Reactivity is the tendency that people have to make more effort to meet guidelines and standards when they know their performance is being assessed. Reactivity can affect a behavioral study when the team is observing the experiment. The team has two choices to maintain reliability of experimenter performance across participants: observe the experimenters covertly or observe them with every participant.
Reactivity can also occur in participant groups who take more than one assessment or who have been exposed to other assessments. Bell (2013) indicates that experimenters need to randomize the order in which they conduct multiple assessments to assess any effect of the order of administration on the results. Any such effect should be accounted for as a confounding factor. If experiments fail to randomize the order of administration, they may unwittingly allow a reactive effect into their data sets, which skews their results.
Reference:
Bell, A. (2013). Randomized or fixed order for studies of behavioral syndromes? Behavioral Ecology, 24(1), 16-20. doi:10.1093/beheco/ars148
Would or do you agree with Bell? (Why or why not?) When he indicates that, "experimenters need to randomize the order in which they conduct multiple assessments to assess any effect of the order of administration on the results."