring, K., & Throckmorton, B. (1993). Social skills, social support, and psychosocial adjustment. Personality and Individual Differences, 15, 275—280. Riggio, R. E., Widaman, K. F., Tucker, J. S., & Salinas, C. (1991). Beauty is more than skin deep: Compo nents of attractiveness. Basic and Applied Social Psychology, 12, 423-439. Riggio, R. E., & Woll, S. B. (1984). The role of nonverbal cues and physical attractiveness in the selection of dating partners. Journal of Social and Personal Relationships, 1, 347-357. Riggio, R. E., & Zimmerman, J. A. (1991). Social skills and interpersonal relationships: Influences on so cial support and support seeking. In W. H. Jones & D. Perlman (Eds.), Advances in personal relation ships (Vol. 2, pp. 133-155). London: Jessica Kingsley Press. Rosenthal, R. (Ed.). (1979). Skill in nonverbal communication. Cambridge, MA: Oelgeschlager, Gunn, & Hain. Rosenthal, R., Hall, J. A., DiMatteo, M. R., Rogers, P. L., & Archer, D. (1979). Sesnsitivity tononverbal communications: The PONS test. Baltimore, MD: Johns Hopkins University Press. Sabatelli, R. M., & Rubin, M. (1986). Nonverbal expressiveness and physical attractiveness as mediators of interpersonal perceptions. Journal of Nonverbal Behavior, 10, 120-133. SOCIAL SKILLS INVENTORY 33 Schutte, N. S., Malouff, I. M., Hall, L. E., Haggerty, D. J., Cooper, J. T., Golden, C. J., & Dornheim, L. (1998). Development and validation of a measure of emotional intelligence. Personality and Individ ual Differences, 25, 167-177. Segrin, C., & Flora, J. (2000). Poor social skills are a vulnerability factor in the development of psychosocial problems. Human Communication Research, 26, 489-514. Snyder, M. (1974). The self-monitoring of expressive behavior. Journal of Personality and Social Psychol ogy, 30, 526-537. Snyder, M. (1987). Public appearances/private realities: The psychology of self-monitoring. New York: W.H. Freeman. Thorndike, E. L. (1920). Intelligence and its uses. Harper's Magazine, 140, 227-235. Tucker, J. S., & Riggio, R. E. (1988). The role of social skills in encoding posed and spontaneous facial ex pressions. Journal of Nonverbal Behavior, 12, 87-97. Zuckerman, M., & Larrance, D. (1979). Individual differences in perceived encoding and decoding abil ities. In R. Rosenthal (Ed.), Skill in nonverbal communication (pp. 171-203). Cambridge, MA: Oelgeschlager, Gunn, & Hain. This page intentionally left blank The Emory Dyssemia Index Marshall Duke Stephen Nowicki, Jr. Emory University INTRODUCTION For more than two decades, we have been examining the relationship between social/interpersonal difficulties and nonverbal language deficits in children and adults (Nowicki & Duke, 1992,1994,2002,2003; Duke, Nowicki, & Walker, 1996). Over this time, we and others (e.g., Elfenbein 8? Ambady, 2002; Feldman & Thayer, 1980) have established a significant link between these two variables. As a result, a variety of assessment and intervention procedures have emerged that are directed at improving social/interpersonal relationships via the strengthening of expressive and/or receptive nonverbal abilities (Feldman, Philippot, & Custrini, 1991). To remediate nonverbal language problems successfully, however, it is first necessary to diagnose them reliably. Whereas the extensive battery of expressive and receptive scales gathered under the rubric of the Diagnostic Analysis of Nonverbal Accuracy (DANVA; Nowicki & Duke, 2003) is suitable for research and for individualized clinical application, we be lieved that there was a need for a method by which larger numbers of people could be screened by psychologists and nonpsychologists with limited time available. Hence, to complement the more time-consuming DANVA, we developed an easy-to-use checklist, The Emory Dyssemia Index (EDI; Love, Duke, & Nowicki, 1994). The EDI, described here, is modeled after other widely used and effective screen ing checklist measures such as the Achenbach Child Behavior Checklist (CBCL; Achenbach 1991). We selected a similar format for two reasons. First, due to their familiarity with the widely used CBCL, teachers would have little difficulty learning to use the EDI. Second, such checklists can be submitted easily to procedures to es tablish their reliability, validity, and usefulness in screening large numbers of peo ple with minimal time requirements. This chapter outlines the development of and uses for the EDI. 35 36 DUKE AND NOWICKI NEED FOR THE EMORY DYSSEMIA INDEX Although children may develop nonverbal expressivity problems for a variety of reasons, there can be both biological and social causes for these difficulties. For ex ample, children may be born with or acquire damage to the neural processing sys tems necessary for interpreting, translating, and expressing nonverbal cues. The limbic system or structures associated with the amygdale are most often impli cated as crucial to the proper processing of nonverbal information. Whereas neural difficulties play a significant role in some relationship problems—especially serious ones—associated with nonverbal cues, we believe that most nonverbal expressivity problems result from simple failures to learn. By this we mean that, although children may be perfectly capable of learning the appropri ate means of communicating nonverbally, they may not have had the opportunity to observe how others express themselves nonverbally and/or be reinforced for showing correct expressions. If it is assumed that children learn proper nonverbal expressivity within their families, it likely is that most parents and others around the children may not show the full range and intensity of emotions nonverbally that children need to see and hear in order to learn. If not corrected with school experi ence, then children's nonverbal problems may continue or perhaps worsen as their social problems increase. For these reasons, the early detection—usually within the education system—of such inappropriate use of nonverbal expressive behavior becomes important. In that the DANVA and other more extensive measures of expressive and receptive nonverbal abilities typically require extended training in their use as well as signifi cant amounts of time and equipment, the EDI was developed in hopes of providing a reasonably reliable and valid method for teacher-based assessment of nonverbal language capacity in large groups of children or in children who appear to have some social/interpersonal difficulties. Researchers and social workers must always be aware of the degree to which they may interrupt ongoing school activities when they are working with children, and maximum effectiveness appears to be related to minimum intrusion coupled with maximum input from teachers. Easy-to-use checklists seem to fit these criteria quite well. Development of the Emory Dyssemia Index The development of the EDI followed traditional test-development guidelines. First, a sample of 20 teachers was drawn from a variety of public and private schools rang ing from elementary through high school. The teachers were asked to generate state ments describing the behavior of children whom they knew to have social difficulties in the form of higher rates of conflict with others, isolation, neglect by others, rejec tion by others, or difficulty establishing and maintaining friendships. Among them, teachers generated more than 300 descriptive phrases, such as "touches others when they don't want to be touched" and "speaks too loudly." THE EMORY DYSSEMIA INDEX 37 A group of psychologists and psychology graduate students (n = 8) familiar with nonverbal channels were asked to select from among the teacher-generated phrases all those items that dealt with one or more nonverbal cues (e. g., touch, fa cial expression, postures, gestures, clothing/jewelry, and paralanguage). The re sult of this selection process produced a list of 112 items. The 112 items were then grouped according to general nonverbal categories (e.g., kinesics, paralanguage, facial expression). Each item was placed next to a rating scale. The rating scale ranged from 1 to 4 on the basis of frequency of occurrence of these behaviors in a child (1 = never; 2 = rarely [once in 2 or more weeks]; 3 = sometimes [weekly]; 4 = often [daily] and 5 = very often [several times daily]). This "Beta" form of the EDI was then given to a sample of 104 teachers who were taking summer continuing education courses and volunteered to help develop the measure. The teachers were asked to think of one child they had known who had significant social/interpersonal difficulties and one child that was an interpersonal "star." The