Although it has been suggested that situational leadership theory (SLT) is a valid theory at the individual and group levels (Kim et al., 2000, 2006; Hersey & Johnson, 1996), most research investigated SLT at the individual level. And except Hambleton and Gumpert’s (1982) study, most studies failed to support SLT. These results suggest that SLT may not be a valid theory at the individual level. However, in this paper we viewed SLT as a valid theory at the group level in the following reasons: First, SLT has been developed from Reddin’s (1967) 3-dimension leadership style theory in which a leader consistently shows behaviors for a certain period time. Accordingly, a leader tends to show similar behaviors toward subordinates in the same group. Second, Hersey et al. (1996) suggested that a leader has two types of leadership style (primary and secondary) and must have one primary style and may have 2 or 3 secondary styles. However, if a leader shows different behaviors in different situations, we cannot identify a leader’s style. Therefore, in SLT a primary style rather than secondary styles should be considered. Third, SLT classified leader behaviors on the basis of two dimensions of leader behavior (consideration and initiating structure) developed from Ohio State University Leadership Studies (Stodgill & Coons, 1957). They proposed that a leader homogeneously behaves toward subordinates in the same group (e.g., Dansereau et al., 1975). So this study theorized and tested a situational leadership theory at the group level of analysis on the basis of the assumption that a leader homogeneously behaves toward members in the same group and the conceptualization of performance and maturity at the group level. In this paper group maturity was defined in terms of group efficacy (group members’ belief of the group’s ability to accomplish group task or goal) and group cohesiveness (group members’ willingness to make effort). The data from 308 members of 83 branch offices of Kwangju and Chonnam in 7 insurance companies showed that a leader homogeneously behaves toward members in the same group and that performance and maturity are group-level variables. The results also showed that a directive leader is effective in the situation of lower level of group maturity and that a delegative leader is effective in the situation of higher level of group maturity. The results also showed that a coach style leader is effective in the situation of high level of group cohesiveness and low level of group efficacy and that a participative leader is effective in the situation of low level of group cohesiveness and high level of group efficacy (medium level of group maturity). These results suggest that situational leadership theory is a valid theory at the group level of analysis. In addition the result suggest the following implications: First, a leader should attempt to show an appropriate behavior that fits into a group situation. Second, an organization should know about group maturity level and leader styles and then place a leader who leadership style fits into the group situation. Limitations and future research directions are discussed in the discussion section.