從自陳性量表結果的正確性來看心理測驗的效益

Enhancement of the Psychological Test Effects: A Study of How Subject Factors Influence the Correct Results of Self-Report Inventory

在矯正機關實施的心理測驗多數皆為自陳性量表，十分容易受到受測者因素而影響其作答反應類型。筆者根據實務經驗將受測者因素概分為「無效作答」、「漏答誤答」、「詐病偽病」、「社會期許性反應」、「真實反應」等五種作答反應類型，同時分析1,015名有效樣本在間隔6個月、5項心理測驗的兩次測驗結果後，有下列五點發現：一、在矯正機關實施自陳性的心理測驗時，受測者較多傾向於「真實反應」和「無效作答」這兩種作答反應類型，而「社會期許性反應」則維持一定的比率，此三種反應類型總和已佔96%以上的受測者。二、過去我們主觀認為發生率可能偏高的偽病詐病則相當少見，低於0.25%。三、僅有588位受測者（佔57.93%）在兩次測驗之間的作答反應型態不變，換句話說就是約有42%的受測者其作答反應型態是會有變化的；其中在「真實反應」與「無效作答」之間相互交替變化者最多。四、確認受測者反應型態，選擇真實反應的前測或後測的結果來收集資料，便可立即提升受測者正確回應的資料比例至近六成（586人，佔57.73%）。五、將測驗結果為真實反應的樣本資料，與所有原始樣本的測驗結果，來進行平均數考驗後，在本研究中的5項心理測驗皆達顯著差異；另外與全國資料來做統計考驗時，亦有3項測驗達顯著差異。最後提供4點拙見，做為各矯正機關改善當前心理測驗結果與提升應用效益的建議。 Most psychological tests use the self-report inventory, but it is very easy to be affected by the subject factors. We analyzed 1,015 valid samples which went through five before-after psychological tests at an interval of six months. Five response types are found, including “ineffective answer,” “missing answer and error answer,” “malingering and factitious disease,” “social desirability answer,” and “real answer.” The findings are as follows: 1. More than 96 percent of the tested samples responded with these three types, including “ineffective answer” , “social desirability answer” and “real answer”. 2. Less than 0.25 percent, relatively low, of the tested samples responded with the type “malingering and fictitious disease.” 3. Only 588 samples, which account for 57.93 percent of all subjects, don't change their response type between the before-after tests, which in turn means 427 samples (nearly 42 percent) give a different response type. Mostly, the change occurs with these two types of “ineffective answer” and “real answer”. 4. To enhance the percentage of accurate response to nearly 60 percent, it is suggested that the real answer type be selected for either before or after test. 5. The test results of the real response type and all samples, to t-Test, have a significant difference in all five psychological tests. Meanwhile, if we use the research data which accounts for 5,555 samples in Taiwan, to t-Test with this test result of the real response type, there is also a significant difference in three psychological tests. In conclusion, four suggestions are provided for the correct agencies concerned to improve the current psychological tests and enhance the effectiveness of future applications.