ABSTRACT
Three methods of data collection have traditionally been used to investigate human behaviour in a social situation: interviewing, participant observation and the small-group experiment. Each of these methods has been subjected to criticism and has fairly serious disadvantages. The information collected from interviews may be clouded by the problem of recall. Participant observation is costly, because it demands a high ratio of observers to observed. Further, this method presents the problem of the selection of relevant behaviour. The small-group experiment removes the behaviour from its usual context and, hence, leads to a lack of clarity about the importance of effects observed to general behaviour.
