Abstract

This study was conducted to examine a possible alternative method to resolve unstable validity of biographical data in small samples. In this study, 219 biographical items were derived from life history with incumbents about early and recent life experience respectedly. The participants of this study were tele-marketers in L company and the survey was repeatedly conducted for three times: 432 cases, 316 cases, 229 cases. Based upon the results of categorical data analysis (χ² test), 88 items among 219 biographical items was selected. Test-retest reliability test showed that the reliability coefficient was .70. But the significant validity shrinkage was observed in the cross-validation and retest data. To establish stable validity of biographical data, the combined weights with first and second data were recalculated. The results showed relatively stable validities in test, retest and cross-validation data.