Table 3.

Respondents' Characteristics Predicting Stigma, Multiple Regression

Dependent variable: stigma, mean value 14.1%
	Coefficient	p-Value	95% CI
Age	0.0009	0.475	−0.0017 to 0.0037
White, non-Hispanic	0.0152	0.659	−0.0526 to 0.0831
Female	0.0401	0.198	−0.0211 to .10143
Education
Less than high school	referent
High school diploma	−0.0389	0.349	−0.1206 to 0.0426
Some college, no degree	0.0706	0.123	−0.0190 to 0.1603
4 year degree +	−0.0450	0.434	−0.1579 to 0.0679
Federal poverty level (FPL)
Below 50%	referent
50%-75%	0.0470	0.392	−0.0606 to 0.1547
75%-100%	−0.0479	0.308	−0.1400 to 0.0442
100%-150%	−0.0387	0.383	−0.1258 to 0.0484
Above 150%	0.0034	0.941	−0.0882 to 0.0951
Depression diagnosis, prestudy	0.0555	0.091	−0.0087 to 0.1198
One or more chronic condition diagnosis, prestudy	−0.0228	0.528	−0.0938 to 0.0481
Obese	−0.0086	0.787	−0.0712 to 0.0540
Medicaid	0.0348	0.344	−0.0374 to 0.1070
Private insurance	−0.0107	0.816	−0.1017 to 0.0802

n = 557. R² = 0.0372. Except for stigma, data source is self-reported survey data collected from September 2009 to December 2010. Seventeen people did not supply complete information and were not included in the model. The stigma code was derived from qualitative interviews conducted from May to September 2011. We fit a linear regression model rather than logistic.

For race, non-white or Hispanic is the referent category. For gender, male is the referent category. Medicaid and private insurance status is at the time of the in-person survey, not at the time of the stigma experience, which could have occurred at any time during the study period, 2008 to 2011.

We asked if the individual was ever diagnosed with depression, diabetes, high blood pressure, congestive heart failure, high cholesterol, or heart attack and, if so, followed up to establish the date of the first diagnosis. Prestudy diagnoses were those made before the drawings for the Oregon Medicaid lottery.

Obesity is defined as a BMI of at least 30. BMI was collected during the in-person survey.