| In this an era of information explosion, it is difficult for us to obtain and extract the news from the vast ocean of information which is we needed, meanwhile, to draw a conclusion according to the news which is valuable at this moment. As a pervasive decision-making process in everyday life, estimation becomes more and more important. The poor quality of children’s performance, and the positive relation between numerical estimation and overall mathematics achievement, have led educators to assign a high priority to improving numerical estimation from ages three to six.The present study regarding the quality of5-to6-year-olds’number line estimation tasks with a range of numbers from0to100, especially on their numerical magnitude representations and numerical estimation accuracy, replicated and modified the measures which were taken to develop shifts from a logarithmic pattern to a linear pattern of estimates in former foreign findings aiding children whose estimates better fit a logarithmic than a linear function. These measures were Aiming at Logarithmic-Linear Discrepancy with Feedback, Number Categorization, and Playing Curvilinear Number/Color Board Games. The present study was designed to demonstrate three trainings’ contributions to both developmental changes and individual differences on0-100number line estimation tasks, to compare these measures’training effects, to extend the linear representation to numerical ranges where kids previously used logarithmic representations, to enrich intervention theories of numerical estimation, and to help teachers explore ways to enhance preschooler’s numerical knowledge and ability to acquire new numerical knowledge.We performed two experiments. The purpose of Experiment1was to examine whether5-to6-year-olds’representations of the magnitudes of numbers between0and100are logarithmic, to identify individual these children whose estimates fit a logarithmic function better than a linear one, and to test the numerical region in which the discrepancy between logarithmic and linear representations is greatest. The analysis of the data suggested the following results:1. The best fitting logarithmic function fit5-to6-year-olds’median estimates better than did the best fitting linear function. This finding with the group medians was mirrored by the function that provided the better fit to individual preschooler’s estimates. The linear function provided the better fit for20%of5-to6-year-olds, whereas the logarithmic function provided the better fit for80%. And the boys’and the girls’numerical estimation accuracies were at the same level on0-100number line tasks.2. The greatest function differences in estimates occurred on numbers in7-31, where the discrepancy between a logarithmic and a linear representation of the values on a0-100number line (with both functions constrained to pass through0and100) was greatest.The children whose estimates fit a logarithmic function better than a linear one subsequently participated in Experiment2. They were randomly assigned to three experimental conditions:Aiming at Logarithmic-Linear Discrepancy with Feedback, Number Categorization, or Playing Curvilinear Number/Color Board Games, were trained to transit from use of a logarithmic representation to use of a linear representation, and completed a posttest. On the posttest, children in all three experimental conditions were presented the same28problems without feedback as in Experiment1. The children’s estimates in Experiment1provided pretest data, which was used as a point of comparison for their subsequent performance. The conclusions were as follows:1. Aiming at Logarithmic-Linear Discrepancy with Feedback promoted extensions of linear representations to new numerical contexts to the extent that the experiences highlight discrepancies between logarithmic and linear representations of numerical magnitudes and make clear the appropriateness of the linear representation in the new contexts.2. The instance of divide-and-conquer approach--Number Categorization--provided a useful means for reducing the cognitive demands of complex tasks and for promoting simultaneous attention to absolute qualities of numbers and their relations, thus making good performances on estimation tasks possible well.3. The benefits of Playing Curvilinear Number Board Games extend to a variety of aspects of early numerical understanding:numerical order, knowledge of numerical magnitudes, counting, and numeral identification. All of these are foundational skills that contribute to the linearity of estimation increasing substantially from pretest to posttest.4. Comparing to Playing Curvilinear Number Board Games, Aiming at Logarithmic-Linear Discrepancy with Feedback and Number Categorization improve numerical estimation from ages five to six significantly, and between these two kinds of intervention methods had no difference. |