Exploration of Multiple Intelligence by Using Latent Class Model

Exploration of Multiple Intelligence by Using Latent Class Model

Shaista Ismat and Junaid Sagir Sidiqui

http://dx.doi.org/10.6000/1927-5129.2012.08.02.28

Abstract: In this article we have tried to explore, “multiple intelligence” in the educated youth through questionnaire items by applying latent class models. A questionnaire consists of 50 questions. These questions have constructed in the light of Howard Gardner theory of multiple intelligence to explore “multiple intelligence”. A survey was conducted on 399 adult students from different regions of Karachi. For statistical analysis we have selected three sets with seven variables, and one set with 4 variables each with binary response.  On these four sets up to three classes latent class models were applied. The Probability of positive response (πiy) in each class were estimated by using E.M algorithm and interpreted the class as on the basis of πiy values. By assessed goodness of fit latent classes/ groups were identified. Two class (two groups of people) model was found in all four data sets. A group (class) consists of the people who think that they have strong verbal expressions abilities, effectively use language to express himself/herself theoretically and poetically, they have good ability to recognize musical pitches, tones and rhythms, we may call this class as “self competence and self esteem” as “musically talented” as “socialize” (having high interpersonal ability).

Keywords: Manifest variables, latent variable, positive response, stochastically independent, prior probability, label, likelihood, Estimation, Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC).