試題項目反應理論與李克特量表法在壓力知覺測量上之差異比較

Abstract

本研究主要目的有：一、比較試題項目反應理論與李克特量尺法在壓力知覺測量計分上的情形。二、比較試題項目反應理論與李克特量尺法在壓力知覺測量信度上的情形。三、探討試題項目反應理論與李克特量尺法，對於國中學生於不同個人屬性（性別、年級、出生序、家庭狀況）在壓力知覺測量結果上的差異情形。以臺北市、臺北縣七所公立國民中學七至九年級893位國中學生為研究對象，以問卷調查法進行資料收集，所使用的研究工具為「生活型態評價問卷的第二部分」。調查所得資料，以描述性統計、皮爾森積差相關、單因子變異數分析等統計方法處理。綜合研究結果，本研究的主要發現如下：一、李克特量尺法和試題項目反應理論等級反應模式在計分上呈現高相關。而試題項目反應理論等級反應模式的計分方式，於極端敏感性較高，當變異量愈大時，欲能顯現其在測量上的優越性、延展性。二、國中學生在壓力知覺測量（生活型態評價問卷的第二部分）的結果顯示所知覺的壓力是偏低的。三、李克特量尺法的平均數和試題項目反應理論等級反應模式的閾值，呈現出相當一致性的變化，在李克特量尺法得分較低的題目中，閾值有偏右的趨向，反之平均數較高的題目，則有偏左的趨勢。四、不管是以試題項目反應理論或是李克特量尺法的計分結果均顯示，不同性別、年級、出生序、家庭狀況的國中學生在壓力知覺測量上，均無顯著差異。最後，研究者歸納研究結論，提出建議以作為在心理測量與輔導工作實務及未來研究的參考。

The main purpose of this research was to 1) compare the scoring of perceived stress measurement between item response theory and Likert scaling, 2) compare the reliabilityof perceived stress measurement between item response theory and Likert scaling, 3) investigate on how junior high school students with different personal properties ( sex, grade, birth order, and family structure ) differ in the perceived stress measurement between item response theory and Likert scaling. The sample of this study was consisted of 893 7th grade to 9th grade students from seven public junior high schools in Taipei city and Taipei county. The data was collected through survey, using “Lifestyle Appraisal Questionnaire-Part II” as instrument. The data obtained was further analyzed using descriptive statistics, K. Pearson's product-moment correlation, and one-way ANOVA.Main findings of this research were as follows: 1) A high level of correlation was found between the scoring of Likert scaling and item response theory’s graded response model. Moreover at the extremes where the variance is higher, item response theory’s graded response model performs an even more sensitive and excellent task in measurement. 2) The measured result revealed that high school students’ perceived stress (Lifestyle Appraisal Questionnaire-Part II) was relatively lower than average. 3) Consistent change was found between the average of Likert scaling and the the threshold points of item response theory’s graded response model. It is also found that survey questions that had lower scores in Likert scaling tended to have a larger threshold, while those that had higher scores tended to have a smaller threshold. 4) Both Likert scaling and item response theory’s measurement on perceived stress showed no significant difference among students of different sex, grade, birth order, and family structure. As a conclusion, the researcher of this study made several implications on psychometric assessment, counseling practice, and for future research.