Research On Validity Of College Entrance Examination In Mathematics In Hunan Province In 2007

Objectives: To provide the theoretical and practical reference for the reform of the College Entrance Examination proposition based on the new curriculum standard; To carry on a research based on a new theory of the validity.Methods: The methods to verify the validity and design the questionnaires are mainly to analysis the documents. The approaches to gathering evidence include test and questionnaire survey. The ways of data analysis include description statistics, correlation analysis, t-test, analysis of variance, confirmation factor analysis as well as some nonparametric tests.Results: The proportion of agreement on the papers of the Liberal art and science which evaluated by experts is .84 and .73 . The differences of the factors investigated which include response process, anxious and stress, understanding the meanings of the titles and using the strategies are significant at the level of 0.01; In the terms of the examinees' train of thought in resolving the questions in the test and the degree on the examinees' train of thought or methods which consistent with the intended cognitive processes and among the six questions which are restricted by the difficulty of the paper, the first three subjective questions in the papers of liberal art and science are in the range of.303～.754 and the others are in the range of.045～.169. Difference of samples in analysis of variance is significant; The Alpha coefficient of the papers of the liberal art is in the range of.67～.78,and the science is in the range of .69～.77. In terms of the correlation coefficient between items and total score, the liberal art is in the range of .11～.88 and the science is in the range of .16～.83.Their correlation coefficients are remarkable. The structure model of the prospective intelligence is supported by the results of confirmatory factor analysis; The result of Pearson correlation is in the range of.2～.4 which is between the college entrance examination scores on the samples of the Liberal art and science and the higher mathematics scores at the university, and the result of the division is in the range of.81～2.93.The results of the Questionnaires of the washback show: The difference in the proportions of the "well" and the "bad" that are chosen by teachers is significant, while the difference of the "well" and "general" is not significant. The difference in the total result of the students' answers is significant and the number of the "well" is at an advantage. The result of the ANOVA suggests that the difference of the samples is significant.Conclusions :The papers of the college entrance examination in Mathematics in Hunan in 2007 met the requirements on the validity of the content; In terms of the response process, the anxious is the important factor that affects students in exerting their intelligence in the test. The students' train of thought in problem-solving fits in with the construction required in the test when the difficulty of the papers is appropriate; In terms of the internal structure and the forecast effect, the indexes in the test met the need required in psychometrics; The positive effects on the teachers and the students are more than the negative effects respectively.