What is it about?
It is not easy to understand strengths and weaknesses of specific domains from individuals' observed responses on subtests. Thus, we introduce a new way to summarize individuals' strengths and weakness in certain domains by identifying a small set of summative profiles.
Featured Image
Why is it important?
The conventional statistical methods, such as t-test, ANOVA, Regression, factor analysis, ... , etc., are not designed to measure strengths and weaknesses of domains (e.g., math and reading, certain cognitive skills or some clinical symptoms) responded by individuals. However, the profile analysis introduced in our study is designed to interpret person response profiles with a small set of summative profiles which are identified by principal component analysis. Once the summative profiles are labeled or defined with their peaks (strengths) and valleys (weaknesses) in their patterns, we regress person profiles onto the summative profiles to estimate person weights. The person weights function as matching statistics (like correlation) between person profiles and core profiles. Thus, those who have substantial weights on a certain summative profile are assumed to carry the same characteristics of the summative profile (note that this summative profile has already been labeled/interpreted based on its pattern). Summative profiles may be considered latent because they are basically estimated from several dimensions (or components); thus, they carry richer information than a single mean profile.
Perspectives
We used educational data as our example but Profile Analysis via Principal Component Analysis (PAPCA) introduced in the study can be applied to any rectangular data sets where rows represent cases/persons and columns measuring variables, such as public health or psychological data, if researchers are interested in identifying summative profiles from numerous person response profiles and studying strengths (improvement in clinical trials) or weaknesses (deterioration in clinical trials) of individual students' or patients' subscores. In a rectangular data set, rows are in fact person response profiles that consist of responses on column measuring variables/sub-scales, and PAPCA summarizes various individuals' subscore patterns with two or three summative profiles and helps understanding the overall strengths and weaknesses patterns with a few summative profiles.
Dr. Se-Kang Kim
Fordham University
Read the Original
This page is a summary of: A Tool Extracting Summative Profiles From Person Score Profiles, Methodology, April 2017, Hogrefe Publishing Group,
DOI: 10.1027/1614-2241/a000125.
You can read the full text:
Resources
Contributors
The following have contributed to this page
