Test-retest Agreement, Random Measurement Error and Practice Effect of the Attention Scale for Elementary School Children

[[abstract]]背景與目的：「國小兒童注意力量表」包含五個注意力分量表，分別評估集中性、持續性、選擇性、交替性與分配性注意力，並以全量表分數判斷兒童的整體注意力表現。本研究之目的為檢驗此量表之再測一致性、隨機測量誤差及練習效應，以提昇其臨床應用價值。方法：104位國小兒童接受「國小兒童注意力量表」之施測，間隔四週再接受重測。結果：全量表之組內相關係數值最高（ICC=0.90），且最小可偵測之變化值之百分比低於判定標準（MDC%=15.7%）。全量表之效應值d 亦極微小（≤0.15）。分量表部分，僅有持續性注意力分量表之MDC%（26.2%）低於判定標準，而集中性、選擇性、交替性和分配性注意力分量表之MDC％（≥33.4%）皆高於判定標準。另外，應用於二年級兒童時，集中性、持續性和分配性注意力分量表之效應值較高（≥0.20）。結論：「國小兒童注意力量表」全量表分數之再測一致性最高、隨機測量誤差最低，且間隔四週之練習效應極微小，顯示此分數最適用於臨床或研究。若臨床人員須得知兒童於五個注意力向度的表現，建議將分量表下的各分測驗分別施測二至三次，再取其平均值作為分量表分數，以降低隨機測量誤差。針對二年級以下的兒童，可在各分測驗之正式測驗前增加練習時間，以降低練習效應之影響。 Background and Purpose: The Attention Scale for Elementary School Children (ASESC) includes five subscales, assessing focused, sustained, selective, alternating and divided attention, respectively. The whole scale score was used to indicate children's overall attention performance. The purpose of this study was to investigate the test-retest agreement, random measurement error and practice effect of the ASESC to improve its clinical utility. Method: One hundred and four elementary school children were assessed using the ASESC twice, 4 weeks apart. Results: The value of intraclass correlation coefficient for the whole scale was the highest (0.90), and its percentage of minimal detectable change (MDC%) was lower than the criterion (15.7%). The effect size d of the whole scale was also very small (≤ 0.15). The MDC% (26.2%) of the sustained attention subscale was lower than the criterion. However, the MDC% (≥ 33.4%) of the focused, selective, alternating and divided attention subscales were all higher than the criterion. The effect size of the focused, sustained and divided attention scale were larger (≥ 0.20) in second grade children. Conclusion: The ASESC whole scale score had highest test-retest agreement, lowest random measurement error and negligible practice effect, which is most suitable for clinical and research use. If clinicians tend to understand children's specific attentional performance, they should administer the subtests of each subscale 2-3 times to obtain an average score to reduce random measurement error. For second grade children or under, more practice before formal test is helpful to reduce the practice effect