Introduction¶
The Iowa gambling task is a study that was created to model human decision making. It seeks to explore our ability to learn and how we may react to varying levels of rewards and punishments due to the decisions we have made. The Iowa Gambling involves four decks of cards (A, B, C and D). Participants must choose a card from these four decks. Each deck choice is a single trial and participants complete a task of ninety five, one hundred or one hundred and fifty trials, depending on which study they are part of. The four decks are not all equal, and although there are variations from study to study, the premise of the study is that there are good decks and bad decks. The good decks reward participants for consistently choosing cards from these decks while the bad decks will result in a loss of money over time if they are consistently selected. There are other intricacies from deck to deck such as a variable win and variable loss function that changes over time. The participants are told their task is to maximise their profit by choosing correctly.
Uses of the Iowa Gambling Task¶
The Iowa Gambling Task has been used to test the decision making of healthy participants vs participants with damage to the prefontal cortex which may impact their decision making. The theory proposed is that healthy participants show an ability to learn from their choices and will tend towards the decks which maximise their earning potential. Conversely the unhealthy participants may not learn from previous choices or may even tend towards more severe punishments from certain bad decks.
Our analysis¶
We seek to explore the Iowa Gambling task using a clustering algorithm. The data used in this project comes from 617 participants across 10 studies. We explore the data to understand any insights or trends and then further investigate any beliefes we have around potential clustering with the k-means clustering algorithm.