Table 3.

Hypothetical example of data with five imputed datasets

Imputed DatasetIDAge (yr)WomanBMI (kg/m2)PPRA (%)StrokeYears Followed
15–18.525–3030–4511–8080–100
1139No01001No8.4
2139No01001No8.4
3139No01001No8.4
4139No01001No8.4
5139No01001No8.4
1244Yes0.35a0.34a−0.21a01Yes10.9
2244Yes0.12a0.45a0.03a01Yes10.9
3244Yes0.21a0.27a−0.47a01Yes10.9
4244Yes−0.01a0.97a−0.44a01Yes10.9
5244Yes0.38a0.80a0.64a01Yes10.9
1467No000−0.21a0.08aNo11.6
2467No0000.04a−0.33aNo11.6
3467No0000.25a0.21aNo11.6
4467No0000.31a−0.04aNo11.6
5467No0000.69a0.07aNo11.6
  • BMI=18.5–25 (normal) and PPRA=0–10 (normal) are used as reference groups and represented by a zero in all dummy variables pertaining to each variable. ID, identification; PPRA, panel reactive antibody.

  • a Multiple imputation may lead to data that are not consistent with the original format; in this case, values imputed for missing observations of categorical (binary) data are continuous. Furthermore, although original categories of a variable may be mutually exclusive, imputed data may not be mutually exclusive, which is appropriate, because the imputed values, per se, do not have any meaning.