Survey of Non-Group Health Insurance Enrollees

Survey Methodology

The Kaiser Family Foundation (KFF) Survey of Non-Group Health Insurance Enrollees is the first in a series of surveys examining the views and experiences of people who purchase their own health insurance, either directly from an insurance company or through a state or federal Health Insurance Exchange created by the Affordable Care Act (ACA). The survey was designed and analyzed by researchers at KFF. Social Science Research Solutions (SSRS) collaborated with KFF researchers on sample design and weighting, and supervised the fieldwork.

The survey was conducted by telephone from April 3 through May 11, 2014 among a nationally representative random sample of 742 adults who purchase their own insurance. Computer-assisted telephone interviews conducted by landline (333) and cell phone (409, including 219 who had no landline telephone) were carried out in English and Spanish by SSRS. Respondents were considered eligible for the survey if they met the following criteria:

• Between the ages of 18-64
• Currently covered by health insurance that they purchase themselves
• Not covered by health insurance through an employer, COBRA, Medicare, Medicaid, or the U.S. military or VA
• If purchase insurance from a college or university, the insurance covers health services received both within and outside the university setting
• If a small business owner, the health insurance they purchase is only for themselves and/or their family, and does not cover non-related employees of their business
• If purchase from a trade association, respondent pays the entire premium themselves
• Respondent was able to answer a question about whether insurance was purchased directly from an insurance company, from a state or federal health insurance marketplace, or through a health insurance agent or broker (Q35 in questionnaire)

Because the study targeted a low-incidence population, the sample was designed to increase efficiency in reaching this group, and consisted of three parts: (1) respondents reached through RDD landline and cell phone dialing to previously uncalled (“fresh”) sample (N=179); (2) respondents reached by re-contacting those who indicated in a previous RDD survey that they either purchased their own insurance or were uninsured (N=186); (3) respondents reached as part of the SSRS Omnibus survey (N=377), a weekly, nationally representative RDD landline and cell phone survey. All RDD landline and cell phone samples were generated by Marketing Systems Group. The “fresh” sample was designed to oversample areas with a lower-income population in order to complete more interviews with respondents who were uninsured prior to the ACA open enrollment period (since being uninsured is negatively correlated with income).

A multi-stage weighting process was applied to ensure an accurate representation of the national population of non-group enrollees ages 18-64. The first stage of weighting involved corrections for sample design, including accounting for the low-income oversample used in the “fresh” sample, the likelihood of non-response for the re-contact sample, number of eligible household members for those reached via landline, and a correction to account for the fact that respondents with both a landline and cell phone have a higher probability of selection. In the second weighting stage, demographic adjustments were applied to account for systematic non-response along known population parameters. No reliable administrative data were available for creating demographic weighting parameters for this group, since the most recent Census figures could not account for the changing demographics of non-group insurance enrollees brought about by the ACA this year. Therefore, demographic benchmarks were derived by compiling a sample of all respondents ages 18-64 interviewed on the SSRS Omnibus survey during the field period (N=4,312) and weighting this sample to match the national 18-64 year-old population based on the 2013 U.S. Census Current Population Survey March Supplement parameters for age, gender, education, race/ethnicity, region, population density, marital status, and phone use. This sample was then filtered to include respondents qualifying for the current survey, and the weighted demographics of this group were used as post-stratification weighting parameters for the “fresh” sample (including gender, age, education, race/ethnicity, marital status, income, and population density). A final adjustment was made to the full sample to adjust for previous insurance status (estimated based on the combined “fresh” and omnibus samples), to address the possibility that the criteria used in selecting the prescreened sample could affect the estimates for previous insurance status.

Weighting adjustments had a minor impact on the overall demographic distribution of the sample, with the biggest adjustments being made based on age (this is common in all telephone surveys, as younger respondents are the most difficult to reach and convince to participate). Weighted and unweighted demographics of the final sample are shown in the table below.

		Unweighted % of total	Weighted % of total
Age	18-25	13%	20%
	26-34	13	17
	35-44	13	17
	45-54	21	19
	55-64	37	25
	Refused	2	2
Gender	Male	47	50
	Female	53	50
Education	Less than high school graduate	5	8
	High school graduate	31	32
	Some college	23	25
	Graduated college	24	20
	Graduate school or more	14	11
	Technical school/other	3	3
	Refused	1	1
Race/Ethnicity	White, non-Hispanic	69	63
	Black, non-Hispanic	13	13
	Hispanic	10	13
	Other/Mixed	6	10
	Refused	2	2
Self-reported health status	Excellent	24	24
	Very good	33	31
	Good	28	30
	Fair	11	11
	Poor	4	3
	Don’t know/refused	*	1
Plan Type	Exchange plans	46	48
	ACA-compliant, non-Exchange plans	17	16
	Compliant, unknown if Exchange	3	3
	Non-compliant plans	33	31
	Unknown if compliant	1	1

 

All statistical tests of significance account for the effect of weighting. The sample size and margin of sampling error (MOSE) for key subgroups are shown in the table below.