Comments in the writing course (11/17)

1. 寫出如何驗證practice effect之定義
2. ANOVA會受樣本數影響，ANOVA為顯著但練習效應之size小，要如何解釋?
3. 先寫出主要架構，然後再寫細節
4. Table的標題太簡略

Outline of the results
1. Descriptive interpretation
2. Practice effect
    (1) Cohen's d
    (2) Patterns of the practice effect
    (3) RCIp
3. Test-retest reliability

原先之寫作

Data analysis

    The practice effects across 4 sessions (i.e., Time 0-1, Time 1-2, Time 2-3) was examined using repeated measures ANOVA. The size of the practice effect was calculated by Cohen’s d. The Cohen’s criteria are: 0.20-0.49, small effect size; 0.50-0.79, moderate effect size; and ≧ 0.80, large effect size¹.

    A cumulative pattern was defined that the mean of the posttest was lower than the pretest and follow this pattern over several sessions, until reaching a plateau. The criteria of a plateau pattern were: (1) consecutive two times of the values of Cohen’s d were lower than previous value (e.g., Cohen’s d of Time 1-2 < Cohen’s d of Time 0-1 and Cohen’s d of Time 2-3 < Cohen’s d of Time 1-2); and (2) two paired t-tests were not significant, one time-point comparing to the two following time-points respectively (e.g., Time 1-Time 2 and Time 1-Time 3).

Results

    Repeated measures ANOVA across the 4 sessions in the FDT revealed significant practice effects on three indices, selective attention, alternating attention, and basic attention and processing speed (Table 2). Table 3 presents the results of Cohen’s d. The Cohen’s d of 5 indices of the FDT ranged from 0.03 to 0.26. The practice effects of 5 indices all showed cumulative patterns across the 4 sessions. Only one index, “basic attention and processing speed” had reached a plateau after the 2^nd assessment.

修改後之寫作

Data analysis

    Effect sizes (Cohen’s d) were calculated for estimating the magnitudes of the practice effects across three comparisons at 4 time-points (i.e., Time 0-1, Time 1-2, and Time 2-3). The Cohen’s criteria are: 0.20-0.49, small effect size; 0.50-0.79, moderate effect size; and ≧ 0.80, large effect size³³.

    A cumulative pattern of the practice effect was revealed when the means of the posttests were lower than the means of the pretests in the 4 serial assessments. The decreasing means of the score (i.e., reduction in reaction time) indicate that the performance gets improvement (fast speed). The criteria of reaching a plateau pattern of the practice effect in order were: (1) two comparisons of Cohen’s d, consecutive two values of Cohen’s d were lower than its previous value (e.g., d of Time 1-2 < d of Time 0-1, and d of Time 2-3 < d of Time 1-2); and (2) two paired t-tests were not significant, one time-point comparing to the two following time-points respectively (e.g., Time 1 versus Time 2, and Time 1 versus Time 3).

Results

    Table 2 presents the magnitudes of the practice effects (Cohen’s d) of 5 indices of the FDT. The small practice effects were shown at Time 2-3 of the index of alternating attention (d = 0.23), Time 1-2 of the index of ability of inhibition (d = 0.26), and Time 2-3 of index of ability of flexibility (d = 0.26). Others’ Cohen’s d were lower than 0.16.

    The comparisons of means between pretest and posttest at the 5 indices in the 4 serial assessments were decreasing (Table 3). It indicated the practice effects of 5 indices all showed cumulative patterns. The two comparisons of the values of Cohen’s d were not lower than its previous value in the 5 indices. Thus, all 5 indices did not reach a plateau pattern (Table 2).