The Individual Group Member Interpersonal ... - Semantic Scholar

11 THE INDIVIDUAL GROUP MEMBER INTERPERSONAL PROCESS SCALE MICHAEL SEAN DAVIS, SIMON H. BUDMAN, AND STEPHEN SOLDZ

In this chapter we describe the development of the Individual Group Member Interpersonal Process Scale (IGIPS), a group therapy process measure based on an interpersonal, interactionally oriented approach to group psychotherapy. We delineate the nature of the group theory underlying the IGIPS as well as the entire developmental history of the instrument. The rationale for the creation of the measure and three successive revisions is then presented. In total, the development of the IGIPS represents the culmination of nearly a decade of research in the area of group psychotherapy. In addition, the structure and properties of the latest version of the system are described and research findings using the instrument outlined. Finally, we summarize and discuss the application of the IGIPS to Group A process data.

HISTORY AND THEORY OF THE SYSTEM The IGIPS was designed to test the assumptions underlying an interpersonal, interactionally oriented approach to group psychotherapy. This

283

approach, described in detail by Yalom (1985) and adapted to a timelimited format by Budman and Gurman (1988), presumes that group therapy has curative possibilities that are indigenous to the group treatment mode. According to this theory, the group becomes a “social microcosm” of the outside world in which maladaptive interpersonal patterns are replayed, recognized, and acknowledged. As group members reenact their interpersonal difficulties in the group, they are likely to receive disconfirming evidence about the beliefs that are causally related to maladaptive interactions. This provides individual group members with a unique opportunity to become aware of problematic interpersonal behaviors and to experiment with new, more effective ones. In this model, therapists assume an active, structuring stance that primarily involves exploring and commenting on “here-and-now” group processes and their associative evolution. In addition, therapists take responsibility for galvanizing the group as a whole, drawing particular group members into interactions, setting limits on destructive behaviors, and tying themes back to central foci. When adapted to a brief, time-limited format, the model also incorporates the importance of a predetermined thematic focus, a sense of explicit time consciousness intended to mobilize change through the acknowledgment of existential limits, and an overt recognition of the group as part of a larger constellation of “therapeutic” experiences extending over the group members’ life span. The most influential factor in whether group members successfully use the opportunities inherent in the group therapy mode is group cohesiveness (Budman & Gurman, 1988). Cohesion is defined as the connectedness of the group, demonstrated by working together toward a common therapeutic goal, constructive engagement around common themes, and a n open, trusting attitude that allows members to share personal material. A cohesive group atmosphere induces a sense of individual safety in group members, enabling them to acknowledge interpersonal weaknesses and to initiate and sustain the difficult process of change. Cohesion also keeps individual members coming to the group in the face of strong or frightening emotionality, such as confrontation or hostility, or during periods without direct rewards. The initial attempt of the Harvard Community Health Plan-Mental Health Research Program (HCHP-MHRP) to study group behavior at the individual level resulted in the development of the “main actor” concept. The m i n actor is defined as the group member whose issues were most focused on in a given 30-min segment. Raters viewing videotapes of therapy sessions could identify main actors with high reliability. These early studies examined the relation of main actor process to pretherapy patient distress and therapeutic outcome. It was found that the number of times a patient was the main actor positively predicted outcome (Soldz, Budman, Demby, & Feldstein, 1990).

284

DAVIS, BUDMAN, A N D SOLDZ

Main actor process was further examined using an adaptation of the Vanderbilt Psychotherapy Process Scale (VPPS). The VPPS was selected because the results of many studies have suggested a connection between VPPS dimensions and outcome within individual therapy (GomesSchwartz & Schwartz, 1978; OMalley, Suh, & Strupp, 1983; Rounsaville et al., 1987; Suh, Strupp, & O’Malley, 1986; Windholz & Silberschatz, 1988). As many of the VPPS dimensions have potential relevance to group therapy, the scale was slightly modified and applied to individual patient behavior in the group setting. In one HCHP-MHRP study that used the modified VPPS suggested the existence of an interaction predictive of outcome between specific patient behaviors and the number of times a patient was the main- actor within the group (Soldz, Budman, & Demby, 1992). Results of another study indicated that the number of times a patient was the main actor was predicted by pretherapy disturbance level, but it did not predict patient outcome (Soldz et al., 1990). Despite these findings, it became clear that there were many limitations to the use of the VPPS as a measure of group process. For example, the VPPS was unable to take into account the various modes of focus an individual patient can adopt within the group setting. A patient in group therapy can choose to discuss personal issues or those of other group members (OGMs). No comparable choice confronts the individual therapy patient. Therefore, the VPPS was insensitive to this distinction. Researchers who have used the VPPS were forced to disregard the other-focused patient behaviors that occurred within the group and to examine only self-focused patient behaviors (main actors). This meant that many of the subtleties and complexities of the group process remained inaccessible to measurement and analysis. Before embarking on the development of the IGIPS, the HCHPMHRP also explored the possibility of using process measures other than the VPPS. Both an early version of the Therapeutic Work Scale (Connelly, Piper, & Braha, 1981) and the Experiencing Scale (Klein, Mathieu, Gendlin, & Kiesler, 1969; Klein, Mathieu-Coughlan, & Kiesler, 1986) were considered. These scales did not meet the specific needs of the intended research. As with the VPPS, the Experiencing Scale was designed for individual therapy and therefore was inherently limited in its application to the group format. Although more suited to group research, the sentenceby-sentence unit of analysis used by the Therapeutic Work Scale was too time-consuming. Moreover, our research team was unable to obtain acceptable reliability in our initial attempt at learning the scale. Most importantly, however, we decided against the other process measures because they were derived from different theoretical perspectives. By developing the IGIPS, the HCHP-MHRP was able to devise an individual-level instrument that was sensitive to both the group format and to our particular theoretical interests. INDIVIDUAL GROUP MEMBER JNTERPERSONAL PROCESS SCALE

285

The IGIPS has undergone numerous revisions. Initially, IGIPS ratings were made on 30-minute segments of therapy groups. Each rater would follow 2 or 3 patients over a 30-minute period and would perform ratings on each patient at the completion of the segment. This methodology was adapted to be as cost-effective and time-effective as possible. Preliminary work with this method suggested, however, that raters had difficulty remembering and averaging the behaviors over such a long period of time. As a result, the reliability of these global ratings was extremely low. To correct this problem, the scale was restructured in a manner that resembles its present form, where ratings are performed on each “statement” made by a group member. Since this fundamental shift in the unit of observation and rating, there have been three revisions of the scale: the IGIPS-I, IGIPS-11, and IGIPS-111.’ Unless otherwise indicated, in this chapter we focus on the most recent version of the scale, the IGIPS-111. The primary goal of each revision was to make the rating protocol more practical and timeeffective and ultimately to increase reliability. The revisions have also increased the scope of the phenomena that can be measured by the scale. For example, in revising the IGIPS-I, it was decided that the therapist should be rated as well as the group members and that the sequence of statements should be recorded. These modifications have made it theoretically possible to perform sequential analyses of group interactions and to explore the relationship between interventions and responses. As a result, the IGIPS has become a multifaceted, flexible group therapy process instrument.

CORRESPONDENCE BETWEEN THE THEORY AND THE SYSTEM For the 5 years before its focus on individual group member behavior, the HCHP-MHRP had been engaged in an ongoing research program. The research program was a reaction to the lack of empirical research on the validity of the “curative factors” putatively associated with group therapy (Kaul & Bednar, 1986). Initially, the program analyzed behavior at the group level. It explored whether the concept of group cohesiveness was as theoretically and clinically useful as the concept of “therapeutic alliance” was in individual therapy. This goal was pursued in a series of studies of time-limited (15-session) outpatient therapy groups (Budman et al., 1987, 1988, 1989) that showed that group cohesiveness was related to outcome measures, such as improved self-esteem and reduced symptoma‘Although more improvements and refinements could be made, because of the disbandment of the Harvard Community Health Plan-Mental Health Research Program research team, there are presently no plans to develop future iterations of the scale.

286

DAVlS, BUDMAN,A N D SOLD2

tology. These and other findings supported the clinical consensus that viewed cohesiveness as an important component of a successful group experience. In these studies, cohesion was measured using the Harvard Community Health Plan Group Cohesiveness Scale (GCS). The scale consists of five subscales and a global scale. The five subscales are (a) Withdrawal and Self-Absorption vs. Interest and Involvement, (b) Mistrust vs. Trust, (c) Disruption vs. Cooperation, (d) Abusiveness vs. Expressed Caring, and (e) Unfocused vs. Focused. The global scale is Global Fragmentation vs. Global Cohesiveness. Raters view 30-minute segments of group therapy process and apply each scale to the group as a whole along a 10-point continuum. The GCS is presented more extensively in Budman and Gurman (1988). After establishing a consistent link between cohesiveness and outcome at the group level, the HCHP-MHRP studied the microevents and individual group member behaviors that could affect cohesiveness and in turn differentiate between positive and negative outcome for groups and individuals. The IGIPS was created for this purpose. The guiding principle in the construction of the IGIPS was to select items derived from the concept of cohesiveness that could also be used and rated reliably at the microevent level. This presented several interesting challenges. The new format required raters to make ratings on brief statements that were sometimes only three words long. Making subtle or complex judgments about such short statements is intrinsically difficult and inevitably resulted in poor reliability. Consequently, items representing constructs from the GCS were often reworded to constrain the use of clinical inference or the reliance on implicit averaging across the poles of a dimension. For example, the bipolar construct Abusiveness vs. Expressed Caring -which was rated reliably on the group global level but presented difficulties on the individual statement level-was broken into two unipolar dimensions, Expresses Negative Sentiment and Expresses Positive Sentiment, which had ratings that indicated the degree to which each characteristic was present in each statement. By subtracting the sum of Expresses Positive Sentiment from the sum of Expresses Negative Sentiment and then dividing the result by the total number of statements within a given unit of time, it was possible to obtain a reliable score that was conceptually similar to the original GCS item. Note, however, that not all the original GCS constructs were able to be translated in this manner. The research team was unable to adapt the construct Focused vs. Unfocused, a construct reflecting the degree to which the group retained a thematic coherence, to the statement format. INDIVIDUAL GROUP MEMBER INTERPERSONAL PROCESS SCALE

287

DESCRIPTION OF THE SYSTEM Rating and Categories

IGIPS ratings are performed on each statement made by a group member. A stuternent is defined as a burst of speech of more than two words by a single person without a pause of more than 10 seconds. A typical statement is bounded by a statement made by other group members. Occasionally, a statement is bounded by a substantial (10-second) pause in an individual’s burst of speech. Each rating indicates to whom the statement is referring (i.e., self, therapist, group as a whole, or other group member), the locational focus of the statement (i.e., life outside the group vs. inside the group), and whether the statement was self-initiated or elicited. In addition to this and other fundamental information, derivative items from the GCS were adapted to the individual statement format. Examples of the adapted items include “Discloses Personal Material” and “Feels Connected to Others.” IGIPS ratings obtained for each statement can be aggregated and combined to form a multitude of possible variables, including ones reflecting the sequential relationships of group interactions. Therefore, the IGIPS is an all-purpose instrument that has the potential to measure a wide range of group phenomena. IGIPS ratings are made with reference to videotapes and transcripts of the therapy sessions. The videotapes help the raters ascertain the emotional tone of a statement, and the transcripts help them remember its content. After each statement, raters typically pause the videotape to determine whether it contained behavior that should be rated on any of the items. IGIPS-I11 consists of 23 items that raters need to consider. Some items are rated on a categorical basis, reflecting the presence or absence of a specific behavior. Other items are rated on a 0-9 Likert scale, reflecting the significance of the behavior. Significance is the intensity or judged psychological importance of a behavior. Given the theoretical orientation of the research team, significance ratings are explicitly associated with the degree to which the behavior reflects a heightened awareness of immediately present feelings and interpersonal processes combined with an increased understanding of the factors responsible for their existence. Although this principle is embedded in all the significance rating protocols, the actual guidelines for making the ratings emphasize different elements of the principle depending on relevance to each item. The section on strengths and limitations contains an example of significance ratings. Many of the items also contain locational and object designations that indicate the locational focus (i.e., inside group or outside group) to which the behavior is referring or to whom the behavior is directed. For example, when using the item “Expresses Negative Sentiment,” the rater

288

DAVIS, BUDMAN, AND SOLD2

is required to record the object toward which this sentiment was directed. The options include the group, the therapist, the self, the ID number of another group member, someone outside the group, or inanimate factors. Once the designations have been determined, the significance level of the sentiment is rated for each of the separate designations on a 1-9 scale. The item “Demonstrates Self-Awareness,” however, uses a different set of designations. For this item, the options include awareness related to either group process or to life outside the group. Again, separate significance level ratings (using the 1-9 scale) are performed for each of the two designations. Each of the 23 items has a unique set of appropriate designations ranging from 0-7 (see Table 11.1 for definitions, types of ratings, and examples). Limitations and Strengths The degree of inference that is needed to perform ratings varies for each item. No inference is necessary for some items, such as “Length of Statement.” It is required for other items, such as “Enhances OGM Awareness.” As previous experience with scale development has demonstrated, greater inference usually means lower reliability. Therefore, a concerted effort was made to control the degree of inference. The challenge was to limit the inference without reducing the clinical relevance of the items. The solution to this dilemma involved the incorporation of rigid, theoretically meaningful guidelines into the significance rating procedures whenever possible. This technique is best exemplified in the items that measure the degree of psychological mindedness, such as “Demonstrates Self-Awareness.” Ratings for this item are made in a two-step process. First, the numerical range within which the rater can operate is established. Ratings are limited to the lower range (1-4) of the 9-point scale if the target patient demonstrates awareness of a feeling or interpersonal enactment without s u g gesting the cause of these occurrences. The midrange (3-7) is used when an ongoing pattern of behavior is acknowledged or when a coexisting cause of a feeling or interpersonal enactment is identified. Finally, the high rate (6-9) is used when a historical cause for a feeling, interpersonal enactment, or pattern of behavior is identified. For example, the statement “I’ve been depressed lately” would be limited to a lower range rating, yet the statement “The incident with my wife has made me depressed lately” would receive a midrange rating. On the other hand, the statement “I’ve been depressed lately. I usually get depressed when I am angry. My family never let anybody get angry, so I don’t know what else to do” would receive a rating in the high range. Once the range is determined, the final rating is adjusted based on examination of other criteria. These include the amount of total awareness INDIVIDUAL GROUP MEMBER INTERPERSONAL PROCESS SCALE

289

Frequency Frequency Frequency

Item 4: Discusses Self Item 5: Discusses OGM Item 6: Discusses Therapist

Item 7: Discusses Impersonal Issues

Item 8: Uses Humor

Item 9: Self-Initiated Statement Rated when statement was not made in response to a direct inquiry or comment. Item 10: Discloses Personal Material Information an “average” person would usually not reveal to strangers in commonplace social situations. Item 11: Expresses Affect Statements delivered in a manner that could be appropriately characterized by an adjective connoting a specific emotion. Significance

Significance with insideand outside-group designations

The ID number of the patient making the statement Each 5-s interval represents 1 length unit Frequency Frequency Frequency

Item 2: Target Patient ID

Item 3: Statement Length

The sequential number of each statement

Type of Ratings

Item 1: Statement Number

Statement Categories Examples

Outside-group designation: “I am uncomfortable in most social situations.” Inside-group designation: “I feel inferior to the people in this group.” Statement delivered with a specific emotional tone, such as sad, happy, angry. Ratings reflect nonverbal behavior.

ing of 1, 2, and 9, respectively. “I was upset about what happened.” “You [OGM] seemed upset by what happened.” “I think you [therapist] should make sure these things don’t happen in here.” “Society needs to provide more resources for children.” A statement that elicits laughter in at least one OGM. Therapist: “Who would like to start today?” Group member: “I have something to discuss.”

3-,8-, and 45-s statements would receive a rat-

The 3rd statement would receive a rating of 3, and the 450th statement would receive a rating of 450. A statement made by a patient with an ID number of 6 would receive a rating of 6.

TABLE 11.1 Individual Group Member Interpersonal Process Scale Categories

Item 15: Feels Connected to Others Overt acknowledgment of similarity, shared experiences, or feelings of intimacy with others. Item 16: Feels Disconnected From Others Overt acknowledgment of feeling dissimilar, alienated, or detached from others. Also includes feelings of loneliness. Item 17: Demonstrates Self-Awareness Overt acknowledgment of a feeling, pattern of behavior, or interpersonal enactment. Significance ratings are scored higher for identification of causes. Item 18: Exhibits Unawareness, Uncertainty, or Confusion Rated when unawareness, uncertainty, or confusion are exhibited in regard to identity, goals, or causes of a feeling, interpersonal enactment, or pattern of behavior. Item 19: InterpersonalSensitivity A communication directed toward an OGM delivered in a warm, tolerant, or sensitive manner.

Item 14: Expresses Positive Sentiment Overt expressions of positive thoughts or feelings.

Item 12: Resistance to Affect Statements delivered in a manner that could be characterized as stiff, flat, detached, monotonic, or withdrawn. Item 13: Expresses Negative Sentiment Overt expressions of negative thoughts or feelings.

Table continues

“I image that was difficult for you [an OGM]. It sounds like you did the best anyone could.”

Significance


Outside-group designation: “I like to keep things of a superficial level until I know someone well.” Inside-group designation: “I am angry because you interrupted me.” Outside-group designation: “I don’t know why I was so mean to him back then.” Inside-group designation: “I’m not sure how I feel about the group.”

POG designation: “I never fit in with anyone in high school.”

Statement delivered with a specific tone that is unable to be described with an adjective connoting an emotion. Ratings reflect nonverbal behavior. OGM designation: “I didn’t think you [an OGM] were helpful last week.” IF designation: “Life sucks. There is no denying it.” Self-designation: “Despite what everyone said, I think I did a good job.” POG designation: “My sister has always been helpful to me.” Group designation: “Everyone in the group has had problems with relationships.”


Significance with group, therapist, self, OGM, POG, and IF designations Significance with group, therapist, self, OGM, POG, and IF designations Significance with group, therapist, OGM, and POG designations Significance with group, therapist, OGM, and POG designations

Significance

98

4U

m

I

$j

~~~~~



Outside-group designation: “Tell your husband you can’t handle that right now.” Inside-groupdesignation: “If you are angry with someone in here, then you should tell them.”

Outside-group designation: “How is your wife doing?” Inside-groupdesignation: “What did you think about last week?” Outside-group designation: “It seems like you go through cycles of high and low energy.” Inside-groupdesignation: “You look sad today.”

Examples

~


~~~

“It doesn’t sound like you [an OGM] were very effective in the situation. Couldn’t you do any better.”

~

Significance

Type of Ratings

(Continued)

Note. OGM = outside group member; POG = people outside group; IF = inanimate factors. In addition to statement number, patient ID, and length ratings, three other categories of ratings are possible for each statement: frequency, significance, and locational-object designations. Frequency ratings are binary ratings indicating the presence or absence of a behavior. Significance ratings are continuous ratings indicating the intensity or judged psychological importance of the behavior. The definitions and guidelines for making significance ratings adhere to a common principle but are implemented differently for each item. Location-object designations indicate the object to which a feeling is directed or the context in which a behavior occurs. The possible object designations include the group as a whole (group), therapist, self, OGMs, POG, and IFs; the context designations include outside the group and inside the group. Different designations are used for different items depending on appropriateness.

Item 20: Interpersonal Insensitivity A communication directed toward an OGM delivered in a cold, intolerant, or insensitive manner. Item 21: Asks Questions Any communication in question form directed to an OGM, the therapist, or the group as a whole. Item 22: Enhances OGM Awareness A communication that highlights the existence of a feeling, pattern of behavior, or interpersonal enactment of an OGM. Significance ratings are scored higher for suggestion of causes. Item 23: Gives Advice A directive or suggestion intended to encourage an OGM to think, feel, or behave a particular way.

Statement Categories

TABLE 11.1

demonstrated within the statement, as well as the degree to which the awareness is differentiated and focused on the immediate present and interpersonal issues. Finally, the raters clinical inference about the importance of the insight can also influence the final rating. Thus, ratings for the above item are constrained by specific theoretical assumptions. The most important of these are that (a) therapeutic growth is related to an elaborated understanding of the relationship between life experience and present behaviors and (b) in the group setting, this elaborated understanding is most effective when applied to interpersonal behaviors occurring within the immediate present. Similar constraints and techniques are used when making significance ratings for other items.

METHODOLOGY FOR ANALYSIS Although the IGIPS-111 contains only 23 items, the numerous designations associated with the items create thousands of formats for aggregating data. To organize this data into manageable units, it is necessary to create reduced items that combine these possible formats. Different researchers may use different reduced items depending on the unit of analysis and the requirements of their research. The criteria should be that their distributions are normal, their interrater reliabilities sufficient, and they measure discrete dimensions that overlap only minimally with other reduced items. Table 11.2 contains an experimental set of reduced items at the individual level, their available reliability coefficients, and formulas indicating how each item was created.’ Most scores for the reduced items are calculated by summing the relevant ratings at the statement level across an entire session and then dividing that sum by an appropriate denominator. Some reduced items use the total number of statements made by the patient as the denominator, whereas others use either the total number of self-focused or OGM-focused statements as the denominator. For example, the reduced item “Feels Connected to Group” is derived by subtracting the sum of “Feels Disconnected From Others” (with in-group designations) from the sum of “Feel Connected to Others” (with in-group designations) and dividing the result by the total number of statements made by the patient. Other reduced items, such as “Demonstrates Insight Into Own Life,” are obtained by summing the ratings of the item “Demonstrates Self-Awareness” (with an out-group designation) and dividing by the total number of statements in which the patient discussed his or her own issues. Still other items use a different ’The items listed in Table 11.2 represent the most recent reduced items constructed by the Harvard Community Health Plan-Mental Health Research Program. This set of items was created to measure behaviors the research team considers to be theoretically important. Beyond their face validity, the validity of this set of items has not been established.

INDIVIDUAL GROUP MEMBER INTERPERSONAL PROCESS SCALE

293

TABLE 11.2 IGIPS-Ill Reduced Items with lnterrater Reliabilities Reduced Items Amount Spoken (Zltem 3) Focuses on Self (Zltem 4nS) Focuses on OGMs (Zltem W S ) Focuses on Impersonal or Abstract Issues (Cltem 7mS) Focuses on Group Process (Zany item with any inside-group designationns) Uses Humor (Zltem 8nS) Initiates Comments (Zltem 9nS) Discloses Personal Material About Own Life (Zltem 10 with Designation OG/TS) Discloses Personal Material About Group (Cltem 10 with Designation IGES) ExpressesAffect [(Cltem 11-Zltem 12)TTSI Quality of Sentiment Expressed Toward Group [(Cltem 14 with Designations G, T, OGM-Cltem 13 with Designations G, T, OGM)nS)] Quality of Sentiment ExpressedAbout Life Outside Group [(Cltem 14 with Designations IF and POG-Cltem 13 with DesignationsIF and POG)/'TS)] Feels Connected to Group [(Zltem 15 with Designations G, T, OGM-Cltem 16 with Designations G, T, OGM)/TS)] Feels Connected to People Outside Group [(Cltem 15 with Designations POG-Xltem 16 with Designations POG)TTS)] Demonstrates Insight Into Own Life Issues (Cltem 17 with Designation OGEltem 4) Demonstrates Insight Into Own Group Issues (Zltem 17 with Designation IGEltem 4) Demonstrates Insight Into OGM Life Issues [(Zltem 21 with Designation OG + Item 22 with Designation 0G)Eltem 51 Demonstrates Insight Into OGM Group Issues [(Zltem 21 with Designation IG + Item 22 with Designation IG)Eltem 51 Amount of Input Received From Group (Cfrequency ratings from Item 14 through Item 16 + Item 21 through Item 23 received from group) Psychological Mindedness of Input Received From Group (Mean significance ratings from Item 21 + Item 22 + Item 23 received from group) Affective Quality of Sentiment Received From Group [(Cltem 14 received from group-Cltem 13 received from group)Einput received from group]

ICC .96 .97 .88 .81 .92 .91 .92 .56 .90 .79 .82 .71 .54

.67 .87 .88 .65 .77 .98 .83 .82

Note. N = 23. The unit of analysis was individual group member behavior across a whole session. Two raters were used to rate eight group members in three different sessions (one member missed one session). ICC = intraclass correlation; Item = item numbers in Table 11.1; X = sum of indicated ratings for each individual group member across a whole session; TS = total statements made by a group member within a session; G = group; T = therapist; OGM = other group member; POG = people outside group; IF = inanimate factors; OG = outside group; IG = inside group.

294


method of calculation altogether. For example, the reduced item “Amount of Input Received From Group” is calculated simply by summing the number of times feedback is directed at a given patient. Once the reduced items have been selected, the final step in the data reduction process involves conducting a principal-components analysis (PCA). The factors that emerge from the PCA should be used to analyze the data. The results of a PCA that was performed on reduced items from IGIPS-I are discussed later. Reliability The interrater reliabilities of the IGIPS-111 reduced items listed in Table 11.3 appear to be good to excellent. Individual rater interclass correlation reliability was calculated from the ratings of two raters who viewed three sessions of an eight-member group. The unit of analysis used in these calculations was the individual group member behavior per session (N = 23; one member missed one session). The reliability ranged from .54 to .98, with a mean reliability of .82 and a median of .83. Reliabilities for these 23 items were also calculated at the statement level. All 1,400 statements of two sessions were rated by different raters and then examined to ensure the correspondence between the sequence of ratings. At this level, reliability was notably lower but still acceptable. The intraclass correlations ranged from .33 to .86, with a mean of .65 and median of .69. Reliabilities for earlier versions of the scale were also acceptable, although they were somewhat lower than for the IGIPS-111. Thirty-five 30minute segments were rated by three different raters using the IGIPS-I. Reliability for the five factors derived from 12 reduced items ranged from .53 to .95, with a mean reliability of .70 (Soldz, Budman, Davis, & Demby, 1993). Reliability for the IGIPS-11, which more closely resembles the IGIPS-111, has also been calculated. Two raters rated 15 different 30minute segments. When aggregated at the group level, the mean intraclass correlation of the nine reduced items used in a study exploring the component parts of cohesiveness was .80, the median was .85, and the range was .46-.94. Because each version of the IGIPS was revised with the goal of increasing reliability, it is not surprising that reliability tends to be higher for each new iteration of the scale.

PREVIOUS RESEARCH WITH THE SYSTEM The three versions of the IGIPS have been used in a variety of studies with varying degrees of validity. The IGIPS-I was applied to the videotapes of the first 4 sessions of seven 15-session interpersonally oriented outpatient INDIVIDUAL GROUP MEMBER INTERPERSONAL PROCESS SCALE

295

?

5

20

c

%

-2

38 0.24 0.42 1.2 2.8 0.12 0.80 2.5

21 2 0.29 0.57 1.3 1.5 -0.36 2.2 1.8

Amount spoken Discusses abstract issues Discusses group process Discloses personal material Expresses affect Feels connected to group Insight into own issues Insight into OGM issues

0.87 1.1 3.6 -0.24 2.2 2.4

0.00

1 40

Diane

47 0.00 0.45 0.30 1.2 0.09 0.00 2.8

Brad

Facilitators

99 0.22 0.38 1.3 3.3 0.14 1.5 1.6

Martha

38 0.00 0.91 1.2 2.4 0.33 2.5 4.0

Pat

64 0.03 0.60 0.20 2.6 -0.08 0.10 2.2

Helen

122 0.05 0.74 0.60 2.7 0.12 0.70 2.2

Alice

Therapists

Note. Approximate cohesiveness rating: midlevel. IGIPS = Individual Group Member Interpersonal Process Scale; OGM = other group member.

Greg

Joe

Reduced items

Main actor

TABLE 11.3 IGIPS Scores From the First Half of Session 3 of Group A

775 0.12 0.61 1.1 2.7 -0.04 1.7 2.2

Group as a whole

therapy groups that contained a total of 52 patients. Twelve reduced items were used in the analysis. A PCA was performed on these reduced items. The literature suggests that a two-dimensional circumplex structure of Dominance and Affiliation underlies interpersonal behavior (Wiggins, 1979, 1982). Therefore, two factors were initially extracted and then subjected to a varimax rotation. The two circumplex dimensions were clearly visible. The first factor, which was theoretically consistent with the Dominance circumplex dimension, had positive loadings from “Expresses Affect,” “Amount of Speech,” and “Discloses Personal Material.” The second factor, consistent with the Affiliation dimension, received loadings from “Connections With Others,” “Responsiveness to Input,” and “Positive Sentiment Toward Others.’’ These two factors, however, accounted for only 39% of the variance. Thus, standard procedures for determining the best factor structure were used, resulting in five factors that accounted for 74% of the variance. After varimax rotation, these five factors were as follows: Activity, Interpersonal Sensitivity, Comfort With Self, Self-Focus, and Psychological Mindedness. (see Appendix ll.A for a list of the reduced items that loaded on these five factors). Interestingly, this more comprehensive five-factor structure resembles the Big Five personality factors consistently found in factor analyses of personality descriptions (John, 1990; McCrae, 1989; McCrae & Costa, 1990). The Big Five has also been related to personality pathology in clinical samples (Costa & McCrae, 1990; Lyons, Ozer, Young, Merla, & Hyler, 1991; Soldz, Budman, Demby, & Merry, 1993; Wiggins & Pincus, 1989). The parallels between Activity and Extraversion, Interpersonal Sensitivity and Agreeableness, and Comfort With Self and the Big Five Neuroticism or Emotional Stability are clear. It also seems reasonable to associate Psychological Mindedness with Openness. Finally, it is possible to argue that the group behavior most representative of the Big Five Conscientiousness factor is the amount of talk directed at other people. This is because conscientious group members are likely to make an effort to further the group task by responding and reaching out to other group members. Because the remaining factor, Self-Focus, is the opposite of discussing others, it seems reasonable to identify this with the Big Five Conscientiousness dimension. This form of convergent validity both with the interpersonal circumplex model and the Big Five factors suggests that the IGIPS is capable of representing relevant and important dimensions of interpersonal beha~ior.~ Other findings from this study also support the validity of the IGIPS. Scores from the Self-Focus factor were positively correlated with the selfreport measures of pretherapy symptomatic distress. The Comfort With Self ’For a more detailed discussion of the first Individual Group Member Interpersonal Process Scale I factor structure and its correspondence with the interpersonal circumplex and Big Five models, see Soldz, Budman, Davis, and Demby (1993). INDIVIDUAL GROUP MEMBER INTERPERSONAL PROCESS SCALE

297

factor scores were correlated with the therapist’s global assessment of patient functioning. Finally, women scored higher than men on the Interpersonal Sensitivity factor score. This replicates other findings that show that women are more concerned with relationships and are more sensitive of the feelings of others (Gilligan, 1982; Lewis, 1976; Spence & Helmreich,

1978).

The IGIPS-I1 was also used in a study (Budman, Soldz, Demby, Davis, & Merry, 1993) designed to explore the component parts of group cohesiveness. In this study, 39 half-hour segments from 12 timelimited psychotherapy groups were measured on both the GCS and the IGIPS-11. Correlations between cohesiveness scores and the IGIPS-I1 reduced items aggregated across all group members suggested that particular group member behaviors are related to cohesion according to the phase of the group therapy. For example, outside-group-focused statements were positively correlated with cohesiveness in the beginning of the group but negatively correlated with cohesiveness at the end of the group. The reverse was true for inside-group-focused statements. These data are consistent with many of the stage-related theories of group development (Beck, Dugo, Eng, & Lewis, 1986; Budman & Gurman, 1988; MacKenzie,

1990).

Finally, there is preliminary, unpublished work with the IGIPS-111. In a long-term group for patients with personality disorders, the IGIPS111 was able to differentiate the early group process of patients diagnosed with avoidant personality disorders from the group process of patients without these diagnoses. These differences were theoretically consistent with the fourth edition of the Diagnostic and Statistical Manual of M e n d Disorders (DSM-IV; American Psychiatric Association, 1994). Compared with the nonavoidant patients, the 2 avoidant patients scored markedly lower on most of the expected dimensions, such as level of verbal activity, initiative, expressed affect, and connectedness. Similarly, IGIPS-111 scores were able to differentiate the group process of histrionic patients. IGIPS-111 ratings were performed on three early sessions of 14 patients distributed across five different time-limited psychotherapy groups. Seven patients received elevated scores on the Histrionic scale of the Millon Clinical Multiaxial Inventory 11 (Millon, 1987), and 7 patients received low scores on this same dimension. Each of the 7 histrionic patients were matched with a same-sex nonhistrionic patient in the same group. As with the avoidant patients, the group process of the histrionic patients differed from the nonhistrionic patients in ways consistent with the DSM-IV. Histrionic patients, for example, expressed more affect, initiated more statements, and expressed more feelings of being different from others. Although the IGIPS has not yet been used in this type of work, it clearly has the potential for analyzing the sequence of interactions among

298

DAVIS, BUDMAN, A N D SOLD2

group members in interesting ways. Unlike most work examining sequential interactions in psychotherapy (Bakeman, 1990; Gottman & Bakeman, 1986; Gottman & Roy, 1990), the IGIPS-111 has quantitative, rather than simply categorical, capacities. This makes it possible to use correlational techniques rather than traditional methods (Gottman & Bakeman, 1986; Gottman & Roy, 1990). An interesting use of the scale, for example, might be to examine the correlations between the type of input received from other group members or the therapist and the emergence of demonstrated self-awareness.

APPLICATION OF THE SYSTEM OF ANALYSIS TO GROUP A, SESSION 3 An analysis of Group A provides an excellent opportunity to illustrate some of the potential applications of the IGIPS-111. Although there are a variety of possible uses for the scale, this illustration focuses on the most elemental ones. The methodological protocol and programming for the more intricate and exciting applications-such as sequential analyseshave not yet been developed. In keeping with earlier HCHP-MHRP research endeavors, the purpose of the methodology that is presented here is to understand the development of cohesiveness at the group level. As referenced earlier, previous work with the IGIPS suggests a relationship between IGIPS scores and the cohesiveness level of the group as a whole. As an extension of this work, the HCHP-MHRP was interested in analyzing the behavior of the individual group members who assume the different roles indigenous to group therapy. These roles consist of the main actor, facilitator, and therapist. Main actors are the individuals whose issues are being explored by the group, and facilitators are those who help the main actor with this e~ploration.~ In particular, we believe that the main actors may play an especially influential role in setting the tone for the group. Therefore, there may be some main actor behaviors that are more conducive than others to the development of group cohesiveness. This analysis is designed to illustrate how this methodology might work and how the IGIPS-I11 could contribute to this type of inquiry. It also provides an opportunity to explore whether clinical impressions of the session are accurately represented in the process. Clinically, members of the HCHP-MHRP felt that one of the notable elements of the third session from Group A was the difference be"To be considered a main actor, it

is necessary that the group focus on the individual's issue for a somewhat sustained period. Previous work with the concept suggests that main factor ratings can be made with extremely high reliability (Soldz, Budman, Demby, & Feldstein, 1990).

INDIVIDUAL GROUP MEMBER INTERPERSONAL PROCESS SCALE

299

tween the first and second half.5 The second half of the session appeared to be especially constructive. Many of the group members worked toward an understanding of their families of origin and explored how these family dynamics may have influenced their present difficulties. Most importantly, these efforts were cooperative. Most of the group members shared their experiences and related them to the overall theme of the discussion. The discussion appeared to be affectively stimulating to most group members. On the GCS, the second half of the session would receive a high cohesiveness rating. The early part of the session would be characterized differently. The discussion was focused primarily on the concerns of Joe. Because he was interested in addressing events related to the group, the topic was necessarily relevant to the other group members. However, the other group members were unable to engage themselves in the discussion with the same amount of energy that was exhibited later. In addition, the focus of the discussion was often tenuous and unclear. Confusion and miscommunication resulted. Although the first half of the session may have been important for Joe, the majority of the group members were unable to engage in significant therapeutic work. O n the GCS, the early part of the session would receive a midrange rating. To illustrate these general differences between the early and later part of the session, we present the scores for eight basic IGIPS-I11 reduced items in Tables 11.3 and 11.4. The individual group member scores for both parts of the session are divided into three categories according to role: main actor, facilitator, and therapist. The aggregated IGIPS-I11 scores and approximate cohesiveness levels for both session halves are also presented. The IGIPS-111 scores accurately represented general clinical impressions. According to the aggregated group IGIPS scores, for example, many of the behaviors commonly associated with productive psychotherapy, such as the disclosure of personal material, the expression of affect, the achievement of insight, and the experience of connectedness, were all higher at the group level in the second half of the session than in the first. In addition, individual member scores represented the different styles of presentation that characterized the main actors. Joe, the only group member who would qualify as a main actor in the first half of the session, discussed his problems in a highly abstract and general manner. This earned him high scores on the reduced IGIPS item “Discusses Abstract Issues.” In addition, by simply attributing the conflict he was addressing to “personality differences,” Joe made it difficult for OGMs or the therapists to relate *The 13th statement from page 98 of the provided transcript was used as the dividing line between the first and second half of Session 3. This particular statement was selected because it is almost exactly in the middle of the session and it is where Diane began the transition to main actor. Also note that the last one-and-a-half pages of the transcript could not be rated because they were not on the audiotape.

300


84 0.19 0.19 2.7 3.9 -0.03 3.0 2.7

234 0.07 0.18 2.8 3.8 0.13 2.9 1.2

Amount spoken Discusses abstract issues Discusses group process Discloses personal material Expresses affect Feels connected to group Insight into own issues Insight into OGM issues

120 0.06 0.13 2.9 3.5 0.11 3.2 1.7

Martha

14 0.00 0.80 1 .o 2.4 0.00 1 .o 7.3

Brad

55 0.26 0.20 3.8 3.0 0.02 2.7 3.5

Joe

Facilitators

0

-

Pat

13 0.00 0.33 0.00 2.4 0.00 0.00 3.5

Helen

Group as a whole

589 0.09 0.21 2.8 3.4 0.08 2.9 2.9

Alice

69 0.02 0.26 0.00 2.6 0.00 0.00 3.5

Therapists

Note. Approximate group cohesiveness rating: high level. Dashes indicate no data. IGlPS = Individual Group Member Interpersonal Process Scale; OGM = other group member.

Greg

Diane

Reduced items

Main actors

TABLE 11.4 IGIPS Scores From the Second Half of Session 3 of Group A

or respond to the issues he was raising. This was reflected in the relatively low facilitator and therapist scores on the reduced item “Insight Into OGM Issues.” Moreover, Joe’s presentation was characterized by a variety of other qualities that appeared to hinder group cohesiveness. As illustrated by his low score on the item “Expresses Affect,” Joe discussed his issues in a flat monotonic voice. Also, although Joe was the main actor for a sustained period of time, the kind of personal information that he disclosed was not highly revealing for someone in the role. This became evident when comparing his score on the reduced item “Discloses Personal Material” with the main actor scores from the second half of the session. The content of Joe’s presentation may have also influenced cohesiveness. One of the main themes that Joe addressed as a main actor was his perception of being different from the rest of the group. This was observed in his negative score on the item “Feels Connected to the Group.” Such expressions of disconnectedness can be understood as an obstacle to positive group bonding, especially early in the group, because they are likely to make group members feel their experiences are irrelevant to one another. In the second half of the session, Diane clearly functioned as the catalyst in establishing a new tone. Initially, Diane alluded to her personal issues in the abstract and general manner that was congruent with the presentation style established earlier in the session. With encouragement from other group members, however, Diane ultimately transcended this mode of expression. She was able to delve into a revealing and detailed discussion of painful childhood experiences and their impact on her life. Moreover, unlike Joe, Diane related her story in a spontaneous and affective manner. The difference between Diane and Joe is clearly visible in their main actor scores. As a main actor, Diane scored notably higher than Joe on “Discloses Personal Material,” “Expresses Affect,” “Feels Connected to Group,” and “Insight Into Own Issues,” whereas she scored much lower on “Discusses Abstract Issues.” The role Diane played in transforming the group atmosphere was implicated in the subsequent main actor episodes of the other group members. Both Greg and Martha were notably responsive to Diane’s presentation and used it as a vehicle to compare and contrast their own personal histories. In addition, both of them appeared to emulate her style of presentation on several important dimensions, including topic, degree of affective expression, level of disclosed personal material, and depth of insight. The similarity in their presentation style was reflected in their main actor scores. Interestingly, Joe also appeared to adopt the new presentation style initiated by Diane and sustained by Greg and Martha. Although too briefly and sporadically to qualify as main actor, Joe did discuss his own issues during the second half of the session. As illustrated by his IGIPS scores, Joe’s behavior during these episodes was markedly different from his pre302

DAVIS, BUDMAN, A N D SOLD2

sentation in the beginning. Respectively, his scores on “Discloses Personal Material,’’ “Expresses Affect,” “Feels Connected to Group,’’ “Insight Into Own Issues,” and “Insight Into OGM Issues” were 1.3, 1.5, -0.36, 2.2, and 1.8, respectively, in the first half, compared with 3.8, 3.0, 0.02, 2.7, and 3.5 in the second half. Joe appeared to adopt the others’ presentation style and, as a result, he became notably more vulnerable, emotional, connected, and insightful. The differences between the two halves of the session can also be understood in terms of stage-related theories of group development that the HCHP-MHRP was interested in exploring. According to these theories, there are stages during group development in which it is most effective to engage the group members around specific tasks. Process-oriented interventions or discussions, for example, are theorized to be more useful late in the group, whereas a focus on personal history and life circumstances are thought to more useful in the beginning of the group. The usefulness of process behavior in understanding personal problems is difficult to discern without sufficient knowledge of personal history. Thus, groups that focus on process issues too early, without first becoming sufficiently familiar with background information, increase the risk of shallow and irrelevant feedback. This idea was illustrated by the early Group A session analyzed here. The process-focused discussion in the first half of the session, which received a .61 score on the aggregated reduced item “Discusses Group Process,” was clearly less effective in engaging the group members than the historically focused discussion of the second half, which received a 21 score on the same item. Moreover, as these discussions were both initiated by the principal main actors-Joe and Diane-it provides an additional example of how main actors can influence the group experience.

ASSESSMENT AND FUTURE DIRECTIONS The IGIPS is a scale that can be used in a variety of ways. The scale has the potential to address research questions at the group, individual, and statement levels. The scale is also sensitive to the roles associated with most forms of group therapy, such as main actor, facilitator, and therapist. In addition, as the scale records the sequence of each statement and the individual to whom it is directed, research inquiries concerning interactive group processes becomes possible. Given this potential arsenal of information, there are many important group research endeavors to which the IGIPS can contribute. These include the relationship between interventions and responses, between individual level and group level behavior, between personality and group process, or between group process and outcome. Finally, although the IGIPS was derived from a particular theoretical orientation, many of the items are general enough to be used by research INDIVIDUAL GROUP MEMBER INTERPERSONAL PROCESS SCALE

303

teams with different theoretical perspectives, thereby making it potentially useful at other research sites.

REFERENCES American Psychiatric Association. ( 1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC:Author. Bakeman, R. ( 1990). Sequential analysis and psychotherapy research. Unpublished manuscript, Georgia State University. Beck, A. P., Dugo, J. M., Eng, A. M., & Lewis, C. M. (1986). Analysis of group development. In L. S. Greenberg & W. M. Pinsof (Eds.), The psychotherapeutic process: A research handbook (pp. 615-705). New York: Guilford Press. Budman, S. H., Demby, A., Feldstein, M., Redondo, J., Scherz, B., Bennett, M. J., Koppenaal, G., Sabin Daley, B., Hunter, M., & Ellis, J. (1987). Preliminary findings on a new instrument to measure cohesion in group psychotherapy. lnternational journal of Group Psychotherapy, 37, 75-94. Budman, S. H., Demby, A., Redondo, J. P., Hannan, M., Feldstein, M., Ring, J., & Springer, T.(1988). Comparative outcome in time-limited individual and group psychotherapy. International j o u d of Group Psychotherapy, 38, 63-86. Budman, S. H., & Gurman, A. S. (1988). Theory and practice of brief therapy. New York: Guilford Press. Budman, S. H., Soldz, S., Demby, A., Davis, M. S., & Merry, J. (1993). What is cohesiveness?An empirical examination. Small Group Research, 24, 199-216. Budman, S. H., Soldz, S., Demby, A., Feldstein, M., Springer, T., & Davis, M. S. ( 1989). Cohesion, alliance and outcome in group psychotherapy. Psychiatry, 52, 339-350. Connelly, J. L., Piper, W. E., & Braha, R. (1981). The Therapeutic Work Rating System (TWS) manual. Unpublished manuscript, McGill University, Montreal, Quebec, Canada. Costa, P. T., 6r McCrae, R. R. (1990). Personality disorders and the five-factor model of personality. Journul of Personality Disorder, 4, 362-371. Gilligan, C. (1982). In a differentvoice. Cambridge, MA: Harvard University Press. Gomes-Schwartz, B., & Schwartz, J. M. (1978). Psychotherapy process variables distinguishing the “inherently helpful” person from the professional psychotherapist. Journal of Consulting and Clinical Psychology, 46, 196-197. Gottman, J. M., & Bakeman, R. (1986). Observing interaction: An introduction to sequential analysis. New York: Cambridge University Press. Gottman, J. M., & Roy, A. K. (1990). Sequential analysis: A gulde for behavioral researchers. New York: Cambridge University Press. John, 0. P. (1990). The “Big Five” factor taxonomy: Dimensions of personality in the natural language and in questionnaires. In L. A. Pervin (Ed.), Handbook of personality: Theory and research (pp. 66-100). New York: Guilford Press.

304

DAVIS, BUDMAN, AND S O D 2

Kaul, T., & Bednar, R. (1986). Research on group and related therapies. In S. Garfield & A. Bergin (Us.),Handbook of psychotherapy and behavior change (pp. 671-714). New York: Wiley. Klein, M. H., Mathieu, P. L., Gendlin, E. T., & Kiesler, D. J. (1969). The Experiencing Scale: A research and training m n u a l (Vol. l). Madison: University of Wisconsin, Extension Bureau of Audiovisual Instruction. Klein, M. H., Mathieu-Coughlan, P., & Kiesler, D. J. (1986). The Experiencing Scale. In L. S. Greenberg & W. M. Pinsof (Eds.), The psychotherapeutic process: A research handbook (pp. 21-71). New York: Guilford Press. Lewis, H. B. (1976). Psychic war in men and women. New York: New York University Press. Lyons, M. J., Ozer, D. J., Young, L., Merla, M. E., & Hyler, S. E. (1991). Relationship of the five factor model of personality to DSM-111 personality disorders. Unpublished manuscript, Boston University. MacKenzie, K. R. (1990). Introduction to time-limited group psychotherapy. Washing ton, DC:American Psychiatric Press. McCrae, R. R. (1989). W h y I advocate the first-factor model: Joint analyses of the NEO-PI and other instruments. In D. M. Buss & N. Cantor (Eds.), Personality psychology: Recent trends and emerging directions (pp. 237-245). New York: Springer-Verlag. McCrae, R. R., & Costa, P. T.(1990). Personality in adulthood. New York: Guilford Press. Millon, T. (1987). Manual for MCMI-II. Minneapolis, MN: National Computer Systems. O’Malley, S. S., Suh, C. S., & Strupp, H. H. (1983). The Vanderbilt Process Scale: A report on the scale development and a process-outcome study. Journal of Consulting and Clinical Psychology, 51, 581-586. Rounsaville, B. J., Chevron, E. S., Prusoff, B. A., Elkin, I., Imber, S., Sotsky, S., & Watkins, J. (1987). The relation between specific and general dimensions of the psychotherapy process in interpersonal psychotherapy of depression. Journal of Consulting and Clinical Psychology, 55, 379-384. Soldz, S., Budman, S. H., Davis, M. S., & Demby, A. (1993). Beyond the interpersonal circumplex in group psychotherapy: The structure and relationship to outcome of the Individual Group Member Interpersonal Process Scale. journal of Clinical Psychology, 49, 551 -563. Soldz, S., Budman, S. H., & Demby, A. (1992). The relationship between main actors behaviors and treatment outcome in group psychotherapy. Psychotherapy Research, 2, 52-62. Soldz, S., Budman, S. H., Demby, A., & Feldstein, M. (1990). Patient activity and outcome in group psychotherapy: New findings. International Journal of Group Psychotherapy, 40, 53-62. Soldz, S., Budman, S. H., Demby, A., & Merry, J. (1993). Diagnostic agreement between the personality disorder examination and MCMI-11. journal of Personality Assessment, 60(3), 486-499. INDIVIDUAL GROUP MEMBER INTERPERSONAL PROCESS SCALE

305

Spence, J. T., & Helmreich, R. L. (1978). Masculinity and femininity: Their psychological dimensions, correlates, and antecedents. Austin: University of Texas Press. Suh, C. S., Strupp, H. H., & O'Malley, S. S. (1986). The Vanderbilt process measures: The Psychotherapy Process Scale (VPPS) and the Negative Indicators Scale (VNIS). In L. S. Greenberg & W. M. Pinsof (Eds.), The psychotherapeutic process: A research handbook (pp. 285-323). New York: Guilford Press. Wiggins, J. S. (1979). A psychological taxonomy of trait-descriptive terms: The interpersonal domain. l o u d of Personality and Social Psychology, 37, 395412. Wiggins, J. S. (1982). Circumplex models of interpersonal behavior in clinical psychology. In P. C. Kendall & J. N. Butcher (Eds.), Handbook of research methods in clinical psychology (pp. 183-221). New York: Wiley. Wiggins, J. S., & Pincus, A. L. (1989). Conceptions of personality disorders and dimensions of personality. Psychological Assessment, 1 , 305-3 16. Windholz, M. J., & Silberschatz, G. ( 1988). Vanderbilt Psychotherapy Process Scale: A replication with adult outpatients. Journal of Consulting and Clinical Psychology, 56, 56-60. Yalom, I. D. (1985). The theory and practice of group psychotherapy (3rd ed.). New York: Basic Books.

306


APPENDIX 11.A PRINCIPAL-COMPONENTS FIVE-FACTOR STRUCTURE OF 12 REDUCED IGIPS-I ITEMS First Factor: Activity How Often Spoke Discloses Personal Material Expresses Affect Second Factor: Interpersonal Sensitivity Makes Connections With Other Group Members Sensitivity When Addressing Other Group Members Positive-Negative Sentiment Third Factor: Comfort With Self Responsivenessto Input Self-Esteem Input Received When Self-Focused Fourth Factor: Self-Focus Discusses Own Issues Fifth Factor: Psychologically Minded Psychologically Minded When Self-Focused Psychologically Minded When Addressing Other Group Members Note. Component loadings were subjected to varimax rotation. IGIPS-I = Individual Group Member Interpersonal Process Scale I.

1NDIVIDUAL. GROUP MEMBER INTERPERSONAL PROCESS SCALE

307

APPENDIX 1l.B METHOD OF CALCULATING REDUCED ITEMS PRESENTED IN TABLES 11.3 AND 11.4 1. Amount Spoken Calculated by summing all the scores on the “Length of Statement” item. Each 5 s of speaking represents a rating of 1. 2. Discusses Abstract Issues Represents the percentage of overall statements containing reference to impersonal or abstract issues. 3. Discusses Group Process Represents the percentage of overall statements that contain an in-group designation rating on any item. 4. Discloses Personal Material Calculated by summing the significance ratings on the item “Discloses Personal Material,” and then dividing that sum by the sum of frequency ratings on “Discusses Self.” For purposes of this presentation, the outside-group and inside-group designations were ignored. 5. Expresses Affect Calculated by subtracting the summed significance ratings of “Resistance to Affect” from the summed significance ratings of “Expresses Affect“ and dividing the result by the overall number of statements. 6. Feels Connected to Group Calculated by subtracting the summed significance ratings of “Feels Disconnected From Others” with inside-group designations from the summed ratings of “Feels Connected to Others” with inside-group designations and dividing the result by the overall number of statements. 7. Insight Into Own Issues Calculated by summing the significance ratings on “Demonstrates SelfAwareness” and dividing by the sum of frequency ratings on “Discusses Self.” Outside-group designations and inside-group designations were ignored. 8. Insight Into OGM Issues Calculated by summing the significance ratings on “Asks Questions,” “Enhances OGM Awareness,” and “Gives Advice” and dividing by the sum of frequency ratings on “Discusses OGMs.” Outside-group designations and inside-group designations were ignored. Note. OGM = other group member.

308

DAVIS, BUDMAN,AND SOLD2