【作者】 熊江玲；

【导师】 范晓玲；

【作者基本信息】 湖南师范大学 ， 基础心理学， 2002， 硕士

【摘要】 目的：编制软件开发人员心理素质测试问卷，使之实现初步的计算机化，以期为我国大型企业选拔具有良好潜质的软件开发人才提供有效的心理测试手段。 方法：1、量表的编制：在对软件开发人员进行工作分析的基础上，确定量表结构，建立本量表的项目库，采用经验法与传统的项目分析技术(如区分度技术)相结合的方式，结合专家评定进行项目筛选，确定项目。2、正式施测：在我国部分著名企业华为公司等取样，样本总数为72例，控制影响其开发水平的影响因素如性别等，对测试结果进行统计分析和逻辑分析。3、信度检验：(1)同质性信度：利用正式施测样本资料，计算量表的一致性系数。(2)分半信度：利用正式施测样本资料，计算量表的奇偶分半相关系数。4、效度检验：(1)内容效度。(2)结构效度：采用因素分析和一致性分析来说明量表的结构效度。(3)实证效度：分别以30名新员工的培训成绩排名、试用期满后转正时的其直接主管的评价排名及转正答辩成绩排名为效标，计算三者与测验分数排名的相关。 结果：1、经过对试测量表及试测结果的测量学和逻辑分析，编制了软件开发人员心理素质测试量表，该量表共有96条项目，分为三个维度：能力、性格、动力。2、项目分析结果表明，各项目的鉴别力指数在0.2以上，区分度较好。3、各维度与总分的相关较高(0.650∽0.882)。4、量表的一致性系数为0.8081；奇偶分半相关系数为0.8362。5、按各指标进行因素分析得到五个特征值＞1的因子。6、心理素质总分与培训成绩排名、试用期满后转正时的直接主管的评价排名及转正答辩成绩排名的相关系数分别为0.406、0.467、0.367。 结论:本研究输出的评定量表共有%个项目，分能力、性格、动力三个维度，适用于我国团体开发模式下的普通软件开发人员。量表的信、效度等测量学指标均较为理想，可以初步用于测评软件开发人员的心理素质。更多还原

【Abstract】 Aim:Design a test questionnaire about psychological quality of software-developing personnel and computerize preliminarily this questionnaire in order to provide large enterprises in China with efficacious psychological testing means for selecting software-developing talents with fine latent quality.Methods:1. Scaling: Settle the construction of the scale, set its item treasury based on the work analysis of software-developing personnel, then sift and settle items by using the experience mode and the conventional item analyzing technology (such as differential validity technology) simultaneously while taking into account expert assessment2. Testing: Take samples, which totals 72, from certain famous software companies (such as Huawei Company) and, controlling the factors that may influence working ability (such as sex), make statistical and logical analysis of the test result3. Reliability testing:(Internal consistency reliability. Calculate the consis-tency correlation coefficient of the scale with the data of the test samples.(2)Split-half reliability. Calculate the split-half correla-tion coefficient of the scale with the data of the test samples.4. Validity testing:(1) Content validity. Analyze and study the content validity of the scale using internal consistency factors.(2) Construction validity. Illustrate the construction vali-dity of the scale using factor analysis and consistency analysis.(3) Validity-criterion correlation validity. Use as three 3 validity criterions the placement of 30 new personnel separately of the training score, of the assessment of the immediate superior when probation expires and there follows a formalemployment, of the score of formal-employment defense, and then calculate the correlation of these 3 with the placement of the test score. Result1. Designs a test scale about psychological quality of software-developing personnel which consists of 96 items and is divided into 3 dimensions as ability, temperament and impetus by making surveying and logical analysis of the test scale and result2. The result of item analysis shows that most of the index of item discrimination were above 0.2.3. The consistency coefficient of the scale is 0.8081 and the Split-half correlation coefficient 0.8362.4. The result effector analysis shows that the 3 dimensions of the scale load 1 factor relatively largely.5. The correlation coefficients of the total score of psychological quality with the placement of training score, of the assessment of the immediate superior when probation expires and there is a formal employment, of the score of formal-employment defense are respectively 0.406,0.467 and 0.367.Conclusion.The assessment scale this research invents, which consists of 96 items and is divided into 3 dimensions as ability, temperament and impetus, is applicable to the ordinary software-developing personnel in the team-work mode in China. The surveying indexes such as reliability and validity are relatively satisfactory thus this scale can be used to test psychological quality of software-developing personnel.更多还原