be demonstrated by speaking out the answers in the group, strategies of deep processing and using words in
context were involved. Touching-and-moving words on the screen or typing in words (T) involved strategies of
using words and retrieving lexical information learned before.
After the coding of 5,304 actions, some sequential patterns were immediately noticed (Bakeman, & Gottman,
1997; Sung, Chang, Lee, & Yu, 2008). A typical sequential pattern observed in the group leaders’ collaboration
in solving crossword puzzles was that they read the questions first to their group members (R) and discussed to
clarify what the questions meant with their group members (D); after the clarification, they decided on answers
(A) and typed in answers (T). To record and tabulate sequences of actions in Frequency Transition Table (Table
2), we began with action codes in the rows, “starting action”, and found their following action codes in the
columns, “subsequent action.” Two raters identified 63 and 38 patterns, respectively, and the inter-rater
reliability reached a substantial level (Kappa=.62).