This thesis models the effects of errors of omission and commission due to increased working memory demands on accuracy rates of spatial relational reasoning problems. The questions modelled are taken from the Multidimensional Relational Reasoning Task (Cortes et al., 2021). Four models were created using Python ACT-R and PRISM theory which demonstrated the negative relationship of increased premises and dimensions per problem on accuracy rates. Results highlight the need for future modelling to consider the individual differences in micro-strategy preferences, how reasoning processes may be affected by different memory errors, and how future measures may be constructed to better address raised concerns.