Table 7.
Comparison of linear (eq. (16)) and non-linear (eq. (9)) regression expressions for predicting body fat fraction from BMI and age.
| Subjects | ± Age | Linear | Non-linear Model I | ||||||
|---|---|---|---|---|---|---|---|---|---|
| a | b | c | MSR | BMI0 | f1 | c | MSR | ||
| Male Caucasians | No | -.166 | .0141 | ---- | 0.00404 | 17.20 | .624 | ---- | .00409 |
| Yes | -.218 | .0129 | .00207 | 0.00263 | 19.15 | .500 | .00194 | .00287 | |
| Male Caucasian +Hispanic+Black | No | -.145 | .0134 | ----- | .00380 | 16.71 | .594 | ---- | .00385 |
| Yes | -.206 | .0127 | .00182 | 0.00270 | 18.73 | .496 | .00172 | .00288 | |
| Male Asian | Yes | -.156 | .0126 | .00169 | 0.00201 | 15.72 | .438 | .00169 | .00212 |
| Male Puerto Rican | Yes | -.155 | .0119 | .00163 | 0.00189 | 17.84 | .536 | .00150 | .00188 |
| Female Caucasian | No | 0.0409 | .0113 | ----- | 0.00391 | 13.50 | .739 | ----- | .00314 |
| Yes | -.0240 | .0104 | .00186 | 0.00281 | 14.39 | .635 | .00151 | .00244 | |
| Female Caucasian +Hispanic+Black | No | 0.0494 | .0109 | ------ | .00351 | 13.50 | .728 | ----- | .00276 |
| Yes | -.0160 | .0104 | .00169 | 0.00260 | 14.37 | .642 | .00132 | .00222 | |
| Female Asian | Yes | -.0903 | .0153 | .00122 | 0.00137 | 12.38 | .573 | .00122 | .00140 |
| Female Puerto Rican | Yes | 0.0718 | .00919 | .000947 | .00159 | 12.82 | .639 | .000737 | .00142 |
The regression parameters (either a, b and c; or BMI0, f1 and c) and the mean square residual error (MSR) for the different ethnic groups are listed.