The TAT Affect Scale and the Inventory of Interpersonal Problems: How Problems Relate to Interpersonal Affect

Abstract: 

This study examined the relationship between the TAT Affect Scoring System and the IIP circumplex in a sample of 167 college students. The TAT Affect Scoring System is designed to measure the affect between characters in Thematic Apperception Test (TAT) stories (Thomas and Dudek, 1985). The Inventory for Interpersonal Problems (IIP) is designed to identify the most common interpersonal problems reported by psychotherapy clients (Horowitz, Rosenberg, Baer, Ureno, & Villasenor, 1988). This study used the IIP circumplex model as developed by Alden, Wiggins and Pincus (1990). Theoretically, both the interpersonal problems reported in the IIP and the affect between characters of the stories told in the TAT are related to a person’s interpersonal attachment. These associations were not significant.

Table of Contents: 

    Introduction

    The Thematic Apperception Test (TAT) is a projective psychological measure that consists of a series of ambiguous pictures. Projective measures are designed to measure an individual’s feelings and attitudes toward a topic or relationship without directly asking the individual. Projective measures are useful because an individual may try to hide his/her feelings if questioned directly or he/she may be unaware of his/her feelings or of a way to express them in words. The TAT uses ambiguous drawings to elicit stories from individuals about the scene and the characters within the scene.

    Thomas and Dudek (1985) hypothesized that interpersonal affect attributed to the characters in the TAT pictures reflects the quality of affect toward characters in similar roles in real life. They developed an Interpersonal Affect Scale, scored from TAT stories, to study spousal relationships. They hypothesized that married partners with the spouse category as their most positive relationship would have higher marital adjustment scores on the Dyadic Adjustment Scale (Spanier, 1976) than would married couples who had the parent or nonspouse category as their most positive relationship. Stories with positive Spouse affect might contain phrases describing what a wonderful marriage the couple had or how much the husband and wife loved each other. Stories with positive Nonspouse affect might say the same thing except that the people are described as friends instead of partners. Stories with positive Parent affect would involve a positive child and parent interaction. Thomas and Dudek (1985) believe that individuals who are properly adjusted to the spousal role also would be separated from the parent-child relationship. Therefore, couples with higher marital adjustment scores, as indicated by the Dyadic Adjustment Scale, would tend to tell stories with characters who also are appropriately adjusted to the spousal role. This hypothesis was supported by their research.

    Beyond investigations of marital adjustment, this scoring system appears to be effective for evaluating other interpersonal relationships. However, to this researcher’s knowledge, no further research has used the TAT Affect Scoring System. The present study looks at the relationship between the TAT Affect Scoring System and the Inventory of Interpersonal Problems as both pertain to attachment theory.

    Horowitz, Rosenberg, Baer, Ureno and Villasenor (1988) developed a measure “designed to help patients and therapists identify interpersonal sources of distress that are often the focus of psychotherapy.” The IIP is a self-report inventory that lists 127 different types of interpersonal problems that people may experience, divided into two groups of ways patients report problems: “It is hard for me to…,” and “These are things I do too much.” The respondent is asked to rate each problem on how distressing it is for them on a scale from 0 (not at all) to 4 (extremely). Through factor analysis, a general Complaint factor and six subscales (Hard to be Submissive, Hard to be Assertive, Too Controlling, Hard to be Sociable, Too Responsible, and Hard to be Intimate) were developed.

    Taking the lead from previous research in the area of interpersonal traits (Wiggins, 1979; Conte & Plutchik, 1981; Kiesler, 1983), Alden, Wiggins and Pincus (1990) developed a circumplex model of the IIP subscales. The IIP Circumplex model consists of eight regions (Domineering, Vindictive, Cold, Socially Avoidant, Nonassertive, Exploitable, Overly Nurturant, and Intrusive) plotted on a graph of dominance (power) and nurturance (affiliation) to form a circle. Out of all the different IIP models, the Circumplex model seems to be the most used. The Gude et al. (2000) model seems to have oversimplified the model, while the Barkham, Hardy and Startup (1994) model results in a larger number of scales to analyze. Additionally, Horowitz has used the Alden et al. (1990) Circumplex model in more recent research regarding the IIP (Horowitz, Rosenberg & Bartholomew, 1993).

    Attachment theory has long been a subject of study to psychologists. Studies suggest that there is a strong causal relationship between a person’s early interactions with their parents and that person’s later ability to make affectional bonds (Bowlby, 1977). Bartholomew (1990) focuses on adults who consistently avoid intimacy. Bartholomew differentiates avoidant adults into two groups: those who want to form attachments but avoid them out of fear and those who say they neither fear nor desire attachments. Bartholomew then proposes a new model of adult attachment based on positive and negative views of the self and others. This results in four categories of attachment: secure (positive view of self and positive view of others), preoccupied (negative view of self and positive view of others), dismissing (positive view of self and negative view of others) and fearful (negative view of self and negative view of others). According to this model, both dependency and avoidance can vary independently of each other. Research by Bartholomew and Horowitz (1991) supported this model.

    In an attempt to combine attachment theory and the study of interpersonal problems, Horowitz, Rosenberg and Bartholomew (1993) investigated the relationship between Bartholomew’s (1990) adult model of attachment and the Inventory of Interpersonal Problems using the IIP circumplex model (Alden et al., 1990; Horowitz et al., 1988). This study found that secure individuals, as defined by the adult attachment model, reported no extreme interpersonal problems. Dismissing individuals reported more interpersonal problems in the Cold and Competitive segments, while Fearful individuals reported more problems dealing with passivity (Introverted, Subassertive and Exploitable segments). Perhaps surprisingly, Preoccupied individuals reported more problems in the area of Expressiveness.

    Based on all of the research, there appears to be an important relationship between an individual’s attachment, his/her later interpersonal problems, and the affect they attribute to story characters. As demonstrated by Horowitz, Rosenberg and Bartholomew (1993), the attachment styles of Dismissing, Fearful and Preoccupied have the highest correlations with the Cold, Socially Avoidant (Introverted) and Intrusive (Expressive) subscales of the IIP Circumplex, respectively. Conceptually, the Dismissing individual does not need nor want to be permanently attached to anyone and therefore presents with interpersonal problems on the Cold subscale. An individual scoring high on the Cold subscale might also tell stories with characters who act similarly with regard to interpersonal interactions. Both the individuals and the characters they describe would reject others in general and might specifically reject the idea of marriage and the spousal role, resulting in negative spousal scores. Similarly, Fearful individuals, though desirous of relationships, actively avoid others for fear of embarrassment or rejection and therefore score high on the Socially Avoidant (Introverted) subscale. Through past experiences the Socially Avoidant (Introverted) person has developed an overall negative outlook on all interactions with other people. These individuals will probably tell stories reflecting this negative attitude toward others. Additionally, Preoccupied people tend to score highest on the Intrusive/Expressive subscale. It seems that these individuals view their peers in such high regard that they try to attract others to them by drawing attention to themselves. Those with Intrusive (Expressive) interpersonal problems will probably tell stories about the positive interpersonal interactions they attempt to attract. Consequently, there may be a relationship between various scales of the IIP and specific category scores of the TAT Affect System because both are related to an individual’s attachment style.

    Additional predictions of correlation between the TAT Affect Scoring System and the IIP Circumplex are expected, though not necessarily directly related to a specific attachment style. Individuals with interpersonal problems in the Nonassertive/Subassertive subscale may be in a submissive role as a result of interactions with controlling parents in early childhood. These people might tell stories that involve parents viewed negatively. People who present with problems on the Exploitable subscale tend to be very trusting and may assume a dependent role. These individuals want and need to believe in the fairy-tale of true love where someone will forever care for them, so they would tell stories of loving, happy relationships with spouses. Those scoring high in the Overly Nurturant subscale try to take care of others, even to their own detriment. They will tell stories that hold the role of a parent or caregiver in very high regard, yielding positive parent scores.

    To summarize, the following hypotheses will be tested:

    1. The net spousal score will be negatively correlated with the Cold subscale.
    2. The net spousal score will be positively correlated with the Exploitable subscale.
    3. The net nonspousal score will be negatively correlated with the Socially Avoidant/Introverted subscale.
    4. The net nonspousal score will be positively correlated with the Intrusive/Expressive subscale.
    5. The net parental score will be negatively correlated with the Nonassertive/Subassertive subscale.
    6. The net parental score will be positively correlated with the Overly Nurturant subscale.

    No specific relationship is expected between the TAT Affect System and the IIP subscales of Domineering and Vindictive. Other relationships may exist between these measures and any such results will be investigated and discussed.

    Methods

    Participants

    This study used data previously collected by Kristin Niemeyer (2000). The sample included 167 undergraduate students enrolled in three different psychology courses at a large southern university. Participants included 122 (73.1%) females and 45 (26.9%) males. All were selected to be between the ages of 18 and 22 years and unmarried. Seventy percent (117) were White, 12.6% (21) were Black, 7.2% (12) were Asian, 6% (10) were Hispanic, and 4.2% (7) were of other ethnicities.

    Measures

    TAT Affect Scoring System. Interpersonal affect was measured by the TAT Affect Scoring System as designed by Thomas and Dudek (1985). The scoring system measures affect by first placing the characters of stories into categories of “spouse,” “nonspouse,” “parent,” “other,” and “nonattributed affect.” Then, the quality of the affect toward that character is scored on a Likert type 4-point scale from -2 for strong negative feelings to +2 for strong positive feelings (0 is used for a lack of affect within the story). The weighting of affective scores is based on the stimulus pull of the picture and the most frequent responses for each card, as well as the emotional content of the response (Thomas, in press). Scores were summed (positive plus negative), resulting in scores referred to as the “net category score” for each category: spouse, nonspouse, and parent. Test-retest reliability was .88 and interscorer reliability ranged from .93 to .97 (Thomas & Dudek, 1985). For the present study, all stories were scored independently by two scorers who were trained using the Thomas (in press) scoring manual and practice materials, and disagreements were discussed and resolved.

    Inventory of Interpersonal Problems. The Circumplex Scales for the IIP as developed by Alden, Wiggins and Pincus (1990) were used in this study. Alden et al. used ipsatized scores during analysis to eliminate the general factor labeled in the original study (Horowitz et al., 1988). Taking data from two normal university populations (N = 197, N = 273), Alden et al. empirically derived a circumplex model of interpersonal problems and then tested it on a third university sample (N = 974). Cronbach’s Alpha was calculated for each of the subscales (Domineering = .77, Vindictive = .80, Cold = .81, Socially Avoidant = .85, Nonassertive = .85, Exploitable = .82, Overly Nurturant = .76, Intrusive = .72). Intercorrelations among the subscales resulted in moderate negative correlations between subscales located on opposite sides of the circumplex (convergent validity) and near zero correlations between subscales located at right angles from one another (discriminant validity). In this study, Cronbach’s Alpha for the subscales ranged from .78 to .84. To test the hypotheses, an individual’s IIP scores were ipsatized by subtracting the individual’s IIP mean score from each subscale score. This allows for each subscale to be compared to every other subscale for that individual. The greatest score above the mean (the highest ipsatized score) represents the highest scoring subscale for that individual and, therefore, the interpersonal area of greatest trouble.

    Procedure

    Six group-administered TAT stories (cards 1, 2, 3BM, 4, 10 and 13MF) were handwritten by all students who chose to participate in a class demonstration of group TAT administration. Each student was given the opportunity to complete an additional questionnaire for extra credit in his or her psychology course. A letter was attached to the front of each packet that stated that return of the packet implied informed consent. Students were asked to return the packet to the researcher or an assistant during a later class meeting or any time in the following four weeks. A numeric code was assigned for each participant’s materials to ensure anonymity and to permit matching of story protocols with questionnaires.

    Hypothesis-Testing Analyses

    Hypothesis 1: The net spousal score will be negatively correlated with the ipsatized Cold subscale. This hypothesis was tested by a Pearson correlation between the ipsatized Cold subscale and the negative spousal category score of the TAT Affect Scoring System.

    Hypothesis 2: The net spousal score will be positively correlated with the ipsatized Exploitable subscale. This hypothesis was tested by a Pearson correlation between the ipsatized Exploitable subscale and the positive spousal category score of the TAT Affect Scoring System.

    Hypothesis 3: The net nonspousal score will be negatively correlated with the ipsatized Socially Avoidant/Introverted subscale. This hypothesis was tested by a Pearson correlation between the Socially Avoidant/Introverted subscale and the negative nonspousal category score of the TAT Affect Scoring System.

    Hypothesis 4: The net nonspousal score will be positively correlated with the ipsatized Intrusive/Expressive subscale. This hypothesis was tested by a Pearson correlation between the ipsatized Intrusive/Expressive subscale and the positive nonspousal category score of the TAT Affect Scoring System.

    Hypothesis 5: The net parental score will be negatively correlated with the ipsatized Nonassertive/Subassertive subscale. This hypothesis was tested by a Pearson correlation between the ipsatized Nonassertive/Subassertive subscale and the negative parental category score of the TAT Affect Scoring System.

    Hypothesis 6: The net parental score will be positively correlated with the ipsatized Overly Nurturant subscale. This hypothesis was tested by a Pearson correlation between the ipsatized Overly Nurturant subscale and the positive parental category score of the TAT Affect Scoring System.

    Results

    Descriptive Statistics

    Descriptive statistics for each variable are listed in Table 1. Alden et al. (1990) developed and tested the circumplex model of the IIP on undergraduates at the University of British Columbia. Though the present sample is not from Canada, there should be little difference between the two populations on this measure. Regarding the TAT Affect Scoring System, Thomas and Dudek (1985) studied married spouses with ages ranging from 28 to 57 years. The present study, however, only includes unmarried students ranging in age from 18 to 22 years. Since the TAT Affect Scoring System is not being used in this study to determine marital satisfaction, the difference in marital status should not affect this study. In contrast, this study is the first to use the Affect Scoring System on a younger population. Differences between the two populations, however, should be most apparent in the attitudes expressed within the story due to lack of maturity and experience, but should not affect the way in which the stories are scored.

    Examination of the frequency distributions for all variables showed approximately normal distributions with no outliers. The relationships between the TAT net category scores were investigated. The net Nonspouse score was positively and significantly related to the Spouse category score (r = .166, p = .033). No significant relationship was found between the net Nonspouse and net Parent scores nor the net Spouse and net Parent scores. Relationships between the ipsatized IIP scores were investigated. The results are listed in Table 2.

    Hypothesis Testing

    Each hypothesis was tested using a Pearson product-moment correlation. It was hypothesized that:

    1. The net spousal score will be negatively correlated with the Cold subscale. This relationship was not significant (r = .012, p = .877).
    2. The net spousal score will be positively correlated with the Exploitable subscale. This relationship was not significant (r = -.046, p = .561).
    3. The net nonspousal score will be negatively correlated with the Socially Avoidant/Introverted subscale. This relationship was not significant (r = -.116, p = .137).
    4. The net nonspousal score will be positively correlated with the Intrusive/Expressive subscale. This relationship was not significant (r = .057, p = .465).
    5. The net parental score will be negatively correlated with the Nonassertive/Subassertive subscale. This relationship was not significant (r = -.026, p = .745).
    6. The net parental score will be positively correlated with the Overly Nurturant subscale. This relationship was not significant (r = .108, p = .169).

    Exploratory Analysis

    Further analysis revealed no significant correlations between any of the ipsatized IIP scores and the TAT category scores. Correlations between all experimental variables were run separately for males and females with the results listed in Table 3. Partial correlations were conducted on all of the above analyses, with gender and ethnicity partialled out. No significant changes were found. Correlations between demographics and each variable can be found in Table 4.

    Discussion

    The statistical evidence did not support the hypothesized relationships between TAT affect and interpersonal problems. One possible explanation for the lack of significance is that over 70% of the participants were female. Bartholomew and Horowitz (1991) found that female subjects received significantly higher scores in the preoccupied attachment style, which they then linked to interpersonal problems in the Intrusive and Overly Nurturant subscales. However, in this study when gender was controlled, there remained no significant relationship between interpersonal problems and affect in TAT stories. Another possible explanation is that this population was both young and unmarried. The fact that all the participants were unmarried might affect their outlook on the spousal relationship and therefore affect the types of spousal stories they tell. Additionally, youth and the lack of more life experiences may affect the kinds of interpersonal problems experienced and reported, as well as affecting the types of TAT stories told.

    The present study found a significant relationship between affect toward spouse and nonspouse characters but no significant relationship between affect toward parent characters and affect toward spouse or nonspouse characters. The relationship between these TAT category scores makes sense when compared with the Bartholomew and Horowitz (1991) study. They found that only peer ratings accounted for variation on both the dominance and affiliative aspects of an individual’s interpersonal problems, whereas they found that parent ratings accounted for variation only on the affiliative axis of the graph of interpersonal problems. These results suggest that although individuals are consistent in their behavior toward their peers (spouse and nonspouse individuals), interpersonal behavior toward parents is determined by something else. An individual chooses the peers he/she associates with, but not the family into which he/she is born. Since a child is born dependent upon parents or caregivers, the dominance aspect of behavior between a parent and child is largely predetermined. At some point in life, the child outgrows the submissive role and conflict may erupt. Since this study dealt with college students, it is likely that most are currently in a gray area between dominance and submission with their parents and the results reflected this by the lack of relationship between parent affect and peer affect.

    The ipsatized IIP scores used in this study were generally consistent with Alden et al.’s (1990) original design of the circumplex. The ipsatized scores of each subscale were most negatively related to their polar opposite and least related to those subscales which were located 90 degrees in either direction. This supports the circular arrangement of the circumplex design.

    One strength of this study was the large sample size. However, over 70% of the participants were female and all were unmarried and currently enrolled in psychology courses. These things may limit the generalizability of these results. Although gender did not appear to affect the lack of correlation, there were some differences in the types of problems reported and the types of stories told between males and females. Specifically, the intercorrelation of the IIP subscales for females exhibited more of the expected correlations according to the circumplex model. This is probably due to the lack of a large sample of males.

    A primary limitation of this study was conducting a secondary analysis and therefore being limited to the previously administered measures. Since attachment style was not originally measured, the relationship of the TAT and the IIP to an individual’s attachment style could not be investigated directly, but only inferred. Another limitation is that all subjects had independently elected to take psychology courses prior to the study. There could be differences in interpersonal relationships between people who study psychology versus those who study other disciplines.

    Future research should be conducted in this area to investigate the relationships between attachment style, interpersonal problems and affect in TAT stories. A study with a more gender balanced participant sample may yield different results. Additionally, the participant pool might contain both married and unmarried individuals to determine if there is a difference in responding on the TAT Affect Scale.

    This study was the first investigating the relationship between an individual’s interpersonal problems and the types of TAT stories they tell. Further research on this topic may result in a greater understanding of affect in TAT stories. It also is important to determine how attachment affects an individual’s interpersonal problems in order to properly treat the problem at its source. In the future, additional assessment measures can be developed to allow for better reporting of the interpersonal problems people face.

    References

    • Alden, L. E., Wiggins, J. S., & Pincus, A. L. (1990). Construction of circumplex scales for the Inventory of Interpersonal Problems. Journal of Personality Assessment, 55, 521-536.
    • Bartholomew, K. (1990). Avoidance of intimacy: An attachment perspective. Journal of Social and Personal Relationships, 7, 147-178.
    • Bartholomew, K. & Horowitz, L. M. (1991). Attachment styles among young adults; A test of a model. Journal of Personality and Social Psychology, 61, 226-244.
    • Bowlby, J. (1977). The making and breaking of affectional bonds. British Journal of Psychiatry, 130, 201-210.
    • Conte, H. R. & Plutchik, R. (1981). A circumplex model for interpersonal personality traits. Journal of Personality and Social Psychology, 40(4), 701-711.
    • Gude, T., Moum, T., Kaldestad, E., & Friis, S. (2000). Inventory of interpersonal problems: A three-dimensional balanced and scalable 48-item version. Journal of Personality Assessment, 74(2), 296-310.
    • Horowitz, L. M., Rosenberg, S. E., Baer, B. A., Ureno, G., & Villasenor, V. S. (1988). Inventory of interpersonal problems: Psychometric properties and clinical applications. Journal of Consulting and Clinical Psychology, 56(6), 885-892.
    • Horowitz, L. M., Rosenberg, S. E., & Bartholomew, K. (1993). Interpersonal problems, attachment styles, and outcome in brief dynamic psychotherapy. Journal of Consulting and Clinical Psychology, 61(4), 549-560.
    • Kiesler, D. J. (1983). The 1982 interpersonal circle: A taxonomy for complementarity in human transactions. Psychological Review, 90, 185-214.
    • Spanier, G B. (1976). Measuring dyadic adjustment: New scales for assessing the quality of marriage and similar dyads. Journal of Marriage and the Family, 38(1), 15-27.
    • Thomas, A. D. & Dudek, S. Z. (1985). Interpersonal affect in thematic apperception test responses: A scoring system. Journal of Personality Assessment, 49(1), 30-36.
    • Thomas, A. D. (in press). Scoring manual for interpersonal affect. In S. R. Jenkins (Ed.) Handbook of clinical scoring systems for thematic apperceptive techniques, Mahwah, NJ: Lawrence Erlbaum.
    • Wiggins, J. S. (1979). A psychological taxonomy of trait descriptive terms: The interpersonal domain. Journal of Personality and Social Psychology, 37, 395-412.

    Table 1: Descriptive Statistics

    Variables M SD Range
    Net TAT Nonspouse Affect -0.6506 2.1458 -7.00 to 4.00
    Net TAT Parent Affect -.0120 2.40703 -6.00 to 8.00
    Net TAT Spouse Affect -.4699 2.03209 -7.00 to 5.00
    Ipsatized IIP Domineering -.3146 0.46304 -1.49 to 1.19
    Ipsatized IIP Vindictive -.3057 0.41670 -1.35 to 1.15
    Ipsatized IIP Cold -.2817 0.49775 -1.61 to 0.95
    Ipsatized IIP Socially Avoidant .1619 0.56351 -0.93 to 2.14
    Ipsatized IIP Nonassertive .2804 0.55898 -1.23 to 2.06
    Ipsatized IIP Exploitable .2393 0.48974 -0.92 to 1.56
    Ipsatized IIP Overly Nurturant .3241 0.56133 -0.92 to 2.10
    Ipsatized IIP Intrusive -.0600 0.52443 -1.61 to 1.54

    Table 2: Correlations between Spouse and Nonspouse TAT Net Category Scores

      Domineering Vindictive Cold Socially Avoidant Nonassertive Exploitable Overly Nurturant Intrusive
    Domineering 1              
    Vindictive .43*** 1            
    Cold .21** .51*** 1          
    Socially Avoidant -.37*** -.07 .27** 1        
    Nonassertive -.62*** -.48*** -.24*** .29*** 1      
    Exploitable -.39*** -.50*** -.49*** -.18* .42*** 1    
    Overly Nurturant -.16* -.47*** -.52*** -.35*** .03 .47*** 1  
    Intrusive .33*** .02 -.37*** -.56*** -.28*** .02 .22** 1

    *. p < .05; **. p < .01.; ***. p < .001.

    Table 3: Correlations between Experimental Variables for Males and Females

     
     
    Domineering
    Vindictive
    Cold
    Socially Avoidant
    Nonassertive
    Exploitable
    Overly Nurturant
    Intrusive
    Male
    Female
    Domineering 1
    1
                 
    Male
    Female

    Vindictive
    .42**
    .43*** 
    1
    1
               
    Male
    Female
    Cold .18
    .20*
    .48***
    .52***
    1
    1
             
    Male
    Female
    Socially Avoidant  -.32*
    -.42***
    -.29
    .02 
    .09
    .32***
    1
    1
           
    Male
    Female
    Nonassertive -.55***
    -.64***
    -.52***
    -.45***
    -.21
    -.22* 
    .35*
    .32***
    1
    1
         
    Male
    Female
    Exploitable -.43**
    -.37***
    -.40**
    -.53***
    -.41**
    -.51***
    -.20
    -.17
    .18
    .47***
    1
    1
       
    Male
    Female
    Overly Nurturant -.21
    -.12
    -.36*
    -.52***
    -.41**
    -.56***
    -.29
    -.36***
    -.08
    .04 
    .53***
    .44***
    1
    1
     
    Male
    Female
    Intrusive -.03
    .46***
    -.17
    .09
    -.56***
    -.31***
    -.26
    -.66*** 
    -.02
    -.36***
    .27
    -.05
    .16
    .24**
    1
    1

    *. p < .05.; **. p < .01; ***. p < .001.

    Table 4: Correlations between All Experimental Variables and Demographic Variables

      Net Nonspouse Score Net Parent Score Net Spouse Score Domineering Vindictive Cold Socially Avoidant Nonassertive Exploitable Overly Nurturant Intrusive
    Gender -.04 .05 -.18* -.13 -.11 -.13 -.13 .17* .11 .11 .05
    Age -.01 -.03 .09 .00 -.03 -.00 .00 .04 .04 -.09 .02
    Ethnicity .09 .05 .09 -.08 -.02 -.09 .14 .04 -.01 -.05 .04

    *. p < .05.

    AttachmentSize
    PDF icon WhiteR.pdf112.41 KB