Table 6.
Differences in congruence coefficients with population components between GPR-Varimax and SPSS-Varimax for double-optimum simple structure.
| 25 iterations |
250 iterations |
|||||
|---|---|---|---|---|---|---|
| GPR start | No | With | No | With | ||
| k | N | loadings | Kaiser | Kaiser | Kaiser | Kaiser |
| 3 | 100 | Unrotated | — | 0.001 | — | 0.001 |
| Random 1 | — | 0.017 | — | 0.017 | ||
| Random 10 | — | 0.017 | — | 0.017 | ||
| 300 | Unrotated | 0.001 | 0.002 | — | — | |
| Random 1 | 0.001 | 0.027 | — | 0.029 | ||
| Random 10 | — | 0.036 | — | 0.030 | ||
| 6 | 100 | Unrotated | — | — | — | — |
| Random 1 | 0.001 | — | — | — | ||
| Random 10 | 0.001 | — | — | — | ||
| 300 | Unrotated | — | — | — | — | |
| Random 1 | — | 0.016 | — | 0.030 | ||
| Random 10 | — | 0.023 | — | 0.032 | ||
| 9 | 100 | Unrotated | -0.001 | -0.001 | — | — |
| Random 1 | -0.001 | -0.001 | — | — | ||
| Random 10 | -0.001 | -0.001 | — | — | ||
| 300 | Unrotated | -0.001 | — | -0.001 | — | |
| Random 1 | -0.001 | 0.027 | -0.001 | 0.026 | ||
| Random 10 | — | 0.034 | — | 0.028 | ||
| 12 | 100 | Unrotated | -0.001 | -0.001 | 0.001 | — |
| Random 1 | -0.001 | -0.001 | 0.001 | — | ||
| Random 10 | -0.001 | -0.001 | 0.001 | — | ||
| 300 | Unrotated | — | — | — | — | |
| Random 1 | — | 0.026 | — | 0.027 | ||
| Random 10 | — | 0.031 | — | 0.029 | ||
GPR, gradient-projection based rotation. k, number of components. With Kaiser, loadings were Kaiser normalized before rotation. No Kaiser, non-normalized loadings. Differences refer to the mean congruence coefficients with population components of 1,000 samples and were computed as GPR-mean – SPSS-mean. Differences of |Δ|≥ 0.001 are reported, otherwise results are considered equal (“—”). A positive sign means that GPR-Varimax was better than SPSS-Varimax and vice versa.