Network structure of responses to a reading-attitude survey

Abstract

In this paper, we present a new approach in analyzing social survey data using networks. From a social survey data related to the attitude of pupils in their early elementary years about reading attitude, we constructed complex networks where pupils are the nodes (N = 955) and they are connected to each other based on their similarity T of response to the survey questions having three- (l = 3) and  five-point (l = 5) scales. The value of l is the number of possible levels of answers from most sad to happiest. Analysis of the networks shows that clustering of pupils with similar responses does not show significant dependence on the pupils' age, sex, reading readiness and survey tool (face iconicity) used. The networks were found to form single clusters for low T values and start to break apart at T = 7 and T = 6 for the three- and five- point scale networks respectively. We show that there are no significant differences in the average degree <k> and average clustering coefficient <C> between the three-point and five-point response scales used. We also show that the resulting network clustering is not simply random.