Reactivity and unreliability of husbands as participant observers

This study examined the reliability and reactivity of participant observation. Twelve couples engaged in two discussions in a laboratory, one on a conflict topic and one on a consensus topic. For the eight couples in the experimental sequences, a baseline discussion was followed by participant observation in which the husband recorded his wife's use of "I" while continuing to converse. The four couples in the control sequences conversed without participant observation. Compared to observations made by a trained observer, the husbands were unreliable observers, underestimating their wives' use of "I." Participant observation was reactive (P<.01), with more wives decreasing their use of "I" in the experimental sequences than in the control sequences (p<.094). The topic of discussion did not differentially influence reliability or reactivity.