Publication Type: Conference Paper/Unpublished Manuscript
Review Method: Peer Reviewed
Abstract: Since 1979, the National Science Foundation has sponsored a biennial survey to assess public attitudes toward and understanding of science and technology. In an effort to reduce the length of the survey with hopes of improving response rates, this study developed a short form for the Factual Knowledge of Science (FKS) scale, one of the primary measures included in the survey. The FKS scale includes 18 questions designed to measure respondents’ knowledge of scientific concepts. Because the FKS scale is used for longitudinal comparisons, a primary concern was that the short form should perform in the same manner as the long form and retain as much information as possible. Data from the 1997, 1999, and 2001 surveys were used to develop and validate the short form. First, the knowledge questions were classified into eight content areas: life science, evolution, geology/geography, astronomy, radioactivity, physics, and chemistry. Exploratory and confirmatory factor analyses indicated that the items form a single factor. Item response theory analyses were conducted to evaluate the difficulty and discrimination of each item. Items for the short form were selected to represent all eight content areas and maintain the distribution of difficulty levels included on the long form. Discrimination parameters were also used in the selection of items. The resulting short form contained 11 items, had a correlation of .95 with the long form, and accounted for 90 percent of its variance. Cronbach’s alpha was nearly identical for the two forms. Group comparisons revealed that the short form detected the same patterns of differences between respondents based on gender, age, education level, and background in math and science as the long form. The results of this study support the reliability and validity of the short form, suggesting its suitability as an alternative to the current scale.