Description: In this study, the researchers want to explore the effects of self-paced dynamic assessment on promoting elementary school children's performances in solving mathematics word problems which require multiple steps of arithmatics. The concepts of dynamic assessments are based on the notion of Zone of Proximal Development, which contends that all children can achieve the highest level of development with adequate instructions (Vygotsky, 1978/1997). Among the models of dynamic assessments, scaffolding is the most effective (Lu, Lin, & Lu, 2004; 2005; Swanson, 2001). In previous studies (Lu, Lin, & Lu, 2004; 2005), the researchers designed models of dynamic assessments and applied the models to children with mathematics learning difficulties. All of these models were easily to apply with minimal trainings, but it cost a lot of efforts to administrate the dynamic assessments for that they need to be administrated individually. In this study, the researchers designed a model of self-paced dynamic assessment which could be administrated in group to minimize the costs of administration. Methodology: The test items and prompts were completed and the participants will take the tests in April 2006.Subjects There will be 120 fourth graders from 2 schools participating in this study. Participants will be randomly assigned to the treatment or controlled group.Treatment sessions: the self-paced dynmaic assessment There will be 4 treatment sessions, in each session, the participants have to solve 10 items of word problems in group. The treatment group will receive the self-paced dynamic assessment in which a series of pre-set prompts of each items are covered by crayon and the examinees can decide if they want to scratch off the crayon to read the prompts during the session. The controlled group receive the same items without any prompts.Dependent measures Participants performaces on a test of 20 word problems will be searved as the dependent measures. There will be two kinds of items: Near Transfer which the items are similar to the items in the treatment sessions but with different numbers; the Far Transfer which the items are different from the items in the treatment sessions but need to apply similar compuation procedures. All of the items have been taught in the fourth grade curriculumn in Taiwan.Procedure The participants will receive the treatment sessions in group during their school days. The dependent measures will be administrated after the treatment sessions. Conclusions: Group diffences of the dependent measures will be compared. If the treatment group perform significantly better than the controlled group, the researchers can contend taht the self-paced dynamic assessments are effective in promoting children's performances in solving mathematic word problems. The procedure of self-paced dynamic assessment used in this study is congruent with the concepts of dynamic assessments and can be administrated in group. As a result, the costs of applying dynamic assessments can be minimized, and children's performances on solving mathematics word problems will be maximized.