China
Africa
Explore chapters and articles related to this topic
Preparing Studies for Statistical Analysis
Published in Lynne M. Bianchi, Research during Medical Residency, 2022
Luke J. Rosielle, Lynne M. Bianchi
If a power analysis was not done in advance and a study fails to find statistical significance, investigators will not know if there really was no effect of the intervention or whether the study lacked the power to reach statistical significance (i.e., a type II error). Thus, null results are a problem because they are ambiguous and cannot distinguish between these two interpretations. One way of partially disambiguating a null result is to perform a retrospective power analysis. A retrospective power analysis will tell you the effect size you could have detected given the number of subjects in your study. If the retrospective power analysis indicates your study was able to detect a large effect size, then you can argue that your findings represent a true negative result, and not a type II error. Of course, doing a power analysis in the beginning is easier and will give others more confidence in your data.
Identifying a Topic and Conducting the Literature Search
Published in Lisa Chasan-Taber, Writing Grant Proposals in Epidemiology, Preventive Medicine, and Biostatistics, 2022
For this section, collect prior studies regardless of whether the studies found positive, negative, or null results. It will be important for you to describe the full state of the science on this topic and not to omit anything that does not support your findings. It can be reassuring to note that a common “research gap” is the simple fact that the prior literature has observed contradictory findings. Also, consider the fact that some of the reviewers will be authors of these contradictory or null papers and would not appreciate having their paper omitted from your summary.
Treatment of Dyslexia
Published in Kees P. van den Bos, Linda S. Siegel, Dirk J. Bakker, David L. Share, Current Directions in Dyslexia Research, 2020
Jack M. Fletcher, Barbara R. Foorman, David J. Francis, Bennett A. Shaywitz, Sally E. Shaywitz
The point is not that the research is flawed because of the definition. However, when definitional variability is introduced, there is much greater possibility for the presence of nonspecific factors unrelated to the treatment which would produce improvement and/or null results. Of particular interest is the need to address the presence of attention deficit disorder, which tends to occur in 25 — 50% of individuals with dyslexia (Shaywitz & Shaywitz, 1988).
The Effectiveness of Didactic and Perspective-Taking Interventions on Reducing Multiple Dimensions of Heterosexism
Published in Journal of Homosexuality, 2023
John A. DeBerry, Eve M. Adams, Cory J. Cascalheira, Tracie L. Hitter
The presence of null results may be attributed to at least two factors beyond the ones mentioned. First, it is possible that reducing hostile, aversive, and amnestic attitudes toward sexual minorities requires different interventions than those used in our study. Given that sexual minorities now have nationwide marriage rights and federal protections against employment discrimination (Movement Advancement Project, 2020), novel interventions need to target subtle forms of sexual prejudice that remain for sexual minorities (e.g., microaggressions; Nadal, 2019). The effectiveness of LGB speaker panels in which facilitators actively engaged their audience (Rye & Meaney, 2009), coupled with evidence that medium-sized effects on sexual prejudice follow educational interventions delivered by LGB people (Bartoş et al., 2014), suggest that future interventions might include young LGB individuals telling their stories of discrimination while inviting the audience to write down moments in which they may have enacted heterosexist prejudice. Perhaps these stories could be recorded and shared by allies to ensure they have utility at scale.
Framing Effects, Social Norm Perception, and Tolerance of Lesbian and Gay Individuals: Experimental Evidence From Slovakia
Published in Journal of Homosexuality, 2023
Andrej Findor, Matej Hruška, John A. Gould, Roman Hlatky, Zuzana Tomková, Miroslav Sirota
When it comes to the presentation of logically equivalent summary information, valence does not seem to affect perceptions of social norms. Respondents who read a positively framed headline about the Slovak population’s level of support for registered partnerships did not perceive social norms to be any more accepting than those who read the negatively framed headline. Given the logical equivalence of the two headlines, this result is not surprising—both groups were presented with the exact same information. It seems that the objective data within the normative signal overpowers any possible valence effects, especially if that data suggests that the majority is not tolerant. However, it is important to note that this null result may not hold for all logically equivalent presentations of information. In the experiment, we juxtaposed a “clear majority” with “every third Slovak.” This is only one way to present the same information. More specific depictions of summary information—e.g., comparing “49%” to 51%”—or more general depictions—e.g., “more than half” versus “less than half”—could evoke different reactions. Some studies use manipulations with embellished or strongly one-sided data to test similar claims. While more likely to identify significant effects, such designs sacrifice external validity. Further research could re-test this hypothesis by identifying the extent of divergence necessary for valence frames to make a difference.
Risk of autism spectrum disorder in offspring with parental schizophrenia: a systematic review and meta-analysis
Published in Nordic Journal of Psychiatry, 2023
Monir Shayestehfar, Amin Nakhostin-Ansari, Amirhossein Memari, Seyed Hossein Hosseini Asl, Faezeh Faghihi
There are also some limitations in the current investigation. Although our analysis represented a significant effect of parental schizophrenia on the risk of ASD in offspring, confounding factors and adjustments were not consistent among all the studies. In addition, we only included articles published in English, so relevant studies in a language other than English may have been missed. We identified considerable heterogeneity across the studies; however, we found that the difference among the studies may not be a major concern because our sensitivity analysis suggested that our result is robust and not dependent on a single study. Also, there were overlaps in the populations included in registry-based studies conducted in Sweden, which is another limitation of our meta-analysis as we included all studies in our synthesis. However, excluding studies did not significantly affect the outcome in our sensitivity analysis. In our systematic review and meta-analysis, most of the evidence came from cohort studies; therefore, there is the possibility that ASD diagnosis in offspring may be in parallel rather than a leading factor to parental schizophrenia. In some studies, psychotic disorders and schizophrenia in parents were not separated, and we included them in our meta-analysis, which is another limitation of this study. Finally, the possibility of publication bias is always a concern because studies with null results are less probable to be published than those reporting positive associations.