The effect of anchors and social information on behaviour.

We use a 'multi-player dictator game' (MDG), with 'social information' about the monetary transfer made by a previous dictator to a recipient, to examine whether average contributions as well as the behavioural strategy adopted are affected by the first amount presented (the 'anchor') using a sequential strategy elicitation method. We find that average contributions are positively affected by the anchor. The anchor is also found to influence the behavioural strategy that individuals adopt, such that low anchors significantly increase the likelihood that players will adopt unconditional self-interested strategies, whereas high anchors increase the likelihood of adopting giving strategies. The distribution of strategies-and hence, the distribution of behavioural 'types'-is therefore affected by the initial conditions of play, lending support to the notion that behavioural strategies are context dependent.

Examining charitable giving in real-world online donations.

The current study uses big data to study prosocial behavior by analyzing donations made on the GoFundMe platform. In a dataset of more than $44 million in online donations, we find that 21% were made while opting to be anonymous to the public, with survey results indicating that 11% of these anonymous donations (2.3% of all donations) are not attributable to any egoistic goal. Additionally, we find that donors gave significantly more to recipients who had the same last name as them. We find evidence that men and women donated more when more donors of the opposite sex were visible on the screen at the time of donating. Our results suggest that men and women were both significantly affected by the average donation amounts visible at the time of their decisions, and men were influenced more. We find that women expressed significantly more empathy than men in messages accompanying their donations.