There are some very surprising numbers here. I'll have to look at the study. As I read it, the entries in the table are 1 for a base rate, with lower numbers providing lower mortality and higher numbers higher. From the way Robin talked about it, the numbers are actually ratios of mortality rates, so 1 vs. 3 is a factor of three (with 3 being worse) in how often people die, and similarly .33 and 1 is a factor of three (with .33 being better). Presumably the numbers are not comparable across the tables.
I believe the numbers in all the other tables, but I didn't realize being male was such a handicap.
Chris
Robin told us:
>> Here are the death rate ratio parameters from their model:
>>
>> Age 25-34 35-44 45-54 55-64 65-74 75+
>> 1.0 2.66 3.46 9.30 16.78 40.00
>> Sex Male Female
>> 1.0 .41
>> Race NonBlack Black
>> 1.0 1.19
>> Residence Rural Suburban City
>> 1.0 1.16 1.52
>> Education 16+yrs 12-15 0 -11
>> 1.0 .95 .90
>> Income 30K$+ 10-29K$ <10K$
>> 1.0 2.14 2.77
>> Smoking Never current former
>> 1.0 1.26 1.28
>> Alcohol drinks/mo. Moderate None Heavy
>> 1.0 1.13 .85
>> Body Mass Normal Underweight Overweight
>> 1.0 2.03 .94
>> Physical Activity Quintiles
>> 5(high) 4 3 2 1(low)
>> 1.0 1.46 1.60 2.25 2.91