Contact Patterns for Contagious Diseases
Leonhard Held, Niel Hens, Philip O’Neill, Jacco Wallinga in Handbook of Infectious Disease Data Analysis, 2019
An alternative approach to deal with the identifiability problem is to use more data, of a different kind: we can use information from social behavior surveys to estimate the contact rates. This use should be done with great care, as information from behavior surveys can be subject to difficulties associated with nonresponse bias, insufficient number of participants, and untruthful answers by respondents. The use of surveys on sexual behavior to inform contact rates in models for sexually transmitted infections, such as infections with gonorrhea and HIV, dates back to the 1990s [5, 6]. The development of social contact behavior surveys to inform contact rates for models of close contact infections or airborne droplet infections, such as measles and influenza, is more recent [1, 7, 8].
Weighting and Complex Sampling Design Adjustments in Longitudinal Studies
Jason T. Newsom, Richard N. Jones, Scott M. Hofer in Longitudinal Data Analysis, 2013
Because the response rate is less than 100% in most social surveys, some type of nonresponse adjustment is almost always used (Little, 1986). Nonresponse in this case is related to the inability to contact individuals or refusal to participate. Some surveys (e.g., Current Population Survey) adjust for nonresponse by dividing the sampling weight by the response rate. Others (e.g., HRS) adjust for nonresponse by dividing the sample weight by the predicted probability of responding to the survey. In the latter example, post-stratification and nonresponse adjustments are combined in a single operation.
Epidemiology and its uses
Liam J. Donaldson, Paul D. Rutter in Donaldsons' Essential Public Health, 2017
Detailed sampling rules govern which members of the household (particularly children) are to be included. Explanatory leaflets describe the purpose of the survey and help to gain compliance. Quality control measures include training of interviewers and nurses, checking of interview and measurement quality, and protocols for interviewing and measuring children. There are rules to govern what to tell people if abnormalities are found and what action needs to be taken. Some information is gathered on nonresponders and reasons for nonresponse.
Records request response rate and vaccination status of first-time college students at a mid-sized Midwestern university
Published in Journal of American College Health, 2022
Alexandra Larsen, Anders Cedergren
Approximately two-thirds of first-time college students in the fall 2018 cohort responded to the vaccination records request from the SHC. While a 66% response rate may be considered acceptable in some types of research, for a process required by state law,35 the response rate should be much higher. When individuals fail to respond to an information request, the impact of nonresponse error needs to be considered. There may be important unknown differences between the groups of non-respondents and respondents which prevent the generalization of findings to the entire study population. This may negatively impact how a university can effectively deal with infectious disease risks on a population level. Actions can be taken to improve response rates, including issuing reminders and providing incentives.36
An Empirical Investigation of Factors Contributing to Item Nonresponse in Self-Reported Bullying Instruments
Published in Journal of School Violence, 2020
Okan Bulut, Jiaying Xiao, Michael C. Rodriguez, Guher Gorgun
Despite their efficiency and ease of use in gathering data, self-reported instruments may fail to provide valid and meaningful results, especially in the presence of missing data (Brick & Williams, 2013; Johnson & Wislar, 2012). Missing data can occur for a variety of reasons in self-reported instruments. For example, some students may refuse to participate in a survey administered in their school due to a lack of interest in the survey topic, demographic characteristics (e.g., age and sex), and personality type (Porter & Whitcomb, 2005). This leads to nonresponse error at the student level (i.e., unit nonresponse) when students who do not respond are different from those who do respond in a way that limits the generalizability of survey findings to the target population (e.g., Standish & Umbach, 2019). Also, students may agree to participate in the survey but refuse to respond to some items on the survey, resulting in item nonresponse (Borgers & Hox, 2001; Brick & Kalton, 1996; Kadengye et al., 2012). Item nonresponse could lead to bias when some students are more likely to skip the questions than others. For example, despite using an anonymous survey, some students prefer to skip questions that ask them to report on the type and frequency of their bullying behaviors (Xiao et al., 2019).
Relationship Between CT Head Findings and Long-term Recovery in Children with Complicated Mild Traumatic Brain Injury
Published in Brain Injury, 2022
Colby Hansen, Laura C. Waller, Dalton Brady, Masaru Teramoto
As for limitations, it is important to note this was not a longitudinal study. Consequently, we cannot comment on the evolution of important prospective indicators of recovery, such as repeated symptom scores and quality of life measures. There are several forms of potential bias due to some of the data for this study being collected by retrospective survey. For example, nonresponse bias may be present due to a relatively low response rate (23.2% in this study), although the demographics of the survey responders and nonresponders were not significantly different. An inherent weakness of retrospective studies is recall bias. This study in particular employed surveys relatively far removed from the incident injury (average 3.9 ± 1.1 years). Caregivers could potentially have biased recall of their child’s recovery from an injury, particularly if they experienced other subsequent unrelated health challenges or injuries between the incident injury and time of survey. Additionally, because of the nature of cross-sectional studies, we cannot make causal inferences between potential injury-related predictors and recovery outcomes after C-mTBI. Certainly there are many non-injury factors that could affect recovery in mTBI, such as racial, ethnic, or socioeconomic factors, for which we did not control (11).
Related Knowledge Centers
- Longitudinal Study
- Sampling Frame
- Selection Bias
- Opinion Poll
- HIV/AIDS
- Self-Selection Bias
- Response Bias