This study examines how social metacognition (including evaluations and questions) affected micro-creativity during group problem solving. Twenty groups of high school students were videotaped as they solved a mathematics problem. Analyses of the 2,951 conversation turns showed that the likelihood of a correct contribution (CC, a measure of micro-creativity) was higher after a group member expressed a wrong, new idea, correctly evaluated an idea, or justified an idea. In contrast, the

more »

... of a CC was lower after a group member disagreed rudely or agreed. Meanwhile, group-level properties (racial diversity, gender diversity, and degree of status differences) did not significantly affect the likelihood of a CC. A CC was more likely after a justification in successful groups than in unsuccessful groups. CCs did not occur uniformly, as some time periods had many CCs while others had few CCs. Furthermore, agreements and correct evaluations had different effects across time periods within a group. Past research on the development of original ideas that are useful or influential (creativity) has largely focused on individuals (Sternberg & Lubart, 1999) . However, the explosion of information and specializations will increasingly require teams with diverse skills and knowledge to create innovations (group creativity; Sawyer, 2004). Still, researchers have not systematically examined how group processes affect creative moments (group micro-creativity). This study takes a step in this direction by analyzing how micro-creativity occurs in twenty groups of students solving an algebra problem. By understanding how group processes affect micro-creativity, group members can work together more creatively. In this paper, creativity refers to the "small c" creativity of ordinary people in daily life, not the "big C" creativity of new knowledge or products that substantially affect society (Sternberg & Lubart, 1999) . Furthermore, micro-creativity is measured via a correct, new idea (correct contribution, CC). Hence, the CC's "new" component is relative to the group members' experiences. The CC's "correct" component refers to the intersection of the problem situation and the relevant field's principles (in this case, algebra). Thus, a CC is a new idea for the group members that is consistent with both the given problem situation and algebra principles. I modeled group, student, and speaker turn characteristics that helped or hindered micro-creativity during group problem solving in a high school algebra class. This study contributes to the research literature in four ways. First, I showed when CCs occur, whether they occur uniformly through a problem-solving session or more frequently in some time periods than in others. Second, this study showed how the local time context created by prior speakers' actions (CCs, justifications, etc.) and interactions affected the likelihood of creating a CC. Third, I tested whether the above effects differed across groups or across time periods. Lastly, I tested these hypotheses with a new statistical discourse analysis method, dynamic multilevel analysis (Chiu & Khoo, 2005) . Group processes and micro-creativity Diverse views and social metacognition can aid CC creation (Chiu, in press; Paulus & Brown, 2003) . Meanwhile, public self-image (face) and status concerns can hinder creation of CCs (Brown & Levinson, 1987) . Group problem solving processes that aid micro-creativity Diversity Groups with diverse views can create more ideas. Specifically, group diversity in nationality or ethnicity can increase the number of perspectives, number of ideas, and quality of ideas (McLeod, Lobel, & Cox, 1996) . Furthermore, groups that create many ideas, representations, and solution proposals, are more likely to find a correct or optimal solution (Paulus & Brown, 2003) . Group members with diverse views can also build on idiosyncratic ideas to create CCs through processes such as sparked ideas, jigsaw pieces, and creative misinterpretations (Paulus & Brown, 2003; Chiu, 1997) . Comments by one person (e.g., a key word) might spark another person to activate related concepts in his or her semantic network and propose a CC. Like fitting jigsaw pieces together, group members also can put together different pieces to create a CC. Lastly, a person might also misinterpret another person's incorrect idea to create a CC. Thus, even wrong, new ideas can lead to CCs. Group members' diverse views can help identify flaws and refine incorrect ideas. (Orlitzky & Hirokawa, 2001) . By creating more ideas (including wrong ideas) and evaluating their validity correctly, groups