Parental expectations are a strong predictor of a gifted student’s academic achievement and higher education attainment. The primary flaw in many parental expectations studies is the over reliance on a single-item to represent this construct. Many of the US national (NELS88, NAEP) and international data bases (TIMSS) use single-item constructs as the only measure of parental expectations.
Research question
The purpose of this study is to synthesize a parental expectations construct and to compare it to a single-item construct in separate path analyses. Which construct is the best predictor in achievement analyses?
Research design & Methodology
This study utilized 6,502 participants from 17 separate studies of students enrolled in elementary, middle, and high schools in the United States (3rd grade through 12th grade). Chin (1988) recommends doing additional factor analyses (either PCA or PFA) using the items isolated in a preliminary factor analysis to isolate different factors at a higher level of abstraction. We used this approach with PCA analyses to synthesize a 9-item expectation factor. We then used the 9-item factor and the single-item construct as independent variables in separate path analyses.
Main result
We analyzed the predictor capabilities of the variables’ with three achievement variables (math, language arts, GPA). The most significant finding of the study is that utilizing a single-item proxy variable to represent parental expectations results in multiple Type I and Type II errors. We agree with Sarstedt and Wilczyski (2009) that multi-item constructs are more productive.
Conclusion
When conducting future research on pre-disposition, parental expectations, and social capital theory, a factor must be utilized to represent parental expectations because complex scio-psychological constructs cannot be adequately represented by single-item measures.