China
Africa
Explore chapters and articles related to this topic
Toolkit for Assessing and Monitoring Leadership and Safety Culture
Published in Cindy L. Caldwell, Safety Culture and High-Risk Environments, 2017
When collecting and analyzing data, be aware of bias that can be introduced through the self-selection bias of the participants. Respondents who chose to participate and provide feedback may have strong views, which could potentially bias the results. This is especially true of the qualitative information in the form of comments from surveys, focus groups, and interviews. If participation is voluntary, staff that have no opinion, or who do not see value in participating, will either not attend or not participate, which means that results may be skewed toward the stronger opinions.
Willingness of Hurricane Irma evacuees to share resources: a multi-modeling approach
Published in Transportmetrica A: Transport Science, 2023
Stephen D. Wong, Mengqiao Yu, Anu Kuncheria, Susan A. Shaheen, Joan L. Walker
The dataset has several key limitations that limits some conclusions. First, the online survey exhibits self-selection bias as individuals opt into the study. We attempted to address this by providing a lottery incentive and asking over 10 agencies with different functions (e.g. transportation, emergency management) and news sources to distribute the survey. Second, we acknowledge online surveys have sampling bias as they only reach individuals with Internet access, often oversampling wealthier populations. Third, we also found that respondents were concentrated in three counties – Brevard, Lee, and Collier. The sample geographies are wealthier, more highly educated, and racially whiter than the impacted area and Florida. While we worked with several agencies in larger counties (e.g. Miami-Dade, Broward, Pinellas, Hillsborough), we found response rates to be lower, possibly due to the lower impact of Hurricane Irma in those areas and the more restrained survey distribution by agencies.
A Three Country Comparative Study of Social Commerce Adoption
Published in Journal of Computer Information Systems, 2023
Chuleeporn Changchit, Robert Cutshall, Joseph S Mollick
Most empirical studies have an innate limitation due to the sample used. Some of the limitations of this study come from a lack of balance in the sample, across dimensions such as age, gender, and employment status. The unbalanced samples cannot reliably represent the entire target populations of the three countries in the study. The unbalanced nature of the samples poses a threat to external population validity of the inferences drawn from the unrepresentative samples. Inferences drawn based on the unbalanced samples may not be true for the populations that have not been properly represented. In addition, the participants for this study were recruited mostly via social media posts. Thus, nonsocial media users were not represented in this study and any generalization must be limited to users of social media. The study also relied on self-reported data by the participants. Therefore, the generalizability of this study’s results may be limited due to self-selection bias and self-reporting bias.
Passive resistance to health information technology implementation: the case of electronic medication management system
Published in Behaviour & Information Technology, 2022
Eui Dong Kim, Kevin K.Y. Kuan, Milan Rasikbhai Vaghasiya, Jonathan Penm, Naren Gunja, Redouane El Amrani, Simon K. Poon
The data collection is conducted using mainly convenience sampling, where participants are selected based on their accessibility to the research. The key advantages of convenience sampling are that it is cheaper, more efficient, less time-intensive, easier, and simpler to implement than other sampling methods (Bornstein, Jager, and Putnick 2013). However, the key disadvantage of convenience sampling is that as not every unit in the population has a chance of being selected in the sample, it can have a self-selection bias, so the scientific generalisation derived from convenience sampling is limited (Bhattacherjee 2012). To complement this disadvantage of convenience sampling, the research team contacted as many clinicians as possible to collect data in person by visiting every ward. As such, the data collection for each survey took more than five weeks (much longer than originally planned), and almost all eligible eMMS users in the hospital were at least given a chance to be included in the sample. As a result, the risk of self-selection bias was minimised, and the shortcomings of convenience sampling are not an issue in this research.