China
Africa
Explore chapters and articles related to this topic
Natural Language Processing for Mental Disorders: An Overview
Published in Satya Ranjan Dash, Shantipriya Parida, Esaú Villatoro Tello, Biswaranjan Acharya, Ondřej Bojar, Natural Language Processing in Healthcare, 2022
Iacer Calixto, Victoria Yaneva, Raphael Moura Cardoso
Having stated the drawbacks and risks of using social media data, we note that the collection of data from participants with verified diagnoses does not come without its own limitations, in addition to its scarcity. Such data can also have biases: clinical texts produced by the clinician and the patient can suffer from reporting bias; observational bias can also be an issue since not all variables involved in predicting a certain clinical outcome are written down in clinical texts; sampling bias can be a serious issue if certain subgroups of a population have limited access to medical or psychological services, etc. There is, therefore, no single best solution and awareness of the limitations of various data sources is crucial. Based on the reviewed literature, such awareness is not always demonstrated and these limitations are not always explicitly discussed.
Online experiments
Published in Catherine Dawson, A–Z of Digital Research Methods, 2019
Are online experiments neutral and objective? Is it possible that your online experiment might be influenced by personal and structural bias, politics, culture, networks and/or society? What other types of bias can influence online experiments (selection bias, design bias, measurement bias and reporting bias, for example)? Is it possible to eliminate such bias? Paying attention to issues of validity and reliability will help to address these issues: see below.
Acute effects of sodium bicarbonate ingestion on cycling time-trial performance: A systematic review and meta-analysis of randomized controlled trials
Published in European Journal of Sport Science, 2023
João Paulo Lopes-Silva, Carlos Rafaell Correia-Oliveira
The risk of bias of the studies included in this meta-analysis was assessed according to the Cochrane Collaboration bias risk tool (Higgins et al., 2011): (a) random sequence generation (selection bias); (b) blinding of participants and personnel (performance bias); (c) blinding of outcome assessment (detection bias); (d) incomplete outcome data (attrition bias); and (e) selective reporting (reporting bias). Furthermore, two additional biases were deemed by the researchers: (1) gastrointestinal symptoms (trough questionnaire) and (2) supplementation identification (asking the participants to indicate which trial was performed – sodium bicarbonate or placebo). These aspects were categorized as unclear risk of bias, low risk of bias, and high risk of bias. It is worth noting that the bias evaluation has been only analysed if the study included in the present review analysed any gastrointestinal symptoms.
The concentration and prevalence of potentially toxic elements (PTEs) in cheese: a global systematic review and meta-analysis
Published in International Journal of Environmental Health Research, 2022
Zahra Hashami, Negar Chabook, Fardin Javanmardi, Reza Mohammadi, Moein Bashiry, Amin Mousavi Khaneghah
Reporting bias assessment includes different biases, including publication, time lag, location, citation, language, and outcome reporting bias. Therefore, any limitation in searching studies may lead to publication bias. In this study, two methodologies were employed to analyze publication bias. The first method, which employed a funnel plot, revealed that the distribution pattern in various chart regions is not uniform or consistent, resulting in biased visibility (Figure 2). Afterward, statistical tests such as Begg and Egger were used. The egger’s test revealed no evidence of bias) (95% CI: −2.03–44.52) and p-value = 0.07(. Whenever the confidence interval does not contain zero in its range, there is no considerable bias in gathered documents during systematic review procedures. Similarly to Pb, the result of the Cd funnel plot showed a bias since the distribution was inconsistent at different points in the graph (Figure 3). However, Begg’s test showed no considerable bias in collected documents during systematic review procedures of the mean concentration of Cd in cheese samples (p-value = 0.6).
The validity of situation awareness for performance: a meta-analysis
Published in Theoretical Issues in Ergonomics Science, 2022
Jonathan Z. Bakdash, Laura R. Marusich, Katherine R. Cox, Michael N. Geuss, Erin G. Zaroukian, Katelyn M. Morris
In general, selection or reporting bias inflates effect sizes and spuriously raises the number of significant results (Ioannidis et al. 2014). Here, we focus on two types of reporting bias; each was only partially addressed in our meta-analysis. The first type of bias is selective reporting, which is the tendency to report only analyses and/or measures that are significant and consistent with the hypothesis (Ioannidis et al. 2014), leaving out non-significant ghost results. The second type of reporting bias is publication bias (i.e., the file drawer problem: papers with significant results are more likely to be published than papers with non-significant results).