Table 3.
Regression models for cut-off corrections of NADL global scores and scores on subtests.
| Intercept (SE) | Age (SE) | Education (SE) | Gender (SE) | R2 | |
|---|---|---|---|---|---|
| Interview with patients | 10.57 (0.21) | −0.02 (0.004) | – | – | – |
| Interview with caregiver | – | – | – | – | – |
| Informal total | 23.38 (0.49) | −0.05 (0.008) | – | – | 0.18 |
| Formal total | 70.78 (2.45) | −0.10 (0.03) | 0.39 (0.11) | – | 0.26 |
| Numerosity comparison | – | – | – | – | – |
| Number line | |||||
| Digit comprehension | – | – | – | – | – |
| Reading number aloud | 4.08 (0.13) | – | 0.04 (0.01) | 0.20 (0.09) | 0.11 |
| Writing numbers on dictation | – | – | – | – | – |
| Mental addition | – | – | – | – | – |
| Mental subtraction | – | – | – | – | – |
| Mental multiplication | 4.45 (0.22) | – | 0.07 (0.02) | – | 0.09 |
| Written rules | 6.27 (0.38) | −0.02 (0.004) | 0.07 (0.02) | – | 0.30 |
| Written addition | 3.42 (0.44) | −0.02 (0.005) | 0.07 (0.02) | – | 0.25 |
| Written multiplication | 4.66 (0.21) | −0.02 (0.004) | – | – | 0.21 |
| Written operations—addition | – | – | – | – | – |
| Written operations—subtraction | 6.49 (0.27) | −0.02 (0.004) | – | – | 0.13 |
| Written operations—multiplication | – | – | – | – | – |
| Total number comprehension | – | – | – | – | – |
| Total reading and writing Arabic numerals | 9.54 (0.17) | – | 0.06 (0.01) | 0.25 (0.12) | 0.14 |
| Total mental calculation | 15.96 (0.35) | – | 0.09 (0.03) | – | 0.05 |
| Total rules and principles | 13.74 (0.85) | −0.05 (0.01) | 0.17 (0.04) | – | 0.40 |
| Total written operations | 17.02 (0.54) | −0.04 (0.01) | – | – | 0.11 |
Each row in the table reports the coefficients of one linear regression model. The first column reports the dependent variable, the following four columns report the parameter for the models (the value within brackets indicates the standard error for the parameter). The last column reports the adjusted R-squared. A missing value in the table indicates that the parameter did not contribute significantly to the model fit and then was removed in the modeling procedure. All coefficients reported (with the exception of Intercept, which is always included in a meaningful model) were selected by mean of backward elimination of non-significant variables. All variables reported significantly improved the fit of the model and their associated t-values had p < 0.05.