USE OF FEEDBACK TO IMPROVE SYMBOLIC-NUMBER MAPPINGS

Purpose: As patients increasingly take an active role in their medical care, they are asked to weigh the risks and benefits of different courses of treatment. Patients with low numeracy are at a disadvantage when making these decisions, as low numeracy may be associated with less exact symbolic-number mappings. The aim of this study was to evaluate whether providing feedback to adult subjects could improve performance on a number line placement task.

Methods: Subjects from two outpatient clinic waiting rooms participated in a three phase number line task. Participants were asked to place eight numbers on a computerized number line ranging from 0 to 1000 (pre-test phase). Participants then placed five additional numbers, but with the correct position being shown after their choice was made (feedback phase). Finally, participants placed the same eight numbers as in the pre-test without feedback (post-test phase). Participant demographics, subjective numeracy, and education level were collected.

   Pre-test and post-test scores, reported as percent absolute error (PAE), were calculated as the average of the absolute differences between the number prompt and the location selected on the number line, divided by the range of the number line. Generalized estimating equations (GEE) were used to model log-transformed scores and to test whether performance improved after feedback and whether the degree of improvement was associated with age, gender, race, education level or subjective numeracy.

Results: There was an overall improvement in task performance following the feedback. The average PAE was 7.32% (SD: 6.00) for the pre-test and 5.63% (SD: 3.71) for the post-test. In the multivariate log-transformed GEE model (Table), college education, non-Hispanic white race, and higher subjective numeracy were associated with smaller absolute errors on pre- and post-test tasks. We found no interaction between subjective numeracy and improvement. There was, however, a significant interaction between college education and post-feedback improvement, with college education moderating the benefits of the feedback. Participants with college education had high pre-test scores and did not significantly improve after feedback, while those without a college education improved significantly.

Conclusions: Adults improve significantly on a number line task when given feedback. Participants without a college education demonstrated the largest improvements. Future studies should address whether feedback, by improving symbolic-number mapping, could improve health-related decision-making in patients with low education.