Assessing Children's Friendship Expectations: Supplementing the Semistructured Interview with Picture Sequence Tasks

Abstract

Seventy-five subjects from Grades 2 and 4 were given two tasks to determine their developmental friendship expectation (FE) levels. The first task employed a semistructured FE interview and the second was a picture sequence task similar in format to the Picture Arrangement (PA) subtest of the WISC-R. To control for sequencing ability on the pictorial task, a modification of PA was also administered. Nine FE picture sequences, corresponding to an extant eight-level developmental sequence, were administered to the children. Interviews were coded along the same dimensions. FE levels based on the interview task were marginally more advanced for both grades: 11 subjects were markedly competent and 8 markedly better on the interview; 11 subjects were markedly competent and 10 markedly better on the visual task. Interview FEs showed no significant relationship to age, whereas picture sequence FEs did. This difference was due to the Type II errors on the visual task at Grade 2 and the interview task at Grade 4. Picture sequencing was unrelated to PA or to task difficulty. The visual task was an apparent underestimate of FE level for 65 subjects who scored s FE level 2 on the visual task. However, sole use of the interview created a large underestimate for 10 of 11 subjects who scored > FE level 2 on the visual index. The risk of underestimation was eliminated by additional use of the visual index, and the more advanced the performance on the visual task, the higher the risk of false negative errors on the interview.