The Study Of Response Category Number Of Personality Test Based On The Framework Of CTT And IRT

Personality test currently has become one of the widely used psychological measurement tool.As a psychological scale,only with good quality can accurately measure the character level of individual.In fact,for personality test,the reliability and validity are not very high all along.Response category,as an important part of test,the number of which is also an important factor affecting test quality.Therefore,in order to further improve quality of personality test,it is necessary to conduct research from the perspective of response category.The researcher firstly got some reasons triggering result differences through analyzing previous studies in-depth,then combined these factors to explore the relationship between response category number and personality test quality.This study chose 7 personality tests,a total of 153 items.After the researcher set response category number from 2 to 7,by a completely random way,these tests were measured to 16 years old and above groups in form of questionnaire.Lastly,1383 valid questionnaires were received.When processing data,in order to compare the advantages and disadvantages of different methods,the researcher used traditional CTT,dominance IRT and ideal point IRT three methods to analyze test data.Specifically,in the analysis of traditional CTT,internal consistency reliability and structure validity were tested.While adopting IRT methods,test data firstly should satisfy unidimensionality assumption.Then the researcher compared test information function and model-data goodness of fit based on the analysis of dominance model(2PLM/GRM)and ideal point model(GGUM).Lastly,the researcher acquired the appropriate number of response category for personality tests.The study results are as follows:(1)The suitable response category number for personality test is more than 2.Based on the analysis of CTT,Traditional CTT order scale with 6 response categories,Dominance IRT order scale and Combined order scale with 5 response categories,E scale with 7 response categories,N scale,Pa scale and Sc scale with 4 response categories,would get high reliability and validity.Based on the analysis of IRT,the suitable numbers for Traditional CTT order scale,Dominance IRT order scale and Ideal Point IRT order scale are 3,4,and 6respectively.For E scale,N scale,Pa scale and Sc scale,the suitable numbers are 7,4,4,and 3respectively.(2)In the aspect of data processing,ideal point IRT method is more appropriate for the new type of personality tests.In view of traditional personality tests,it is advisable to adopt dominance IRT or traditional CTT method.(3)Measurement purpose and test length two factors can affect the relationship between response category number and test quality to some extent.However,the effect of item homogeneity is not obvious.