CONFIABILITY OF REPEATED-MEASURES OF THE PATERNITY INVENTORY FOR ADULT AND ADOLESCENTS (PIAA)

This study was carried out as part of a research line on parenting att itudes. Previous studies in Mexico have tested the Adult-Adolescent Pa renting Inventory (AAPI) in terms of its validity, reliability, standa rization, and discriminant validity in order to detect either child ab users (< 20 years of age) or teens (12-19 years of age) who have been abused. The AAPI measures four parenting constructs: inappropriate dev elopmental expectations, lack of empathy, use of corporal punishment, and parent-child role reversal. The aim of the present study is to obt ain a test-retest reliability of the AAPI-Spanish version; which has s hown an adequate validity and reliability to assess parenting attitude s, particularly for adults. Lack of variance of adolescents responses contributed to obtain smaller alpha coefficients (Solis-Camara & Diaz, 1991). So far, the studies conducted in Mexico (Diaz et al., 1990; Di az et al., 1991) have administered the AAPI-Form B, but there is an al ternate Form (named A). In view of the existence of two Forms, authors are considering the possibility of conducting a test-retest study wit h alternate Forms. However, those subjects of the studies who have ans wered Form B, have commented on the high reading level of the items of Form B. Apparently, the wording of some items in negative Form and in past tense (e.g., item 6 of Expectations: children sould not be expec ted to talk before the age of one year) suggested some people to mark the ''disagree'' answer. Written comments by these people indicated th at they were in disagreement with the expectation of children verbally expressing themselves before the age of one year, but with their answ er the meaning conveyed to the statement was that they do believe chil dren should be expected to talk before the age of one year, but whit t heir answer the meaning conveyed to the statement was that they do bel ieve children should be expect to talk before the age of one year. As Form A is worded in present tense non-negative statements, it seems ea sier to answer (Bavolek, 1989). In this study, authors changed the wor ding of several Form-B items (19 items), thus, the negative form was k ept but the forms of verbs was changed to present tense; also some wor ds were changed to more common ones. The authors expected to obtain hi gher scores by improving Form B. Therefore, authors studied the AAPI t est-retest reliability on each separate Form (A and B), for the purpos e of obtaining more information from the differential scores by Forms, and the temporal stability of each one. A one-week period was establi shed as adequate for measuring the stability of the inventory. A rando m sample of 300 subjects with restrictions of sex (50% of each) was se lected from one elementary school (parents only) and two high-schools (both parents and adolescents); the sample came from middle-SES in acc ordance with Mexican standards. The AAPI is a Likert-type instrument w ith 5 alternative responses (from strongly disagree to strongly agree) . Raw scores are computed by adding item scores of each scale. Two-hun dred and fifty three subjects (1 39 parents and 114 teens) answered th e first administration of the AAPI; 167 subjects (66%) answered twice the AAPI (84 parents and 83 adolescents; there were 65 mothers, 19 fat hers, 40 female teens, and 43 male teens). Each AAPI scale was analyze d by sex by a group controlling type of Form. The main effects were fo und on the first testing of punishment (F[1,56] = 9.8, p < .01) and em pathy (F[1,56] = 6.9, p < .05) for adolescents who answered Form A; in both cases, girls scored significantly higher (p < .05) than boys. As only a few differences were found, and they were not important for th e main purpose of the study, further analyses were not controlled by s ex. Comparisons of AAPI scores by type of Form (i.e., Form B first ver sion, B modified, and A) indicated that scores tended to increase from the first B version (i.e., the one used is the validity study) to the Form-B, modified version used in this study, and to Form A; however, t tests indicated that most increases were not significant, with the e xception of role reversal parents' scores between both Forms B (t[406] = 2.6, p < .01), and role reversal and expectations teens' scores bet ween Forms B (t[592] = 27, p < .01). Regarding Form A, parents' scores were significantly higher on each scale as compared with Forms B; for adolescents, however, empathy and role reversal scores were similar b etween Forms. Comparisons of AAPI scores, for Form A, between parents and teens, indicated parents scored higher than teens on expectations (F[1,124] = 6.3, p< .05), empathy (F[1,124) = 28.4, p < .01), punishme nt (F[1,124] = 10.2, p < -01), and role reversal (F[1,124] = 86.7, p < .001); similar results were found for Form B. Solely parents' expecta tions scores were not higher than adolescents'. The main purpose of th e present study was to test the temporal stability of subject response s to both AAPI Forms. First and second administrations of the inventor y scales were compared with repeated measures t test. We did not find statiscally significant differences between administrations of Form A, in parents (n = 41), but we did find a significant increase on expect ations scores of Form B (t[42] = 2.0, p < .05). In contrast, we found higher scores in adolescents on Form A second administration, in empat hy and role reversal, and in expectations and role reversal scores of Form B. Pearson correlations were separately performed for test-retest scores of parents and teens by Forms. For parents, all correlations w ere significant, either in Form A (range from .54 to .80, p < .001) or Form B (.35 -.73, p < .05), and also for teens (Form A: .48 -.82, p < .001, Form B: .39 -.68, p < .01). The results, of this study support p revious findings (Solis-Camara y Diaz, 1991) where parents have scored higher than teens on AAPI scales, whit the exception of expectations in Form B, and where there were no sex differences In this study, auth ors explored the assumption that Form A of the AAPI is easier to under stand than Form B. Results indicated that Form A, indeed, is easier th an Form B, even with the im proved wording of several items, particula rly for parents. Our results support the test-retest reliability of th e AAPI scales in their two Forms; however, further studies are needed to examine test-retest reliability with alternate forms and the differ ential pattern of responses between parents and teens.