The <i>P</i> Value Problem in Otolaryngology: Shifting to Effect Sizes and Confidence Intervals

Vila, Peter M.; Townsend, Melanie; Bhatt, Neel K.; Kao, W. Katherine; Sinha, Parul; Neely, J. Gail

doi:10.1177/0194599816677735

Cited by 16 publications

(7 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, CONSORT guidelines do not recommend the elimination of p -value reporting. This too has been discussed elsewhere [4,7,29,31,32,33]. While we reference CONSORT frequently, it was not our intention to compare reporting in sports nutrition to the complete list of CONSORT recommendations.…”

Section: Limitationsmentioning

confidence: 99%

Reporting Characteristics in Sports Nutrition

Earnest

Roberts

Harnish

et al. 2018

Sports

View full text Add to dashboard Cite

Purpose: To examine data reporting characteristics in sports nutrition. Methods: We examined 236 papers from ten journals published in 2016. The primary outcome was statistical variance associated with treatment (SD (correct) vs. SEM or CI). Secondary outcomes included the reporting of: (a) effect sizes (Y/N); (b) outcome prioritization (Y/N; primary, secondary, etc.) and (c) statistical variance relative to change from baseline (CI (correct) vs. SD or SEM). As tertiary/exploratory outcome, we examined whether authors stated a directed hypothesis. Statistical evaluation was performed using chi-square analyses. Results: We observed significant trends for all analyses (p < 0.001) and between category comparisons (p < 0.002). For the primary outcome, 128 (59%) articles correctly used SD to denote treatment variance, while 79 (36%) and 11 (5%) used SEM and CI, respectively. For secondary outcomes, 63 articles (29%) reported effect sizes, while 155 (71%) did not. Additionally, 188 articles (86%) did not prioritize outcomes, 134 articles (61%) stated no hypotheses and 40 (19%, out of 100) articles used CI to denote change scores vs. SD (19%, n = 41) and SEM (n = 10, 5%). Eight articles (4%) reported no variance terms. Conclusions: Overall, there are gaps regarding reporting in sports nutrition. Editors, journal publishers, and the field of exercise science alike should consider these outcomes and provide editorial staff, reviewers and authors with more concrete guidelines.

show abstract

Section: Limitationsmentioning

confidence: 99%

Reporting Characteristics in Sports Nutrition

Earnest

Roberts

Harnish

et al. 2018

Sports

View full text Add to dashboard Cite

show abstract

“…Further, some researchers suggest that instead of relying solely on the p- value or adjusted alpha values in the case of multiple testing (e.g. Bonferroni adjustments, which make the alpha criterion more stringent), researchers should rely on effect sizes, logical conclusions and encourage replication studies (Bender and Lange, 2001; Moran, 2003; Perneger, 1998; Thomas et al , 2017; Vila et al , 2017). Therefore, we encourage readers to focus on effect sizes with the results presented below, particularly in relation to the context of the study and similar research within the field, as has been recommended by Vache-Haase and Thompson (2004) as an alternative to Cohen’s benchmarks, when possible.…”

Section: Correlations Among Perceived Personality and Judgments Of Vu...mentioning

confidence: 99%

Perceived personality traits and presumptions of vulnerability to victimization in women

Stewart

Smodis

Forth

2022

JCP

View full text Add to dashboard Cite

Purpose In women, having a history of sexual victimization has been linked to certain personality traits (e.g. low levels of assertiveness) and nonverbal behaviors (e.g. fewer head movements). The majority of research in this area, however, has considered how self-reported personality traits and gait relate to victimization. As such, the present study aims to examine how observers’ perceptions of personality impact judgments of targets’ vulnerability to sexual and violent victimization, and how the nonverbal behaviors used when making these judgments may vary depending on perceived personality traits. Design/methodology/approach A total of 309 participants watched eight audio-less videos of a woman speaking. Following each video, participants rated each woman on varying personality and emotionality traits, as well as their perception of how vulnerable the woman was to future victimization, and how they came to their decision according to a number of predetermined nonverbal cues. Findings Consistent with previous research, observers’ perceptions of sexual vulnerability were negatively related to perceptions of targets’ self-esteem and confidence, and positively related to anxiety. While violent vulnerability displayed a similar pattern of results, the nonverbal behaviors cited during the vulnerability appraisal process varied between personality traits. Though few results emerged within the latter query, anxiety exhibited the majority of all significant relationships, including being positively associated with facial expressions and upper and lower body movements. Originality/value Results suggest that different behavioral and personality interventions (e.g. increasing self-esteem) may serve to increase self-efficacy, autonomy and confidence, as well as help women feel more in control of their destiny and interpersonal communications.

show abstract

“…Not only the common used p-value for the CA-test is provided but also a compatible confidence interval for a well-chosen effect size are available in the line of the recent p-value controversy [29,2].…”

Section: Introductionmentioning

confidence: 99%

A modified Armitage test for more than a linear trend on proportions

Hothorn¹,

Schaarschmidt²

2020

Preprint

View full text Add to dashboard Cite

The Armitage test for linear trend in proportions can be modified using the multiple marginal model approach for three regression models with arithmetic, ordinal and logarithmic dose scores simultaneously, to be powerful against a wide range of possible dose response relationships. Moreover, it can be used for particular designs in the generalized linear (mixed) model for the three common effect sizes odds ratio, risk ratio and risk difference. The related R package tukeytrend allows simple generalizations, e.g. the analysis 2-by-k table data with a possible plateau shape or analysing overdispersed proportions. The evaluation of further real data examples are available in a vignette to that R package.

show abstract

The P Value Problem in Otolaryngology: Shifting to Effect Sizes and Confidence Intervals

Cited by 16 publications

References 8 publications

Reporting Characteristics in Sports Nutrition

Reporting Characteristics in Sports Nutrition

Perceived personality traits and presumptions of vulnerability to victimization in women

A modified Armitage test for more than a linear trend on proportions

Contact Info

Product

Resources

About