Developing working memory and numeracy skills in primary-aged school pupils by computerised training programmes : the effectiveness of adaptive multi-domain working memory training

  • Sharon Louise Brown

Student thesis: Doctoral Thesis


Systematic reviews indicate that while working memory training programmes may yield ‘near-transfer’ effects with improved scores on measures of working memory, there is less convincing evidence for ‘far-transfer’ effects generalising to academic skills. This thesis aims to further investigate training-transfer effects by evaluation of two computerised training programmes to develop working memory and numeracy in primary school-aged pupils.Three studies are reported. The first is a quasi-experimental design comparing effects upon 7-8 year old pupils of a 7-week computerised working memory training programme Memory Quest Flex (Junior) (Truedsson & Strohmayer, 2010) (n=25) with a 7-week computerised numeracy training programme Study Ladder (Study Ladder Holdings Ltd., 2013) (n=27), both relative to a non-active control condition receiving their regular curriculum (n=24).The computerised numeracy training programme served as both treatment and active control conditions, to control for the Hawthorne Effect. Training-transfer effects are measured using performance scores on dependent variables of working memory, fluid reasoning and numeracy, and independent variables of treatment condition, time-point and gender. Children’s scores on the Pupil’s Perceptions of Numeracy Questionnaire (PPNQ), a short scale developed for this study, were considered to illuminate findings of the main study.Findings from the main study change score analysis revealed the adaptive computerised working memory training programme yielded significant near-transfer effects to a complex working memory task, [F(2,73) = 13.9, p <.001, ŋp² = .28], with Cohen’s d +1.34 relative to other conditions. Significant transfer effects werer eported for fluid reasoning, [F(2, 73) = 31.6, p <.001, ŋp² = .46], with Cohen’s d +1.34 relative to other conditions, and numeracy performance, [F(2, 73) = 4.63, p =.013, ŋp² = .11], with Cohen’s d + .65, relative to the control condition.However, Study Ladder yielded only near-transfer effects to numeracy performance, with Cohen’s d + .56, that approached conventional significance levels (p = .053). No significant main effect was found for gender across conditions for any dependent variables (all p-values > .251).In the second study, Principal Components Analysis confirmed the PPNQ had a unidimensional structure and good internal consistency reliability (Cronbach’s a = .806). Findings revealed a positive shift in PPNQ scores following intervention in understanding numeracy instruction. Findings for the control condition support conclusions that the PPNQ was sufficiently sensitive to detect increasingly negative perceptions of numeracy. This study suggests how the PPNQ might be used to identify children’s perceptions of numeracy and access to the numeracy curriculum.The third study used qualitative methods to investigate issues of implementation fidelity important for acceptability and feasibility of computerised training programmes in the primary schools.This thesis identifies dual benefits of adaptive, multi-domain computerised working memory training not fully explained by the Hawthorne Effect. Recommendations are discussed towards the development of a computerised working memory training paradigm in schools.
Date of Award13 Nov 2017
Original languageEnglish
Awarding Institution
  • University Of Strathclyde
SupervisorJim Boyle (Supervisor) & Lisa Woolfson (Supervisor)

Cite this