Development Of Forced Selection Personality Test For Internet Recruitment

With the development of information technology and the increase of Internet penetration,online recruitment has become mainstream.In order to better serve job seekers and employers,Online recruitment companies are continuously trying to improve the accuracy of recruitment.Not only to match the resume and work experience,but also more attention should be paid to match implicit traits such as personality.The iceberg model of competence and Holland job-matching theory theoretically put forward the importance of personality for recruitment.However,for a long time,Online recruitment companies have measured personality traits through traditional Likert questionnaires such as the 16 PF and Big Five Personality questionnaires.The Likert scale is the single stimulus format.In high-stake situations,candidates will choose the options that are beneficial to themselves in order to obtain the ideal job.That will cause the response bias,which seriously affects the effectiveness and fairness of recruitment.This article takes the candidates as the research object.The purpose of the study is to develop the forced-choice scale to solve the problem of traditional likert format scale and to select terrific candidates for companies.This article consists of three studies:Study 1 use questionnaire compilation technology was used to collect questionnaires in two stages.First,30 questionable questionnaires were collected(15 males and 15 females,aged 22-27 years),and then 425 Likert questionnaires were collected online(76 males and 249 females,aged 19-31 years old),and finally through the definition of dimensions,analysis of difficulty and discrimination,social desirability statistics,and considering the balance of dimensions,24 questions under four dimensions are retained to form 3(Number of questions)× 8(number of question groups)forced selection questionnaire.Study 2 uses an experimental paradigm named“known groups test,” Through the recruitment situation,a fraud group(377 people,247 males,130 females,age 21-38 years old)and an honest group(297 people,121 males,176 females,age 18-25 years old).Likert questionnaire responded,verifying that the Likert scale is easy to falsify in the recruitment situation,and there are social biases and other reaction biases.Study 3 introduces the Thurstone IRT model,and solves the limitations of ipastive data.Additionally,we use(377 people,247 males,130 females,age 21-38 years old)questionnaires and verify the test performance of self-compiled mandatory questionnaires,and estimates the model parameters.The study draws the following conclusions:(1)Under the recruitment situation,candidates have response bias when using the Likert personality scale.(2)The forced-choice personality scale can fit the Thurstone IRT model well,and the model parameters can be well estimated,besides the reliability of IRT scoring method is better than the traditional scoring method.(3)The self-edited personality scale has a good fitting index,which can provide a reference for online recruitment.