--- - |2+ Abstract Several studies suggest infants and young children prefer prosocial over antisocial individuals and expect individuals will selectively approach prosocial rather than antisocial others. There is, however, confusion regarding (a) the methods used to assess preference versus those used to assess expectations and (b) the interpretation of previous studies. In the current study, utilizing eye‐tracking technology, we dissect children and young adults' preferences and expectations after watching the hill paradigm, in which one character (the Helper) helps a protagonist ascend a hill whereas another character (the Hinderer) prevents the protagonist from doing so. Neither children nor young adults displayed a looking preference for the Helper over the Hinderer. With respect to expectations, children shown an animation in which the protagonist approached its Helper looked significantly longer than children shown an animation in which it approached its Hinderer, whereas young adults' looking times were comparable across the two scenarios. We argue longer looking time to the Helper‐approach animation may reflect that children expect the Helper to move in concert/coordination with the protagonist and are surprised when this does not occur. We discuss our findings in relation to arguments about boundaries between different techniques used to measure preferences versus expectations in developmental research. Highlights We address whether clear boundaries exist between developmental paradigms that aim to test preferences, predictions, and expectations. Eye tracking was used to collect data on children's and young adult's looking behaviour on a commonly used paradigm in social evaluation research. We demonstrate that measures of preferences, predictions, and expectations can be extracted from a single paradigm and, therefore, that the boundaries between paradigms are far from clear cut. - Infant and Child Development, Volume 27, Issue 4, July/August 2018.