Research into working memory has developed two models; the multi-component model of working memory and the embedded processes model, which differ on their theories of capacity. Reporting k values and adjustment of load have been introduced to try and gain a clearer picture of working memory capacity, however the models are based on old studies which do not utilise these techniques, such as Cowan and Morey, 2007. Therefore this study will replicate the Cowan and Morey (2007) study, but introduce load adjustment to see if the results are replicated and help resolve some of the debate over capacity. It was hypothesised that the introduction of load would alter the results to show no differences in recall and support the embedded processes model. This was tested by analysing accuracy between 5 conditions which differed in the number of stimuli and, if two stimuli were present, whether they were from similar or different domains. 24 student participants completed an experimental test, based on Cowan and Morey (2007), which had 127 trials including verbal and visual stimuli. This test was adjusted for the participant’s own verbal and visual memory spans. The data showed a significantly better accuracy when recalling one stimulus over two stimuli, and a significantly better accuracy when recalling stimuli from similar domains than from different domains. The ANOVA showed that these differences were highly significant, F(2.64, 63.12) = 96.04, p < .05 The results also suggest the accuracy is affected by the stimuli to be recalled and not the stimuli to be encoded. Neither the hypothesis nor the arguments of Cowan and Morey (2007) were supported; instead the results indicated that older studies should be replicated using current techniques to check reliability, if they still support either of the models mentioned, or is a new model required.