嬰幼兒綜合發展測驗之預測效度研究

Predictive Validity of Comprehensive Developmental Inventory for Infants and Toddlers (CDIIT)

王天苗(Tien-Miau Wang)

《特殊教育研究學刊》 29期 (2005/09) Pp. 1-24

https://doi.org/10.6172/BSE200509.2901001

摘要

本兩年期研究主要目的在探究「嬰幼兒綜合發展測驗」(CDIIT)的預測效度。除探討該篩選測驗結果對同時期所得診斷結果的預測效果及學前測得CDIIT的結果對日後入學表現和特殊教育需要之短期、中期和長期預測力之外，並比較該篩選測驗、「幼兒問題轉介表」和「學前兒童發展檢核表」(PCDC)三種篩選工具的預測力。研究樣本包括：(l)短期追蹤個案：包括第一年前測個案242名（就讀北市幼稚園大班，一般兒173名、遲緩兒69名）和第二年追蹤所得209名個案（一般兒l54名，遲緩兒55名）；(2)中期追蹤個案：從參加八十八學年度北市幼稚園優先入園鑑定的遲緩幼兒中，四年後追蹤得103名幼兒；(3)長期追蹤樣本：即八十四年至八十五年間CDIIT建立常模時的效度樣本，經約八年後追蹤，共得61名一般幼兒和59名遲緩幼兒。本研究工具包括CDIIT、PCDC、「幼兒問題轉介表」和「學生學習狀況調查表」四種。個案資料以相關分析、分類分析、X^2差異比較進行統計分析。
本研究主要的發現包括：(l)CDIIT篩選測驗結果能有效預測同時期由特教老師綜合研判的診斷結果。(2)「幼兒問題轉介表」和CDIIT（任一分測驗Z分數低於－l切截點）較PCDC更能有效篩選出問題個案，但錯誤接受率也較高。(3)學前測得之CDIIT結果對日後特教需要與在校表現的預測力：具短期預測力，也有長期預測效果，但相隔時間愈長，測驗結果與日後特教需要和在校問題間關係愈弱；無論由篩選或診斷測驗所得總分愈低或被判定有問題者，日後比較需要特殊教育，也有較多學業學習、人際關係或團體參與等問題，不過對特教需要和學業學習的預測力較高；以篩選測驗「任一分測驗」結果來預測的正確率要比以「總分」的預測效果大，不過有較高過度轉介率；預測無特教需要和無在學問題者的正確率要比預測有需要或有問題者要高。

關鍵字

發展測驗 ； 預測效度

並列摘要

The main purpose of this two-year study was to examine the predictive validity of the Comprehensive Developmental Inventory for Infants and Toddlers (CDIIT), the well-designed standardized development test in Taiwan. It was intended to validate CDIIT for predicting diagnostic results and later school problems or special education needs, based on the data from 242 pre-test preschool children (NH=173, DD=69), 209 1-year follow-up cases (NH=154, DD=55), 103 4-year follow-up DD cases, and 120 8-year follow-up cases (NH=61, DD=59). Comparisons were made on the predictive power of CDIIT, teacher's referral ratings on Child Problems Referral Survey (CPRS), and Preschool Children Development Checklist (PCDC). Correlation analyses, classification analyses, and X^2 group comparisons were also conducted.
The two-year study has found that CDIIT has fairly good predictive validity for diagnostic results and later school performances or special education needs. The results from CPRS and CDIIT Screening Test with Z scores below cut-off point of-1.0 in any single subtest has the hightest sensitivity and lowest false negative rate than PCDC and CDIIT with Z score derived from Total Score to predict cases having later school prblems or speical educational needs, however, higher overreferral rate was noted. The evidence, therefore, has given strong supports to suggest that the result of CDIIT judged from any single subtest rather than Total Score can be used confidently in mass screening for developmental delays, however, CPRS, designed based on classroom situations, might be more appropirate to be used by preschool teachers.

並列關鍵字

developmental test ； predictive validity

參考文獻

(1997)。身心障礙保護法。

(1993)。兒童福利法。

(1997)。特殊教育法。

Barnes, K. E.(1982).Preschool screening: The measurement and prediction of children at-risk.

Buck, A.,Gart, J.(1966).Comparison of a screening test and a reference test in epidemiological studies.American Journal of Epidemology.83,586-592.