Even though a great deal of research has examined the relationship between demographic variables and interview outcomes such as ratings and referrals (Arvey, 1979; Arvey & Campion, 1982; Dipboye, 1992; Harris, 1989), it is still difficult to make decisive judgments about these relationships (Arvey, 1979; Arvey & Campion, 1982; McIntyre, Moberg, & Posner, 1980; Raza & Carpenter, 1987). In fact, many argue that the modest validity coefficients reported in the literature are caused, in part, by this lack of understanding (Ugbah & Majors, 1992; Wiesner & Cronshaw, 1988). Similarly, a number of researchers suggest that methodological flaws have contributed to this lack of consistency across studies (Arvey & Campion, 1982; Harris, 1989; Jablin & McComb, 1984; Schmitt, 1976). First, many studies used fictional environments (typically the behavioral lab of the university), and fictional job seekers (typically college students) who are vying for fictional jobs. When college students are used as both applicants and interviewers, we must immediately question the external generalizability of these studies (Parsons & Liden, 1984). Second, past selection interview studies have been criticized for failing to gather an in-depth awareness of interview dynamics (Eder & Buckley, 1988; Liden & Parsons, 1986). Specifically, most studies have focused exclusively on correlating post-interview ratings or rankings and subsequent job performance, with little regard for what actually occurred during the course of the interview (Dipboye, 1992). To some extent, past research concludes that the events occurring during the interview have little effect on subsequent behaviors. Unfortunately, this narrow focus has permitted researchers to totally disregard individuals' ability to encode and decode information. Also, this suggests 'individuals are unable to adjust their behaviors during the course of the selection interview. Many have suggested that gathering a richer understanding of the patterns of both the applicant and the interviewer would tell us much more than we already know about the selection interview process (Dipboye, 1992; Harris, 1989). Structure and Purpose The reason for examining patterns occurring during employment 'interviews stems from the importance that has on interview outcomes (DiSalvo & Larsen, 1987; Hollandsworth, Kazelskis, Stevens, & Dressel, 1979). For example, Bjorkquist (1987) suggests that the most influential factor affecting applicant is oral skills. Similarly, Kinicki and Lockwood (1985) report that an applicant's ability to express thoughts and ideas explains a significant portion of variance in the applicant's rating score. Einhorn (1981) concludes that verbal, nonverbal, and rhetorical skills are the most influential predictors of applicant in a sample of college job seekers. In sum, a number of studies echo Cissna and Carter (1982), who conclude that communication is the single most important determinant of success (p. 57) in the employment interview process. In the present study, we looked on the interview as a dyadic process rather than an outcome. Specifically, we analyzed selection interviews on a statement-by-statement basis. A coding scheme developed by Ellis (1976) was used to code statements made during the interview. Trained field researchers coded videotaped employment interviews by examining each statement in relation to the statement that came before it. This comparison resulted in each statement's placement into one of five categories: dominance, structuring, equivalence, deference, or submission. Once this was completed, an interviewer's coded statement and the applicant's coded response to that statement were paired and rated for effectiveness, one pair at a time. Coding statements in pairs allows the interactive nature of the interview (i.e. …