Reliability of Variables

　A reliability check was performed to determine the internal consistency of the Intercultural Behavioral Assessment Indices and the instruments used to measure Personal Attributes, Communication Skills, Psychological Adaptation, and Cultural Awareness. Cronbach's (1951) coefficient alpha was computed for each of the following variables: selfconsciousness (including private selfconsciousness, public selfconsciousness, and social anxiety subscales),  selfdisclosure (including consciously intended disclosure, amount of disclosure, positiveness/negativeness of disclosure, depth/intimacy of disclosure, and honesty/accuracy of disclosure subscales), communication adaptability (including adaptability and rewarding impressions subscales), social situations, interaction involvement (including communication responsiveness, communication perceptiveness, and communication attentiveness subscales), cultural awareness, communication competence, and intercultural communication competence. Table 2 lists the alphas of these variables and the reliability coefficients obtained in previous research (discussed in Chapter I).

　SelfAwareness

　Selfconsciousness, as a measure of selfawareness, was operationalized via 23 statements which grouped into three dimensions. A coefficient alpha of .70 was found for private selfconsciousness, .88 for public selfconsciousness, and .77 for social anxiety. Previous research reported coefficient alphas of .79 for private selfconsciousness, .84 for public selfconsciousness, and .73 for social anxiety, and .80 for overall selfconsciousness (Fenigstein et al., 1975). Reliability for overall selfconsciousness was .84 in the present study.

Table 2. Reliability Coefficients for Intercultural Behavioral Variables

　SelfDisclosure

　Wheeless's (1978) 31ietm General Disclosiveness Scale was used in the present study to measure subjects' general tendency to disclose to others. The scale consists of five dimensions: amount of disclosure, consciously intended disclosure, honesty/accuracy of disclosure, positiveness/negativeness of disclosure and depth/intimacy of disclosure. The coefficient alphas for the five dimensions in this study were .83 for amount of disclosure, .72 for consciously intended disclosure, .86 for honesty/accuracy of disclosure, .75 for positiveness/negativeness of disclosure, .84 for depth/intimacy of disclosure, and .79 for the total scale.

　These results are similar to those reported by Wheeless (1978) who calculated coefficient alphas of .83 for amount of disclosure, .65 for consciously intended disclosure, .84 for honesty/accuracy of disclosure, .90 for positiveness/negativeness of disclosure, .78 for depth/intimacy of disclosure.

　Communicative Adaptability

　The present study used Wheeless and Duran's (1982) 20item Communicative Adaptability Scale to measure two prominent dimensions of communicative competence: adaptability and rewarding impressions. Principal axis factoring with oblique rotation was used to analyze the scale. The results showed that two factors were extracted from the 20 items. Items 5, 6, and 7 were significantly loaded in the first factor, communication adaptability, which accounted for 33.5% of the variance. Items 11, 12, 13, 14, 15, 16, 17, and 19 were significantly loaded in the second factor, rewarding impressions, which accounted for 13.1% of the variance. The coefficient alphas calculated in this study were .88 for adaptability, .85 for rewarding impressions, and .90 for the overall scale.  Previous research (Wheeless & Duran, 1982) reported coefficient alphas of .86 for adaptability and .87 for rewarding impressions. 

　Social Situations

　The social Situation Questionnaire developed by Furnham and Bochner (1982) was used in this study to measure subjects' Psychological Adaptation in a new environment. A coefficient alpha was not reported. In the present study, a coefficient alpha of .92 was calculated for the overall scale. 

　Interaction Involvement

　The present study used Cegala's (1981) 18item Interaction Involvement Scale to measure communication responsiveness, perceptiveness, and attentiveness. Principal axis factor analysis with oblique rotation was used to identify the main dimensions of interaction involvement. Three factors were extracted accounted for 51.3% of the total variance. The first factor, communication responsiveness, included items 9, 14, 15, 16, and 17, and accounted for 29.6% of the variance. Items 1, 4, 12, and 13 were significantly loaded in the second factor, communication perceptiveness, which accounted for 14.7% of the variance. Items 2, 5, and 8 were significantly loaded in the third factor, communication attentiveness, which accounted for 7.0% of the variance. Item 5 was eliminated in this study in order to achieve a higher reliability coefficient for the third factor. 

　The coefficient alphas of the three components of the Interaction Involvement Scale were .80 for communication responsiveness, .82 for communication perceptiveness, .65 for communication attentiveness, and .83 for the total scale. Previous research by Brunner and Hannigan (1984) reported coefficient alphas of .79 for responsiveness, .67 for perceptiveness, .82 for attentiveness, and .86 for the total scale. Cegala's (1981) study also reported internal reliability levels: .88 for communication responsiveness, .86 for communication perceptiveness, .87 for communication attentiveness, and .90 for the total scale.

　Cultural Awareness

　The Test of American Culture, developed by Kitao (1981), was used in this study to measure the degree of subjects' Cultural Awareness. Kitao's study reported only splithalves reliability. The coefficient alpha of the instrument used in the present study was .74 for the total instrument when answers were coded as either right or wrong.

　Communication Competence

　A version of the Communication Competency SelfReport Scale created by Rubin (1985), adapted for the otherreport mode, was used to measure subjects' general communication competence ability. The present study found the internal reliability of the instrument to be .90, compared to the .87 reported by Rubin (1985) for the selfreport version and the .91 reported by Wilson (personal communication, May 1986) for the otherreport mode.

Intercultural Behavioral Assessment Indices

　This instrument, developed by Ruben (1976), was used in this study to measure seven constructs of intercultural communication competence: display of respect, interaction posture, orientation to knowledge, empathy, relational role behavior, interaction management, and tolerance of ambiguity. The instrument has demonstrated its reliability and usefulness. In the present study, the coefficient alpha of the instrument was calculated at .80. Ruben failed to report alphas in his previous research. 

Test of Hypotheses

　Hypothesis 1 predicted significant positive correlations among measures of Personal Attributes, Communication Skills, Psychological Adaptation, and Cultural Awareness. Pearson productmoment correlations were computed to test this hypothesis. The results are summarized in Table 3.

Table 3. Pearson Correlation Coefficients for Variables

　Results indicated that private selfconsciousness was significantly correlated with public selfconsciousness (r = .53, p < .001), consciously intended disclosure (r = .25, p < .01), depth/intimacy of disclosure (r = .24, p < .01), rewarding impressions (r = .22, p < .05), and with communication perceptiveness (r = .36, p < .001).

　Public selfconsciousness was significant correlated with social anxiety (r = .31, p < .001), consciously intended disclosure (r = .17, p < .05), communication perceptiveness (r = .28, p < .001), and with communication attentiveness (r =  .18, p < .0501). Social anxiety was significantly, but negatively correlated with consciously intended disclosure

　(r =  .53, p < .05), positiveness of disclosure (r = .27, p < .001), honesty/accuracy of disclosure (r =  .33, p < .001), communication adaptability (r = .67, p < .001), social situations (r = . 41, p < .001), communication responsiveness (r = 1 .49, p < .001), and with communication attentiveness (r =  .19, p < .05).

　For the selfdisclosure dimensions, the results indicated that consciously intended disclosure was significantly correlated with honesty/accuracy of disclosure (r = .54, p < .001), communication adaptability  (r = .42, p < .001), rewarding impressions  (r = .41, p < .001), social situations (r = .32, p < .001), communication responsiveness (r = .35, p < .001), and communication attentiveness (r = .36, p < .001). Amount of disclosure was significantly correlated with depth/intimacy of disclosure (r = .44, p < .001). Positiveness/negativeness of disclosure was significantly correlated with communication adaptability (r = .30, p < .001), communication responsiveness (r = .18, p < .05), and communication attentiveness (r = .19, p < .05). Lastly, honesty/accuracy of disclosure was found to be significantly correlated with communication adaptability (r = .45, p < .001), rewarding impressions (r = .47, p < .001), social situations (r = .34, p < .001), communication responsiveness (r = .55, p < .001), communication perceptiveness (r = .25, p < .01), communication attentiveness (r = .37, p < .001), and with communication competence (r = .24, p < .01). 

　Communication adaptability was found to be significantly correlated with rewarding impressions (r = .40, p < .001), social situations (r = .53, p < .001), communication responsiveness (r = .59, p < .001), communication perceptiveness (r = .17, p < .05), communication attentiveness (r = .25, p < .01), cultural awareness (r = .17, p < .05), and with communication competence (r = .19, p < .05). Rewarding impressions was significantly correlated with social situations (r = .32, p < .001), communication responsiveness (r = .46, p < .001), communication perceptiveness (r = .26, p < .01), and communication attentiveness (r = .30, p < .001), and with communication competence (r = .28, p < .01).

　Psychological adaptation to various social situations was significantly correlated communication responsiveness (r = .46, p < .001), communication perceptiveness (r = .20, p < .01), and communication attentiveness (r = .29, p < .001).

　The interaction involvement measures were also found to be significantly correlated to other dimensions. For example, communication responsiveness was significantly correlated with communication perceptiveness (r = .27, p < .001), communication attentiveness (r = .46, p < .001), cultural awareness (r = .27, p < .001), and with communication competence (r = .27, p < .01). In addition, communication perceptiveness was significantly correlated with communication attentiveness (r = .43, p < .001).

　In sum, the results showed significant relationships among measures of selfconsciousness, selfdisclosure, communication adaptability, social situations, and interaction management. However, the results also showed that the dimension of Cultural Awareness was only significantly correlated with measures of communication adaptability and communication responsiveness. This indicates that significant relationship exists only between Cultural Awareness and Communication Skills.

　To test Hypothesis 2, canonical correlation analysis was used to test the relationship between the linear combination of one set of variables (i.e., display of respect, interaction posture, orientation to knowledge, empathy, relational role behavior, interaction management, and tolerance of ambiguity) and a second set of variables (i.e., private selfconsciousness, public selfconsciousness, social anxiety, consciously intended disclosure, amount of disclosure, positiveness of disclosure, depth/intimacy of disclosure, honesty/accuracy of disclosure, communication adaptability, rewarding impressions, communication responsiveness, communication perceptiveness, communication attentiveness, cultural awareness, social situations, and communication competence). The structure coefficients, which indicate the correlation of the variables in one set with the function in the other set, were used to define the variable sets. The .05 level of confidence was used for all statistical tests. 

　Table 4 shows that Hypothesis 2 was confirmed. One significant canonical root was found. The canonical correlation (Rc) between the first and second set of variables was .70. An examination of the canonical loadings revealed that the first set was defined by communication competence. Results of the canonical correlation analysis also indicated a significant high positive relationship between display of respect, interaction posture, orientation to knowledge, empathy, relational role behavior, interaction management, tolerance of ambiguity, and communication competence. In addition, results showed a significant but small positive relationship between display of respect, interaction posture, orientation to knowledge, empathy, relational role behavior, interaction management, tolerance of ambiguity and both cultural awareness and social anxiety. 

　

Table 4. Canonical Analysis Summary for Intercultural Behavioral Assessment Indices and Intercultural Behaviors

Exploration of Research Questions

　To find out which of the sixteen measures of Personal Attributes, Communication Skills, Psychological Adaptability, and Cultural Awareness best predict the seven dimensions of intercultural communication competence, stepwise multiple regression analyses were conducted. Each of the seven dimensions was regressed onto the 16 measures of intercultural communication competence. The results are presented in Table 5.

　　　　

Table 5. Summary of Stepwise Multiple Regression Analysis


　Results indicated that display of respect was best predicted by communication competence, communication perceptiveness, cultural awareness, and social anxiety. Interaction posture was best predicted by communication competence and honesty/accuracy of disclosure. Orientation to knowledge was best predicted by communication competence and amount of disclosure. Furthermore, empathy, relational role behavior, interaction management, and tolerance of ambiguity were each best predicted by communication competence.

　The purpose of the second research question was to find a model that would best explain intercultural communication competence. In order to answer this research question, factor analysis was first used to test (a) the seven dimensions of intercultural communication competence (i.e., the seven IBAI measures), and (b) the 16 measures of intercultural communication competence. Second, multiple regression analysis was used to investigate which factor (of those created from the 16 measures of intercultural communication competence) best predicts intercultural communication competence. Finally, factor analysis was performed to generate a new model that might best explain intercultural communication competence by combining all intercultural communication competence measures.

　Table 6 reports the results of the first principal axis factor analysis. Only one factor was extracted from the seven measures of intercultural communication competence; it accounted for 46.4% of the variance. This, along with the canonical correlation results, and a coefficient alpha of .80 for the seven measures, indicates that the IBAI items should be combined into one measure of intercultural communication competence. Zscores were computed and summed for further analyses.

　　　　

Table 6. Principal Axis Factor Analysis of the Seven Components of Intercultural Communication Competence

　Table 7 reports the results of the second principal axis factor analysis. Three factors with eigenvalues of 1.00 or higher were extracted for the 16 measures of intercultural communication; these accounted for 42.2% of the variance. The first factor accounted for 22.6% of the common variance. Those variables having loadings of at least .50 with secondary loadings no higher than .35 were: social anxiety, social situations, communication adaptability, and communication responsiveness. These variables are concerned with the ability to adapt appropriately and effectively to the social interaction. On the bases of these data, Factor 1 was labeled Social Adjustment. The coefficient alpha for this factor was. .39.

　

Table 7. Oblique Rotated Factor Matrix for the Measure of Intercultural Communication

　The second factor accounted for 10.2% of the common variance. Those variables having high loadings were communication perceptiveness, private selfconsciousness, and public selfconsciousness. These variables are mainly concerned with a person's ability to monitor or understand his or her own behavior. Therefore, this factor was labeled SelfAwareness. The coefficient alpha for this factor was .61.

　The third factor accounted for 6.3% of the common variance. Two variables had a significant loading on this factor: amount of disclosure and depth of disclosure. Since both variables are concerned with selfdisclosure, this factor was then labeled SelfDisclosure. The coefficient alpha for this factor was. .61.

　Factor scores were computed for these three factors and the new IBAI total was regressed onto the new three dimensions of intercultural communication. Results indicated that the first factor, Social Adjustment, best predicted intercultural communication competence (r = .19, p = .13). The amount of variance accounted for was very small. 

　To test the original model of intercultural communication competence composed of the four dimensions of Personal Attributes, Communication Skills, Psychological Adaptation, and Cultural Awareness, the 16 measures comprising the original four dimensions and the IBAI total (using zscore procedures for summation) were submitted to principal axis factor analysis. Table 8 reports the results, and Table 9 reports the factor correlation coefficients. Five factors were extracted accounting for 47.4% of the variance. After oblique rotation, the first factor accounted for 21.4% of the common variance and had an eigenvalue of 3.63. Those variables loading on the first factor were: social anxiety, communication adaptability, social situations, and communication responsiveness. This factor, identified with the first factor on Table 6, was labeled Social Adjustment. The coefficient alpha for this factor was .59.

Table 8. Oblique Rotated Factor Matrix for the 16 Measures of Intercultural Communication and Intercultural Communication Competence

　The second factor accounted for 10.1% of the common variance and had an eigenvalue of 1.72. Only one variable, i.e., private selfconsciousness, loaded cleanly on this factor, yet public selfconsciousness was also clearly involved. This factor was labeled SelfConsciousness. The coefficient alpha for this factor was .65.

　The third factor had an eigenvalue of 1.20 and accounted for 7.0% of the common variance. Two variables loaded significantly on this factor, communication competence and intercultural communication competence. This factor was labeled Communication Competence. The coefficient alpha for this factor was .36.

　The fourth factor has an eigenvalue of .98 and accounted for 5.8% of the common variance. Amount of selfdisclosure and depth of selfdisclosure were the two variables significantly loading in this factor. Therefore, this factor was labeled SelfDisclosure. The coefficient alpha for this factor was .61.

　Finally, the fifth factor had an eigenvalue of .52 and accounted for 3.0% of the common variance. Communication attentiveness and perceptiveness were the only variables loading clearly on this factor. Interaction Involvement was then used to label this factor. The coefficient alpha for this factor was .35. 

　Therefore, the results of the second research question indicate that a new model can be generated to explain intercultural communication competence. The new factors found in this study were Social Adjustment, SelfAwareness, SelfConsciousness, Communication Competence, and Interaction Involvement. As shown in Table 9, these new factors were not strongly interconnected. This new model might be used to replace the model originally proposed in the first chapter and the model originated from the 16 measures of intercultural communication competence.

Additional Analyses

　Subjects' demographic data were also analyzed in this study. As evident on Table 10, length of time subjects had lived in the United States was significantly correlated only with rewarding impressions (r = .23, p < .01), communication responsiveness (r = .17, p < .05), and with communication perceptiveness (r = .17, p < .05). Level of acquaintance was significantly correlated only with communication adaptability (r = .20, p < .05), rewarding impressions (r = .28, p < .001), communication responsiveness (r = .28, p < .001), communication competence (r = .44, p < .001), and with intercultural communication competence (r = .35, p < .001).
　
Table 10. Pearson Correlation Coefficients for Demographic Variables
、
　Ttests were used to analyze differences between men and women, and native and nonnative speakers of English on the 16 measures of intercultural communication and intercultural communication competence. The results showed (a) males (M = 4.78) were significantly more positive about their selfdisclosure than females (M = 4.45), t(141) = 2.08, p < .05); (b) subjects with English as native language (M = 3.65) reported significantly greater depth in selfdisclosure than subjects without English as a native language (M = 2.47), t(143) = 2.23, p < .05); (c) subjects with English as a native language (M = 5.98) were significantly higher in communication adaptability than those without English as a native language (M = 5.00), t(140) = 2.26, p < .05; (d) subjects with English as a native language (M = 5.93) reported that they were significantly more perceptive in their interaction than subjects without English as a native language (M = 4.90), t(139) = 2.31, p < .05; and (e) subject with English as a native language (M = 23.17) were significantly more culturally aware than subjects without English as a native language (M = 20.16), t(146) = 2.11, p < .05.

　Lastly, oneway analysis of variance (utilizing Turkey's post hoc test) was used to investigate differences of nationality and marital status on the 16 measures of intercultural communication and intercultural communication competence. The results showed significant differences between subjects (a) from the Far East (M = 2.81) and Middle East (M = 1.96) on social anxiety, F(5,139) = 3.60, p < .01; (b) from Asia (M = 4.93), and Europe (M = 4.06), Africa (M = 5.37) and Europe (M = 4.06), and Africa (M = 5.37) and the Far East (M = 4.55) on positiveness of selfdisclosure, F(5,136) = 4.31, p < .01; and (c) from Africa (M = 5.37) and the Middle Ease (M = 3.94) on communication perceptiveness F(5,134) = 3.36, p < .01. No significant differences were found for persons of different marital status.