Skip to main content

Table 7 Construct validity: narration

From: Expressive language sampling as a source of outcome measures for treatment studies in fragile X syndrome: feasibility, practice effects, test-retest reliability, and construct validity

  1. Note that all values are bivariate zero-order correlations except for that between dysfluency and the SB5 VWM score, which is a partial correlation controlling for syntax (MLU). n = 85–89 across correlations. Uncorrected p values for individual tests are marked with asterisks as follows: ***signifies p ≤ .0005; **signifies p ≤ .005; **signifies p ≤ .050. Shaded cells contain values that were significant at p ≤ .050 after correcting for multiple tests through the FDR procedure. The boldfaced values represent convergent validity relationships; all other values represent discriminant validity relationships