Research On Cognitive Diagnostic For Reading Of Grade 3-4 Students

Objectives:(1)From the perspectives of education and psychometrics,based on the existing reading cognitive diagnosis test,integrating A and B sets of cognitive diagnostic tests that contain three cognitive levels of words,sentences and chapters to provide a more comprehensive diagnostic basis for children with reading difficulties.(2)Analyze the cognitive errors that occur during the reading process of students and classify their attribute mastering patterns.(3)Equivalence study on reading cognitive diagnosis test A and B volumes.Methods: Based on the cognitive cognitive process theory of reading,the main error types and wrong rules in primary school students' reading are used as the basis for the selection.Based on the existing research on reading cognitive diagnosis,this study analyzes 14 readings.Cognitive attributes are constructed from three levels: words,sentences,and chapters,and finally combined analysis.The test is divided into two sets of A and B,each set has 100 items.The expert assessment method is used to determine the demarcation points of the test.A total of 2measurements were carried out.A total of 1326 test rolls were issued and a total of 1265 were recovered.The recovery rate was 95% and the effective rate was 87%.Among them: 591 boys,510 girls;609 in the third grade,492 in the fourth grade.The anchor test non-equal design wasused to perform Tucker linear equivalence on the two sets of A and B sets.The student's attribute mastery probability,reaction mode and error rules were analyzed using the DINA model.Results:(1)Project quality: The difficulty coefficient of A volume is between [-3.183,2.608],the average difficulty coefficient is-0.262;the discrimination coefficient is between [0.406,1.225],and the average discrimination coefficient is 0.741;Between [0.103,0.40],the average guess factor is 0.249.The difficulty coefficient of B volume is between[-3.964,3.437],the average difficulty is-0.160;the discrimination parameter is between [0.408,1.427],the average discrimination coefficient is 0.719;the guessing parameter is between [0.105,0.350],The average guess factor is 0.247.(2)Test quality: A-volume Cronbach's? coefficient is 0.900,split-half reliability is 0.629,maximum information function is 17,and CDA-based test reliability is 0.84.The B-volume Cronbach's ? coefficient is 0.907,the split-half reliability is 0.701,the maximum information function is 16,and the reliability based on the cognitive diagnosis test is 0.83.Confirmatory factor analysis showed that the relationship model between each part of the project and the property was reasonable(RMSEA was less than 0.05,CFI and TLI were both greater than 0.90);The correlation between the scores of the two sets of tests and the final scores of the Chinese language were 0.497 and 0.652respectively;there were significant differences in the pass rates of the twotypes of schools(emphasis and non-emphasis)on each attribute.(3)Test demarcation points and detection rate: A volume demarcation is divided into 45 points,detection rate is 7.3%,B volume demarcation is divided into 43 points detection rate is 6.8%;(4)anchor By quantifying the equivalent results,the linear conversion equations under W = 1 and W =0.456 weights are: B = 1.018 A + 0.46,B = 1.018 A + 0.45.(5)Cognitive diagnosis results based on the DINA model: A1(shape-to-sound combination),A2(sound-sense combination),A3(synonymous combination),and A7 of the sentence part of the word part in the average grasp probability of all 14 attributes(Sentence structure understanding)and chapter A10(Articles and phrases understanding),A11(content analysis summary)several attributes are best mastered;A6(emotional color understanding),A8(sick sentence analysis)two attributes are the worst.From the typical property mastering model,72.67% of the subjects were classified in the word part,84.38% of the subjects were classified in the sentence part,and 86.13% of the subjects were classified in the text part.Almost all of the 51 children with dyslexia were classified into non-ideal mastery models.Conclusions:(1)The quality of the test items is good.The test has good reliability and validity and the indicators reach the requirements of Psychometrics.(2)The equivalent values of the two sets of A and B volumes are realized,and the results of the equal values under differentweights are not much different.(3)Students have the worst grasp of emotional color understanding and disease sentence differentiation.(4)The 51 students with dyslexia selected were almost all classified into atypical mastery mode,There are different levels of cognitive deficits in terms of words,sentences,and texts.