Imitation, or mimicry, is a common occurrence during social encounters, and involves 58 spontaneous copying of others' actions and gestures [4]. Although such behaviour rarely 59 reaches conscious awareness for either interaction partner, it subconsciously signals a desire to 60 affiliate and build rapport [5]. For example, people who are imitated are bigger tippers [6], 61 donate more to charity [7], engage in prosocial behaviours [7,8,9, 10] and indicate liking 62 people who imitate them more than those that do not [6]. Clearly, then, imitation can play an 63 important role in guiding social interactions. To clarify the role imitation can play across 64 different social contexts, recent research has started to identify its antecedents [1, 2]. For 65 example, prosocial priming can increase imitative behaviour [1]. Thus, there exists a bi-66 directional relationship between imitation and prosociality; those who are imitated behave 67 more prosocially and those who are prosocially primed imitate more. These studies have all 68 employed observational techniques to study imitation, with the measurement being the 69 frequency of observed copying behaviours during live social interaction. 70Other researchers devised an index of imitation that is based on reaction time measures. 71The automatic imitation task [11,12] is an example of a stimulus-response-compatibility (SRC) 72 paradigm, referring to the fact that people cannot help but be affected by the presence of an 73 irrelevant stimulus feature [13,14]. In one well-established automatic imitation task, 74 . CC-BY 4.0 International license It is made available under a (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint . http://dx.doi.org/10.1101/333880 doi: bioRxiv preprint first posted online May. 29, 2018; 4 individuals are instructed to respond to a number cue by lifting their index or middle finger. 75Concurrently, participants either observe a congruent or incongruent finger movement. 76Reactions times (RT) are longer in the incongruent compared to congruent condition and this 77 difference is thought to signify the cost of inhibiting an imitative response [1,15]. Here, then, 78 imitation is captured as the time it takes to suppress the urge to copy an observed action and 79 prioritise one's own action. The tendency towards imitation (incongruent RT less congruent 80 RT) will hereafter be referred to as the congruency effect. 81A handful of studies have explored the effects of prosocial priming on automatic 82 imitation [16,17, 18]. Priming is thought to operate by subtly triggering a goal that 83 unconsciously guides behaviour [19]. The logic of these studies is that a prosocial prime would 84 activate a goal to affiliate and that this goal would be achieved through increasing the tendency 85 to imitate [16]. Despite using slightly different variants of the automatic imitation task and 86 different experimental designs (see Method section, T...

Imitation and perspective taking are core features of non-verbal social interactions. We imitate one another to signal a desire to affiliate and consider others’ points of view to better understand their perspective. Prior research suggests that a relationship exists between prosocial behaviour and imitation. For example, priming prosocial behaviours has been shown to increase imitative tendencies in automatic imitation tasks. Despite its importance during social interactions, far less is known about how perspective taking might relate to either prosociality or imitation. The current study investigates the relationship between automatic imitation and perspective taking by testing the extent to which these skills are similarly modulated by prosocial priming. Across all experimental groups, a surprising ceiling effect emerged in the perspective taking task (the Director’s Task), which prevented the investigation of prosocial priming on perspective taking. A comparison of other studies using the Director’s Task shows wide variability in accuracy scores across studies and is suggestive of low task reliability. In addition, despite using a high-power design, and contrary to three previous studies, no effect of prosocial prime on imitation was observed. Meta-analysing all studies to date suggests that the effects of prosocial primes on imitation are variable and could be small. The current study, therefore, offers caution when using the computerised Director’s Task as a measure of perspective taking with adult populations, as it shows high variability across studies and may suffer from a ceiling effect. In addition, the results question the size and robustness of prosocial priming effects on automatic imitation. More generally, by reporting null results we hope to minimise publication bias and by meta-analysing results as studies emerge and making data freely available, we hope to move towards a more cumulative science of social cognition.

A variety of subtle social cues, including gaze behaviour, are used to form impressions of others. For example, if another’s eye-gaze reliably helps or hinders us while we complete a task, we incidentally form a positive or negative impression about them. In real life, people are rarely so consistent in their behaviour, and they are often encountered in dynamic group contexts. To date, however, it is not yet known how incidental impressions are affected by either changes in target individual’s behaviour over time group, or by the group’s behaviour. To better understand how impressions are formed when subtle social behaviours change valence over time, we manipulated helping behaviour both at the level of the individual (Experiments 1-3) and the wider group (Experiments 4 & 5). Contrary to the idea that first impressions are hard to change, we found no evidence that impressions were driven by initial behaviour (primacy effects). Rather, people tended to form impressions based on the most recent behaviour, with some influence from the overall, average behaviour. In addition, we found that individuals’ behaviours appear to be viewed more or less favourably, depending on the behaviour of the wider group. Overall, we demonstrate that impression formation based on subtle social cues is not dominated by a single process, but instead reflects a complex product of cognitive mechanisms that integrate average valence over time, the direction of behaviour changes, the recency of observed behaviour, and the group context in which the behaviour is observed.