Extended recruitment efforts minimize attrition but not necessarily bias

Häring, R.; Alte, Dietrich; Völzke, Henry; Sauer, Sybille; Wallaschofski, Henri; John, Ulrich; Schmidt, Carsten Oliver

doi:10.1016/j.jclinepi.2008.06.010

Cited by 80 publications

(89 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Despite the higher response rates, however, almost none of the estimates from the survey changed significantly. Similar outcomes have been observed numerous times, for example by Curtin et al (2005), Haring et al (2009), andIngen et al (2009). While there are several possible explanations for the lack of an effect on the estimates, these examples point out gaps in our understanding of the effects of data collection efforts on biases.…”

Section: Brick: Unit Nonresponse and Weighting Adjustmentssupporting

confidence: 74%

Unit Nonresponse and Weighting Adjustments: A Critical Review

Brick¹

2013

Journal of Official Statistics

121

View full text Add to dashboard Cite

This article reviews unit nonresponse in cross-sectional household surveys, the consequences of the nonresponse on the bias of the estimates, and methods of adjusting for it. We describe the development of models for nonresponse bias and their utility, with particular emphasis on the role of response propensity modeling and its assumptions. The article explores the close connection between data collection protocols, estimation strategies, and the resulting nonresponse bias in the estimates. We conclude with some comments on the current state of the art and the need for future developments that expand our understanding of the response phenomenon.

show abstract

Section: Brick: Unit Nonresponse and Weighting Adjustmentssupporting

confidence: 74%

Unit Nonresponse and Weighting Adjustments: A Critical Review

Brick¹

2013

Journal of Official Statistics

121

View full text Add to dashboard Cite

show abstract

“…Details on the SHIP study design and data have been published. 19,20 In brief, from the 212,157 inhabitants living in the area, a representative sample of 7008 subjects aged between 20 and 79 years was selected using population registries where all German inhabitants are registered. Only individuals with German citizenship and main residency in the study area were included.…”

Section: Methodsmentioning

confidence: 99%

Ultrasonographic hepatic steatosis increases prediction of mortality risk from elevated serum gamma-glutamyl transpeptidase levels # †

et al. 2009

Self Cite

View full text Add to dashboard Cite

The aim of the present study was to investigate the association of serum gamma-glutamyltransferase (GGT) levels with all-cause mortality and to assess the impact of ultrasonographic findings of hepatic hyperechogenicity in that association. We used data from 4,160 subjects (2,044 men and 2,116 women) recruited for the population-based Study of Health in Pomerania (SHIP) without baseline hepatitis B and C infections or liver cirrhosis. GGT was divided into age-and sex-dependent quintiles to calculate overall and sex-specific crude incidence mortality rates. Hepatic steatosis was defined by elevated GGT levels (>80%) and the presence of hyperechogenic liver ultrasound. We used multiple-adjusted Cox proportional hazards regression models, first, to assess the direct effect of GGT on all-cause mortality, second, to stratify according to the ultrasonographic finding, and third, to investigate potential mediating effects of cardiometabolic risk factors. During 29,810 person-years (7.3 years, median) of follow-up, 307 individuals (7.5%) died, resulting in a death rate of 0.86 deaths per 1000 person-years. Elevated GGT levels were associated with increased risk of mortality in men (hazard ratio [HR] 1.49; 95% confidence interval [CI], 1.08-2.05), but not in women (HR 1.30; 95% CI, 0.80-2.12). This association was even stronger in men with hepatic steatosis (HR 1.98; 95% CI, 1.21-3.27). Cause-specific mortality analysis by cardiovascular disease deaths confirmed the sex-specific association. Adjustment for cardiometabolic risk factors did not affect the estimates. Conclusion: In the case of increased GGT levels, liver ultrasound should be performed, not only for diagnosis, but also for further risk stratification.

show abstract

“…Survey data may be biased because people who complete surveys tend to be significantly different to those who do not, often on variables central to the study questions [e.g., 8,9]. One method of dealing with this bias is to assume that late responders are similar to nonresponders and make adjustments based on the attitudes or characteristics of late participants [10]. This method is consistent with the continuum of resistance model [11,12], which suggests that people who respond in the first contact lie at one end of a continuum and non-respondents at the other end.…”

Section: Introductionmentioning

confidence: 99%

“…Late participants are seen to be "more like" non participants in that they would never have participated had they not received a reminder [13]. However, a number of studies have thrown doubt on the continuum of resistance model, suggesting that non-participants are not more similar to late participants than they are to early participants [10,12,14,13]. For example, Haring et al [10] found that late respondents were generally younger, single, better educated, and employed than early respondents but found no support for the hypothesis that late respondents were similar to non-respondents.…”

Section: Introductionmentioning

confidence: 99%

“…However, a number of studies have thrown doubt on the continuum of resistance model, suggesting that non-participants are not more similar to late participants than they are to early participants [10,12,14,13]. For example, Haring et al [10] found that late respondents were generally younger, single, better educated, and employed than early respondents but found no support for the hypothesis that late respondents were similar to non-respondents. This result is consistent with the notion that rather than constituting a single continuum where people differ on overall readiness, speed of return of kit may reflect subgroup differences among those contemplating action (c.f., the transtheoretical model, [15]) with these differences potentially reflecting different scores on the social cognitive and other variables related to screening participation.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation