In Mischel et al.’s (1972) famous ‘marshmallow test’ four year olds were given a choice between eating a marshmallow (or similar treat) now or waiting and receiving an extra marshmallow at the end of the experiment. The marshmallow was placed on a table in front of the children and left unattended to maximise temptation. In this way self-control is seen as a finite resource which can be depleted. In a follow-up study a positive correlation was found between delayed gratification and SAT scores, with the correlations stronger for quantitative test scores than verbal test scores (Shoda et al., 1990). Importantly, studies have shown that self-control is a better predictor of academic outcomes than IQ (Duckworth & Seligman, 2005).
However, self-control is also shaped by social interactions, particularly perceptions of trust. In a modified marshmallow experiment Kidd at al. (2013) added a preceding stage where perceptions of researcher reliability could be shaped. In this pre-stage, children were promised new crayons to draw with while they waited for the marshmallow experiment. One group received the new crayons as promised while the other group received old, clearly used crayons. Children who received the promised new crayons waited significantly longer than those who received the old crayons. Suggesting that self-control is strongly shaped by reasoned beliefs of the reliability of promises made by the researchers. Michaelson et al. (2013) found similar results for experiments with adults that showed social trust having a causal role in the willingness of individuals to delay immediate gratification.
The misBehaving Economist 