Poster #43 - What predicts middle school students’ growth in symbolic number comparison performance?

• In Event: 3-003 - Poster Session 09
  In Poster Session: PS 09 Section - Cognitive Processes

Integrative Statement

Symbolic number comparison (SNC), such as determining whether 2 or 8 is larger, is commonly used to measure numerical magnitude processing and its relation to mathematics achievement (e.g., Schneider et al., 2017). Performance on SNC tasks differs across age groups, young students improve on SNC (e.g., Matejko & Ansari, 2016), and performance is better in adults than children (e.g., Holloway & Ansari, 2008). However, open questions remain about whether students improve their SNC performance beyond elementary school, and the extent to which individual differences in related skills may contribute to any change. The present study addressed these questions in order to illuminate the developmental trajectory of SNC between the early grades and adulthood, and to clarify the cognitive processes that relate to SNC task performance.

Participants were from a longitudinal sample of middle school students (n = 376). In Grade 5, students (Mean age = 11.02 years, SD = 0.32) completed single-digit SNC, nonsymbolic number comparison, and executive function (EF) tasks, and a mathematics competence assessment. Students completed double-digit SNC in Grades 6-8. We used a multi-level model for change to first examine whether students’ SNC performance (i.e., response time adjusted for accuracy) changed over time generally. We then investigated the effects of single-digit SNC and nonsymbolic performance, EF, and mathematics competence in Grade 5 on Grade 6 SNC performance and rate of change in double-digit SNC performance from Grades 6-8.

Results showed that students’ SNC performance improved from Grades 6-8, on average (PE = -84.61 ms, SE = 5.81, p < .001). Table 1 displays the final model of predictors of SNC from Grades 6-8. Results show that Grade 5 SNC performance and Grade 5 EF predicted Grade 6 SNC performance, controlling for nonsymbolic comparison performance (Table 1, first panel). However, only Grade 5 nonsymbolic comparison performance predicted SNC rate of change from Grades 6-8, controlling for SNC performance and EF (Table 1, second panel). Interestingly, the relation was inverse, in which students with relatively worse performance on nonsymbolic comparison improved more in SNC from Grades 6-8, compared to students who performed relatively better, controlling for Grade 5 SNC and EF. Lastly, controlling for number comparison performance and EF, Grade 5 mathematics competence did not predict Grade 6 SNC performance (PE = -6.79, SE = 4.44, p = .13) or rate of change (PE = -0.12, SE = 2.43, p = .96).

Taken together, results suggest that students refine their basic number processing skills well beyond when mathematics education explicitly targets these skills. Results show that SNC and EF skills independently predict SNC performance, and that predictors of SNC performance and SNC rate of change differ. However, the mechanisms that underlie SNC rate of change remain opaque. When measuring SNC rate of change, nonsymbolic comparison performance may capture numerical magnitude processing and EF skills better than related single-digit SNC and EF tasks. Further, single-digit SNC may not be sensitive enough to detect change in SNC beyond elementary school, suggesting that the use of double-digit SNC tasks may be preferable.

Authors

• Courtney Pollack, Massachusetts Institute of Technology
  Presenting Author

• Eric Wilkey, Western University
  Non-Presenting Author

• Gavin R. Price, Vanderbilt University
  Non-Presenting Author