| The reasons for the development of career orientation assessment scale for junior middle school students in this study are as follows: Firstly,in the past decade,the country has attached more and more importance to vocational education.The Party and the state have issued a series of policies and systems aimed at promoting the development of vocational education,in order to release educational dividends and talent dividends through promoting the development of vocational education,among which the vocational diversion policy of the middle school entrance examination is one(Hu Yao et al.,2022).In 2021,the Ministry of Education clearly put forward the policy of vocational diversion for junior middle school students,and junior middle school students will face various choices after the vocational diversion.Secondly,the title and expression of the assessment scale need to be updated to more in line with the characteristics of contemporary junior middle school students;Thirdly,the development of the digital age requires the addition of a new dimension of career orientation.In 2022,the newly revised Occupational Classification Code of the People’s Republic of China,based on the 2015 edition of the Code,will include 158 new occupations that have been published in recent years,and marked 97 digital occupations for the first time.Therefore,based on the binary item response theory and text mining method,this study aims to form a localized assessment tool for junior middle school students’ career orientation in the digital era and enrich effective measurement methods for junior middle school students’ career orientation.In study 1,the career orientation scale of junior middle school students was developed based on item response theory,which was divided into three parts: first,the self-compiled topic formed the preliminary career orientation scale of junior middle school students;The second is to compile a formal scale of career orientation of junior middle school students based on item response theory.The subjects selected junior middle school students from 3 schools in Shandong Province,with a total of 2224 effective subjects.1133 students were randomly divided into the first half,and the formal scale was formed through one-dimensional test,model selection and item parameter analysis.Third,the brief version of the career orientation scale for junior middle school students was compiled by item response theory.The participants were selected as the last half of the scale,1091 of which were randomly divided into half.After the analysis of item differentiation,item content,and item information function peak value,the short version scale was formed and compared with the formal scale.In Study 2,based on the method of text mining,the career orientation of junior middle school students was predicted through the Chinese compositions of junior middle school students.On the basis of the scale data of study 1,the compositions were subsequently collected,and a total of 2180 valid matching texts,with an average length of 605 words,were established to establish the classification prediction model of career orientation traits.In this process,the difference between different classification labels and the prediction results of the model is compared,and the machine learning classification algorithm most suitable for this study is selected,and the validity of the short version scale is further verified.The results of this study are as follows:(1)A preliminary scale of junior middle school students’ career orientation,consisting of135 items and nine dimensions,including nature type,operation type,power type,research type,art type,business type,conventional type,information type and social type.(2)The subscales of the formal scale of junior middle school students’ career orientation meet the one-dimensional test,and the two-parameter Logistic model fits better in the model selection.(3)There are 124 items in the formal scale of junior middle school students’ career orientation.The differentiation of each subscale ranges from 0.419 to 2.078,the difficulty ranges from-1.265 to 2.999,and the peak value of average information volume ranges from 0.457 to1.090.There was a moderate positive correlation between scale and similar dimensions of criterion scale.(4)A simplified version of the career orientation scale for junior middle school students,including 50 items,the differentiation of each subscale ranges from 0.659 to 2.374,the difficulty ranges from-0.501 to 0.970,and the peak value of average information content ranges from0.686 to 1.720.There was a significant positive correlation between the scale and similar dimensions of the criterion scale.(5)In the comparison between the short version and the formal scale of Junior middle school students’ career orientation,the short version and the formal scale reached a significant high correlation,the short version scale’s each subscale item characteristic curve and test information function curve are similar to the formal scale.(6)In the process of establishing the trait model for predicting career orientation,the model with dimension score Z-score as 1 is the best predictor.(7)In the machine learning classification model,all dimensions of career orientation achieve good prediction effect on six different classifiers,among which the Logistic regression model has the best performance.For the prediction of medium and below inclination,the average accuracy rate is 0.85,the average recall rate is 0.90,and the average F1 is 0.87.(8)The logistic regression model predicted results of the brief version of the career orientation scale for junior middle school students were close to the results of the formal scale.The average accuracy rate of the model was 0.77,the average recall rate was 0.84,and the average F1 value was 0.81.This study draws the following conclusions:(1)The self-designed formal scale of junior middle school students’ career orientation and the short version of the scale are constructed with nine dimensions.The performance of the scale in discrimination,difficulty,and the peak value of item information function has reached an acceptable level,and the validity of the short version of the scale is good.The effectiveness of the short version of the scale has been verified in comparison with the formal scale.(2)Through text mining,a categorical prediction model of career orientation of junior middle school students is established.The logics-tick regression model has the best performance and effectively predicts the category of medium and below tendencies.The model prediction results of the compact version of the career orientation scale for junior middle school students are close to those of the formal scale,and the validity of the abbreviated version scale has been verified again. |
