Table 4.
Characteristics of covariates by level of catecholamine ( CAT ) in observed data and weighted data. Evans County study of 609 men
| Exposed ( CAT = 1) | Unexposed ( CAT = 0) | Total | |
|---|---|---|---|
| Observed data | ( n = 122) | ( n = 487) | ( n = 609) |
| Age in years (mean) | 61 | 52 | 54 |
| Cholesterol in mg/dl (mean) | 199 | 215 | 212 |
| History of ever smoking | 63% | 64% | 64% |
| Coronary heart disease | 22.1% | 9.0% | 11.7% |
| Weighted data | (∑ weightsa = 111) | (∑ weights = 487) | (∑ weights = 598) |
| Age in years (mean) | 55 | 54 | 54 |
| Cholesterol in mg/dl (mean) | 206 | 212 | 211 |
| History of ever smoking | 58% | 64% | 63% |
| Coronary heart disease | 27.3% | 10.8% | 13.8% |
a Sum of weights. Weights were calculated from a logistic regression model for the dependent variable, CAT ; weights were stabilized by a numerator equal to the marginal probability of ( CAT = 1) in the study population. The mean weight was slightly less than 1 (0.98), therefore the total sample size for the weighted analysis is slightly smaller than for the unweighted analysis.