Model with all explanatory variables
![]() |
−18.62 − 12.74 wt + 0.059 age − 0.037 ht − 0.441 neck − 0.017 chest + 0.886 abdomen − 0.180 hip + 0.244 thigh − 0.014 knee + 0.168 ankle + 0.154 biceps + 0.434 forearm − 1.508 wrist |
Models with one explanatory variable
Use this diagram to explain the difference between the coefficient of X when it is the only variable in the model (marginal relationship) and when there are also other variables in the model (conditional relationship).
The coefficient of wt is negative in the full model, suggesting that heavier men tend to have lower body fat. However the coefficient is positive when wt is the only variable in the model.
Other variables have different signs than expected in the full model.
Percent body fat and several other measurements of body dimensions from a group of 252 men.