Content Analysis of Verbal Behavior

New Findings and Clinical Applications

Louis A. Gottschalk

• Publisher: Routledge
• Year: 2014
• eBook ISBN: 9781317843917
• Topic: Psicología
• Subtopic: Psicología clínica

Focusing on language and the assessment of its meaning, this volume concentrates on a method of content analysis developed by the author and Goldine Gleser. Applicable to transcripts of speech or verbal texts, this method uses the grammatical clause as its smallest unit of communication, considers whether or not a verb is transitive and involves an object, or is intransitive and describes a state of being. It derives scores on many scales that have been tested for reliability of scoring and for construct validity with concurrently administered measures, such as rating and self-report scales as well as biochemical and pharmacological criteria. Finally, this volume provides detailed descriptions of the clinical and basic research establishing the validity of these scales, so that a reader can locate studies that have pertinence to any special interest area.

A major achievement described in this book is the development of computer software that understands grammar and syntax, can parse natural language, knows most of the words in the Merriam-Webster dictionary, has been taught to identify idioms and slang, and is capable of continuing to learn. The program can score all the scales, report whether the scores obtained from a verbal sample are one to three standard deviations from the norms, and suggest APA DSM-IIIR diagnostic classifications the clinician might consider in assessing the patient.

1	Introduction

The content analysis of language as a method of assessing what people say or write about in speech and texts and how strongly they may feel about their subject matter has undergone considerable growth, especially in the past decade. A perusal of refereed journal articles published over this period of time reveals articles originating from multiple disciplines and fields, for example, medicine, surgery, psychiatry, nursing, psychoanalysis, psychology, sociology, history, marketing, advertising, business management, and so forth.

What subject matters are currently being dealt with or illustrated by content analysis? Let us here limit such a survey to refereed journal articles published over the past 10 years from a worldwide selection in which an English abstract has been made available. An abbreviated listing of some of the subjects covered will give the reader an idea of their diversity and breadth.

Alexithymia–the split brain

Affective verbal measures

Anxiety—preoperative, prediction of postoperative outcome

Behavioral geography

Cancer, breast implant

Cancer, quality of life

Cardiac disorders, myocardial infarction and content analysis

Cardiac problems, nursing care

Cardiac recovery, nurses’ recordings

Children, attitudes toward lesbians and gays

Children, Chinese versus American children

Children, concepts of grandparents or old age

Children, conversational behavior of accepted and rejected children

Children, coping behavior in hospitals

Children, delinquency—early memories

Children, hyperactive

Children, imaginary companions

Children, of alcoholics

Children, self-help groups of parents

Children, suicide

Computer analysis, altered states of consciousness

Computer analysis, anxiety

Computer analysis, nursing research

Computer analysis, poetry of Emily Dickinson

Content analysis, autobiographies of famous psychologists

Content analysis, buyer–seller interaction

Content analysis, nursing and initial clinical experiences

Content analysis, surviving loneliness

Content analysis, diaries of a feminist novelist

Content analysis, dream anxiety and cerebral glucose metabolism

Content analysis, concepts of God

Content analysis, good life in advertising

Content analysis, guide for raters of two-person conversations

Content analysis, health and women’s images of health

Content analysis, nursing and initial clinical experiences

Content analysis, Shakespeare’s sonnets

Depression and anxiety, prediction

Depression, aggression in the dreams of blind women

Dreams, cultural differences

Drug addiction, explanatory style

Fear of success

Geriatrics, depression

Happiness

Helper secrets, from nurses

Hope

Hospitalization, impact on family

Managerial communications, crises

Posttraumatic stress disorders, holocaust effects

Primary process, measurement in paintings

Psychopolitics

Quality of life

Schizophrenia

Sexual arousability

The methods used in these content analysis studies range from simple counts of words, to counts of parts of speech, counts of classes of words, or counts of categories of subject matters. Some content analysis methods avoid doing any counts, but rather they record recurring sequences of emotions or feelings, which they may or may not link up with changes in vocal qualities (paralanguage variables) or body movement (kinesic variables). Some methods disregard issues of the magnitude of any types of themes or subject matters. Other methods focus on establishing a measure of the intensity or magnitude of a subject matter or psychological or behavioral dimension.

The Gottschalk–Gleser content analysis method, since its inception, has aimed at measuring the magnitude of the psychobiological dimension on which it is focusing. It has selected the grammatical clause as the unit of communication from which it derives the semantic message being conveyed by the speaker or writer. This book largely limits its focus to the Gottschalk–Gleser content analysis method, but it touches here and there on how other methods of content analysis or methods of measuring psychological or behavioral dimension relate to it.

AIM OF THIS BOOK

The overall aim of this book is to familiarize readers with the origins and development of the Gottschalk–Gleser method of measurement of psychological dimensions in children and adults through the analysis of the content and form of their verbal behavior. The method evolved originally as a research tool and instrument, and it has been used in this way for several decades. Over this period of time, many clinical applications have become apparent and have been demonstrated, but these potential applications have not been broadly utilized.

There are two major reasons for this limited application of the method. One reason is that the peer-refereed journals and books written on the research findings involving the method are not widely read by clinicians. The other reason is that the use of this method of measuring various psychological dimensions through verbal behavior analysis requires learning how to score the typescripts of spoken or written natural language up to a level of interscorer reliability of .80 (in comparison to scorers who have established expertise using the method) in order to achieve maximal validity in any research enterprise. The progressive development of computer-driven artificial intelligence software programs that are capable of scoring these verbal analysis scales (Gottschalk & Bechtel, 1982, 1989, 1993; Gottschalk, Hausmann, & Brown, 1975) with a reliability equal to expert human scorers (namely, ≥ .80) resolves the latter problem. This book is intended to help remedy both of these problems.

The procedure described in this book is the fruit of the amalgamation of different scientific disciplines and technical knowledge. Hence, because these sciences, as well as the technical fields, are themselves growing and evolving, one must expect that this offspring, this methodology, of such modern and diverse scientific and technical fields is in a state of change with continuing new discoveries. The intention here is to provide information about ongoing or recently completed research regarding new content analysis scales based on the content analysis methods described. In addition, the aim is to report not only on past but also on new studies involving the application of these scales to clinical problems, as well as to further construct validation research. Finally, the plan is to share with readers what verbal behavior analysis scales have definitely become scorable by computerized technology. Examples are provided of computerized assessments of 5-min speech samples. These examples will include actual scores obtained per grammatical clause of a speech sample, comparisons of computer-derived scores to norms, and suggestions of clinical diagnostic classifications from the DSM-III-R the clinician might consider in making a diagnostic formulation.

BRIEF HISTORY OF THIS METHOD OF ANALYSIS OF VERBAL BEHAVIOR EXTENDING FROM INITIAL AIMS TO CURRENT RESEARCH GOALS

While a research associate at the Institute for Psychosomatic and Psychiatric Research and Training of the Michael Reese Hospital in Chicago (1948–1951), Dr. Louis Gottschalk became interested in exploring emotional factors that contribute to or trigger epileptic convulsions. The results of his studies provided strong evidence that the arousal of psychological events played a causal role in the occurrence of the epileptic seizures in the children with whom he worked (Gottschalk, 1953, 1956). On moving to the National Institute of Mental Health to become a research psychiatrist in 1951, he decided to continue to pursue psychophysiological factors contributing to epileptic seizures in adults. He located military personnel at Walter Reed Army Institute of Research who had developed seizures during their tour of military duty. He selected patients who had frequent discrete abnormal paroxysms (about one every 1–3 min) of readily distinguishable, high-amplitude, fast or slow waves in their electroencephalograms. He had these patients free-associate while their electroencephalogram was being recorded, and he synchronized the content of their free associations with the electroencephalographic findings. He found that merely speaking as well as doing silent mental arithmetic tended to suppress the frequency of this paroxysmal electroencephalographic activity. But he discovered that the arousal of separation anxiety in the content of their spoken speech precipitated much of the epileptic seizure activity in the electroencephalograms of these patients (without their having overt seizures) or their grossly observable seizures (Gottschalk, 1955). Such observations strongly motivated him to try to develop more objective ways than the usual careful empathic listening of a neuropsychiatric and psychoanalytic clinician to assess the intensity of separation anxiety or other subjective affects. He decided to do so by focusing on the development of an objective method of measurement of psychological states from the content analysis of speech.

Research With Gove Hambidge, Jr.

At the National Institute of Mental Health, he teamed up with a fellow psychiatrist and psychoanalyst, Dr. Gove Hambidge, Jr., to pursue developing a method of analyzing and measuring communicative processes. They explored a number of ways of eliciting and recording different channels for the communication of affects and other psychological states, for example, by lexical features in speech, by paralanguage features (vocalizations, voice quality, voice set by linguistic features (junctures, stress, pitch), and by kinesic features (posture and gestures) of behavior. Although both Gottschalk and Hambidge were accustomed, as clinicians, to using cues from all possible channels of communication with respect to how a person is feeling and thinking, they finally decided that it was not possible for them to analyze and measure systematically all the elements in these diverse communication channels. They decided to focus their attention on characteristics of speech, especially lexical content features. They devised two standardized techniques of eliciting speech, both attempting to maximize the human tendency toward projection of intrapsychic qualities, response sets, and attitudes.

The verbal method of eliciting speech used instructions that told subjects that this was a study of speaking and conversational habits and asked the subjects to talk for 5 min about any interesting or dramatic personal life experiences they had ever had. The visual method of eliciting speaking involved asking subjects to tell stories for 5 min about certain pictures from Murray’s Thematic Apperception Test (1943). The first paper published from their early investigations indicated that these two different methods of eliciting speech evoked from the same individual significantly different percentages of several different classifications of words (Gottschalk & Hambidge, 1955). That is, these two different methods had different effects on the form of speech (i.e., such variables as rate of speech, frequency and duration of pauses, frequency of incomplete words and of nonlexical vocalizations) and the content (considerations of the meanings, relationships, objects, concepts, and processes symbolized by words) as well as the grammatical classifications of words (including adjectives, adverbs, nouns, pronouns, verbs, prepositions, conjunctions, and interjections).

Their next study involved an examination and comparison of the differences in the form and content of the speech patterns of a small group of psychotic and nonpsychotic individuals (Gottschalk, Gleser, & Hambidge, 1957). Although the research data in this study were collected in 1952 and 1953 by Gottschalk and Hambidge while at the National Institute of Mental Health, they were not fully analyzed and prepared for publication until 1956 after Dr. Gottschalk had moved to Cincinnati. Dr. Goldine Gleser assisted in the statistical analysis of the data.

Differences were found between the two groups in both the form and content of 5-min verbal samples elicited by the verbal instructions which asked subjects to talk for 5 min about any interesting or dramatic personal life experiences they had ever had. Formwise, the psychotic subjects showed a decreasing word rate per minute as compared to the nonpsychotic subjects. Contentwise, the psychotic patients showed a greater number of references to the self and to negations; more use of verbs; less use of we and us, of words referring to inanimate objects, and of qualifying words; and significantly less use of words that denote or connote location or spatial relationship (regardless of grammatical use) as compared to the nonpsychotic subjects.

Another study carried out by Gottschalk and Hambidge involved the verbal analysis of 80 verbal samples obtained from a single patient on a once-a-week basis over a period of 1 year while this individual underwent a personal psychoanalysis by Gove Hambidge (Hambidge & Gottschalk, 1958). Forty of the 80 verbal samples were elicited by the verbal method, which is purposely ambiguous and simulates the psychoanalyst’s request of the patient to free-associate. Forty other verbal samples were elicited by the visual method, which simulates the projective test situation. Evidence was obtained that indicated that the psychoanalytic experience and the state of therapeutic progress at the time of giving the verbal sample influence the way a patient expresses himself or herself with regard to the choice and use of language variables in the verbally induced and visually induced verbal samples.

After moving to Cincinnati to join the Department of Psychiatry as a Research Professor of Psychiatry in September 1953, Dr. Gottschalk began to explore the use of content features in language as a possible tool for the quantitative analysis of psychological states. He began to examine thematic content categories, using the grammatical clause (“molecular” approach) instead of a single word (“atomistic” approach) as the unit of communication. With new collaborators, he found—in a double-blind drug-placebo study—that a psychoactive pharmacological agent, pipradrol—a medicament similar in action to dextroamphetamine—produced significantly increased thematic references in brief samples of speech to achievement strivings (Gottschalk et al., 1956).

Research With Stanley Kaplan

In another study, a very complex psychological coping mechanism (a masochistic, problem-solving approach to obtaining support and love) assessed from the content analysis of 5-min speech samples was found to be significantly associated with the percentage of streptococcal bacteria that could be cultured from the oropharynx of a woman with rheumatic heart disease (Kaplan, Gottschalk, & Fleming, 1957). A cross-validation study confirmed this hypothesis by finding a significant correlation (.40, p < .05) between a frequency count of relevant themes in 20 5-min verbal samples and the relative percentage of streptococcal bacterial colonies cultured from the patient’s throat (Gottschalk & Kaplan, 1958; Kaplan & Gottschalk, 1958). The thematic categories selected for measuring the dimension of “masochistic problem-solving approach to obtain support and love” from the typescripts of 5-min speech samples were arrived at by adapting the “focal conflict” paradigm first described by French (1954) and later elaborated on by Gottschalk (1976a) and Kepecs (1979).

Collaboration With Goldine C. Gleser and Carolyn N. Winget at the University of Cincinnati

The most prolific and creative collaboration for Louis Gottschalk in developing a method of measuring psychological states through the content analysis of verbal behavior took place over more than a decade with Goldine Gleser. Dr. Gleser, also on the faculty of the University of Cincinnati, was a psychologist with considerable experience in interdisciplinary research and statistics. They decided to study speech at both the “atomistic” level (i.e., on the basis of words) and at the “molecular” level (on the basis of the grammatical clause and themes). An initial step in their collaborative research was to obtain 5-min verbal samples from 90 “normative”¹ individuals, all medically healthy and capable of gainful employment. The group was stratified for sex and intelligence. From this investigation there came a contribution to the atomistic approach to content analysis, giving in detail the relationship of sex and intelligence to the choice of words (Gleser, Gottschalk, & John, 1959). Data from this sample of subjects were also used in a number of reports featuring the thematic (and grammatical clause) approach to content analysis (Gleser, Gottschalk, & Springer, 1961; Gottschalk et al., 1960; Gottschalk, Gleser, Magliocco, & D’Zmura, 1961; Gottschalk, Gleser, & Springer, 1963; Gottschalk & Gleser, 1964).

The “atomistic” approach was applied to a set of false and genuine suicide notes collected by Shneidman and Farberow (1957) in an investigation to determine factors contributing to suicide. Analysis...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Dedication
• Table of Contents
• 1. Introduction
• 2. Review of Reliability and Validity Studies Using the Gottschalk-Gleser Method of Content Analysis
• 3. Range of Some Types of Applications of This Method of Verbal Content Analysis
• 4. Psychosocial Research Using This Method
• 5. Content Analysis Studies Involving the Neurosciences, Neuropsychopharmacology, and Biological Psychiatry
• 6. Applications of New Findings Involving the Content Analysis of Verbal Behavior to Clinical Medicine and Clinical Psychiatry
• 7. Measurement of the Content Analysis of Natural Language by Computerized Artificial Intelligence as a Means of Facilitating and Speeding Up Verbal Behavior Analysis
• References
• Appendix
• Author Index
• Subject Index