Abstract:

Statistical training is important in a wide range of fields, but many students dislike and even fear it. Negative attitudes towards statistics can reduce academic performance and disrupt the teaching-learning process, making it important to better understand these attitudes so that effective interventions can be designed. Here we provide an exploratory and descriptive study of the relationships between different dimensions of attitudes towards statistics, a preliminary step in the process of designing better teaching strategies . We analysed attitudes towards statistics in a sample of 76 Canadian (age: mean = 21.74, SD = 1.57; 76.3% female) and 81 Spanish (age: mean = 19.99, SD = 3.1, 76.5% female) university students. Participants filled in an electronic questionnaire containing the English or Spanish versions of the Auzmendi Scale to Measure Attitudes towards Statistics (ASMAS). The ASMAS measures five dimensions of attitudes about statistics: perceived utility-value, anxiety, security-confidence, pleasantness, and motivation. Results of network analysis (EBICglaso estimator, normalized centrality measures and tuning parameter = 0.5) suggest that anxiety items are the most strongly interrelated. The motivation and confidence items also cluster together but their relationships are slightly weaker. The items from the utility and pleasantness dimensions are partially mixed in our sample data. Analyses using an acyclic directed graph (ADG) on the complete scales provided initial insights into how these dimensions relate to one another. Results indicated that the pleasantness dimension predicted the motivation, security-confidence, and value-utility dimensions. In turn, security-confidence predicted value-utility, and both combined with motivation predicted anxiety. We also found differences when comparing the networks generated by country. Our results are useful to understand the complex relationships between attitude dimensions, and to plan and design teaching strategies to improve the teaching-learning process.

Keywords:

Attitudes, statistics, network analysis, university, Bayes nets.