An Empirical Evaluation of a Test Measuring Children's Attitudes towards TV Advertisements

Abstract

In a recent article, Rossiter ( 2 ) developed a 7-item scale to measure children's attitudes towards television commercials. The items assessed: ( a ) perceived truthfulness, ( b ) potential annoying qualities, ( c ) objectivity in describing advertised products, ( d ) over-all Likability, ( e ) perceived persuasive power, ( f ) believability of characters, and ( g ) trustworthiness as guides to product purchase. The scale was designed to be either selfor interviewer-administered, the child responding to a $-point Likert scale, strength of agreement or disagreement being verbally and visually reinforced through use of large and small response words and boxes. When given to 208 4th to 6th grade children the scale showed high internal consistency reliability ( a = .69) and moderate test-retesc reliability ( r = .67). The present study had two objectives: to examine the relationship between age and socio-economic status, on the one hand, and the attitudes of children towards Rossiter's items on the other; and ro test Rossiter's assumption that the seven test items would empirically load onto one factor. A sample of 165 children was drawn from two elementary schools in the metropolitan area of Perth, capital of Western Australia. Rossiter's attitude test was interviewer-administered in each school to two age groups: 6to 8and 9to 12-year-olds, n = 88 (M.,. = 7.06 yr., SD = .68) and n = 77 (M.,. = 11.43 yr., SD = .65), respectively. The socio-economic scams of each child was determined using Duncan's Socio-economic Status Scale ( 1 ) . Age correlated significantly ( a < ,005) with each of the seven items: the older the child the more negatively was each perceived. Socio-economic status was significantly correlated with items d ( a = .001), f ( a = .04), and g ( a = ,003). Of the 21 inreritem correlations, all but two were significantly intercorrelated ( a < .04). To test Rossiter's one-factor assumption a varimax rotation of the principal factor matrix with Kaiser normalization was performed. Two factors emerged, Items b, d, and g loaded on Factor 1 which appears to measure the attraction of commercials to the audience, while Items a, c, e, and f loaded on Factor 2, interpretable as viewing the commercial content of advertisements as input to the decision to purchase.