AB+-+Wright+(2006)

The only procedure that is always correct in this situation is a scatterplot comparing the scores at time 2 with those at time 1 for the different groups. In most cases you should analyse (sic) the data in several ways. If the approaches give different results. . . think more carefully about the model implied by each. ( Wright, 2003, p. 130 ) The paradox occurs because different assumptions are made and different questions are asked corresponding to each test.  t test: posti = prei + β1groupi  + β0 + ei  ANCOVA: posti = β2prei + β1groupi + β0 + ei The t test “ asks whether the average gain in score is different for the two groups ” ( p. 666 ), while ANCOVA “ asks whether the average gain, partialling out pre-scores, is different between the two groups ” ( p. 666 ). If the assumption is that without treatment measures will remain constant or if the assignment tests and is based on ability then the t test is the best tool, but if the assumption is that even without treatment measures are changing, or if assignment is from the pretest then the ANCOVA is the best tool. ||
 * **Reference: ** || <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">Wright, D. B. (2006). Comparing groups in a before-after design: When t test and ANCOVA produce different results. //British Journal of Educational Psychology, 76//, 663-675. doi:10.1348/000709905X52210 ||
 * **<span style="font-family: 'Times New Roman',Times,serif; font-size: 130%;">Author's: ** || <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">Wright, D. B. ||
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 130%;">**Title:** || <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">Comparing groups in a before-after design: When t test and ANCOVA produce different results. ||
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 130%;">**Year:** || <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">2006 ||
 * **<span style="font-family: 'Times New Roman',Times,serif; font-size: 130%;">Journal: ** || //<span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">British Journal of Educational Psychology // ||
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 130%;">**Retrieval Information**: || <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">http://dx.doi.org/10.1348/000709905X52210 ||
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 130%;">**Bibliography**: ||  ||
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">This article focused on nonrandomly select groups assumed to be non-equivalent. Identified in this article are the “ <span style="color: #0000ff; font-family: 'Times New Roman',Times,serif; font-size: 120%;">two most common statistical approaches <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">” ( <span style="font-family: 'Times New Roman',Times,serif; font-size: 90%;">p. 663 <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">) used in this situation; t test on the gain scores and an analysis of covariance (ANCOVA). The author states that
 * **<span style="font-family: 'Times New Roman',Times,serif; font-size: 130%;">Journal: ** || //<span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">British Journal of Educational Psychology // ||
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 130%;">**Retrieval Information**: || <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">http://dx.doi.org/10.1348/000709905X52210 ||
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 130%;">**Bibliography**: ||  ||
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">This article focused on nonrandomly select groups assumed to be non-equivalent. Identified in this article are the “ <span style="color: #0000ff; font-family: 'Times New Roman',Times,serif; font-size: 120%;">two most common statistical approaches <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">” ( <span style="font-family: 'Times New Roman',Times,serif; font-size: 90%;">p. 663 <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">) used in this situation; t test on the gain scores and an analysis of covariance (ANCOVA). The author states that
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 130%;">**Bibliography**: ||  ||
 * <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">This article focused on nonrandomly select groups assumed to be non-equivalent. Identified in this article are the “ <span style="color: #0000ff; font-family: 'Times New Roman',Times,serif; font-size: 120%;">two most common statistical approaches <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">” ( <span style="font-family: 'Times New Roman',Times,serif; font-size: 90%;">p. 663 <span style="font-family: 'Times New Roman',Times,serif; font-size: 120%;">) used in this situation; t test on the gain scores and an analysis of covariance (ANCOVA). The author states that
 * **<span style="font-family: 'Times New Roman',Times,serif; font-size: 130%;">Additional References: ** ||  ||
 * N/A ||
 * N/A ||