Date of Award

Spring 1-1-2016

Document Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Yuko Munakata

Second Advisor

Naomi P. Friedman

Third Advisor

Timothy Curran

Fourth Advisor

Richard Olson

Fifth Advisor

Michael Mozer


Working memory (WM) capacity – the ability to store and process information in primary memory – is a core ability that is central to most human behavior. Theoretical characterizations of WM capacity aim to identify the cognitive constructs that underlie this ability and to explain the close relationship between WM and short-term memory (STM). These characterizations are heavily based on behavioral research within the field of cognitive psychology. A relatively unexplored method that may provide a deeper understanding of the cognitive constructs that underlie WM is the use of a twin design. Twin designs allow for more detailed examination of underlying processing in that common and unique variance across constructs can be further decomposed into variance that is driven by the same or distinct sets of genes or the same or distinct environmental factors. This dissertation contributes to existing theories of WM in three ways: 1) The first aim was to determine whether existing research suggests that genes more strongly influence individual differences in WM task performance than STM task performance. Results of the meta-analysis revealed a tendency for WM tasks to be more heritable than STM tasks. 2) The second aim was to determine whether WM and STM constructs are distinguishable by decomposing their relationship into candidate cognitive subprocesses - attention control (AC) and long-term memory (LTM) - and then further decomposing those findings into genetic and environmental influences. The majority of the overlapping variance between WM and STM also covaried with AC and was almost entirely genetic in nature. 3) The final aim evaluated how much of the WM-intelligence relationship can be accounted for via mediation with the candidate cognitive constructs from Aim 2 and then determines the extent to which genes versus environmental factors are contributing to that relationship. AC and LTM significantly mediated the WM-intelligence relationship and STM played a negligible role. The genetic models revealed that the majority of overlapping processing for WM and intelligence was driven by genetics, which was mainly due to overlap with AC and only minimally due to STM.