Skip to main content
. Author manuscript; available in PMC: 2011 Aug 24.
Published in final edited form as: Demogr Res. 2005 Oct 17;13(18):455–484. doi: 10.4054/DemRes.2005.13.18

Table 4.

Coefficients for OLS Regression Models of the Logged Total Mortality Ratio, by Age, High Income Nations 1975–2000 (N = 105, 21 nations, 5 years)

Coefficients 0–34 35–69 70+ All Ages
Fixed Effectsa Fixed Effectsa Fixed Effectsa Fixed Effectsa
No Yes No Yes No Yes No Yes
Predicted Logged Non-Smoking Mortality Ratio
b 1.000** 1.000** .383 .368** 4.128** .525** .802** .697**
se .213 .099 .244 .064 .721 .099 .203 .074
beta .420 .707 .132 .311 .530 .520 .374 .580
Predicted Logged Smoking Mortality Ratio
b -- -- .084** .097** .011 .024** .087** .100**
se -- -- .011 .007 .006 .005 .019 .012
beta -- -- .702 1.020 .190 .625 .474 .711
Lagged Cigarette Consumption
b -- -- −.004 .018* .016** .009 .040 .059**
se -- -- .011 .007 .006 .006 .021 .011
beta -- -- −.031 .162 .272 .162 .465 .692
Predicted Logged Smoking Mortality Ratio × Lagged Cigarette Consumption
b -- -- .030** .036** .023** .016** .043** .050**
se -- -- .009 .005 .005 .003 .016 .009
beta -- -- .261 .436 .410 .536 .580 .593
Const .000 .000 .495 .480 −.213 .217 .314 .356
R2 .176 .500 .444 .720 .324 .285 .291 .630
Predicted Proportional Change from 2000 to 2020
Δb .176 .176 .047 −.006 −.329 −.083 .137 .081
Δc .088 .088 .015 −.036 −.169 −.065 .071 .024
Δd .000 .000 −.017 −.066 −.009 −.046 .006 −.032
Δe −.088 −.088 −.048 −.096 .152 −.028 −.060 −.089
*

p < .05

**

p < .01

a

Based on predicted values from OLS regression of logged smoking and non-smoking ratios from models without and with fixed effects in Table 4

b

Based on assumed linear rate of increase (or decrease at ages 70+) in the logged ratio of nonsmoking mortality;

c

based on assumed falling rate of increase (or decrease at ages 70+);

d

based on assumed no change;

e

based on assumed falling rate of decrease (or increase at ages 70+)