Non-response bias and hazardous alcohol use in relation to previous alcohol-related hospitalization: comparing survey responses with population data

Table 1 Percentages of people with alcohol-related hospitalization in the surveys, and in the corresponding population data, and the resulting estimates of non-responders in the surveys

Material	Percent	95% CI¹	*Hospitalized n*	*Total N*
2006 cross-sectional survey	1.13²	0.99, 1.22	362	32 626
2006 population data	1.64		22 298	1 358 176
2006 non-responders	2.48³		511	20 597
2002-2007 longitudinal survey	0.92⁴	0.78, 1.02	210	23 369
2007 population data	1.68		22 841	1 361 698
2007 non-responders	2.45³		580	23 651
Both surveys	1.04⁵	0.95, 1.14	572	55 995
Population data	1.66		22570	1 359 937
Non-responders	2.47³		1119	45 262

¹Confidence intervals are given for the survey responders, since it is their estimates that are being compared with the corresponding population parameters.
² Since the surveys were stratified by gender, municipality and district, this estimate was weighted for sampling probability. The corresponding un-weighted estimate was 1.11 percent.
³ The estimate of the non-responders was based on un-weighted estimates of the responders. It was obtained by assuming that survey estimates correspond to population data.
⁴ The un-weighted estimate was 0.90 percent.
⁵ The un-weighted estimate was 1.02 percent.

ISSN: 1747-597X