Doliński (2018, this issue) deplores the decline of behavior observation in social psychology since the 1960’s and asks whether (social-) psychology is still a behavioral science. I question both, that there was a decline and that direct behavior observations are essential for a science of behavior. After all, behavior can also be inferred from outcomes and other traces of behavior. During the alleged heydays of behavioral observation, social psychology was threatened by a crisis partly precipitated by Wicker’s (1969) demonstration that verbal attitude measures were often unrelated to behavioral responses towards attitude objects. His critique was devastating, because social psychology at that time relied heavily on rating scales as dependent measure. The advance of the social cognition movement in the 1970’s was to provide social psychology with new techniques (e.g., priming, cognitive load, reaction time techniques) that eased the reliance on rating scales. At the same time, it became insufficient to merely show a relationship between an external event and a behavioral response and to rely on speculations about the internal processes that might have been responsible for this relationship. Instead, studies had to assess the cognitive and motivational processes assumed to link those external events, typically – but not always – using social cognition techniques. This required additional studies leading to a decline in the proportion of studies reporting behavioral observations. I illustrate this development with one of my own research programs and also suggest that in this example an outcome may be a more valid measure of behavior than behavioral observations.