Outlier Impact and Accommodation Methods: Multiple Comparisons of Type I Error Rates

Liao, Hongjing; Li, Yanju; Brooks, Gordon P.

doi:10.22237/jmasm/1462076520

Cited by 48 publications

(31 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Outliers in charge amounts were addressed by Winsorizing data to the next highest data point within three standard deviations (SD) of the mean. 19 A test of the assumptions prior to conducting a one-way analysis of variance to identify differences in charges by injury location and injury type revealed heterogeneity across groups; therefore, we used non-parametric tests. The Kruskal-Wallis test was used to examine statistically significant differences in median charges by areas of injury on the body and charges by types of injury.…”

Section: Methodsmentioning

confidence: 99%

Injuries Associated with Hoverboard Use: A Case Series of Emergency Department Patients

Weingart

Glueckert

Cachaper

et al. 2017

WestJEM

View full text Add to dashboard Cite

IntroductionSince hoverboards became available in 2015, 2.5 million have been sold in the US. An increasing number of injuries related to their use have been reported, with limited data on associated injury patterns. We describe a case series of emergency department (ED) visits for hoverboard-related injuries.MethodsWe performed a retrospective chart review on patients presenting to 10 EDs in southeastern Virginia from December 24, 2015, through June 30, 2016. We used a free-text search feature of the electronic medical record to identify patients documented to have the word “hoverboard” in the record. We reported descriptive statistics for patient demographics, types of injuries, body injury location, documented helmet use, injury severity score (ISS), length of stay in the ED, and ED charges.ResultsWe identified 83 patients in our study. The average age was 26 years old (18 months to 78 years). Of these patients, 53% were adults; the majority were female (61.4%) and African American (56.6%). The primary cause of injury was falls (91%), with an average ISS of 5.4 (0–10). The majority of injuries were contusions (37.3%) and fractures (36.1%). Pediatric patients tended to have more fractures than adults (46.2% vs 27.3%). Though 20% of patients had head injuries, only one patient reported using a helmet. The mean and median ED charges were $2,292.00 (SD $1,363.64) and $1,808.00, respectively. Head injuries resulted in a significantly higher cost when compared to other injuries; median cost was $2,846.00.ConclusionWhile the overall ISS was low, more pediatric patients suffered fractures compared to adults. Documented helmet use was low, yet 20% of our population had head injuries. Further investigation into proper protective gear and training is warranted.

show abstract

Section: Methodsmentioning

confidence: 99%

Injuries Associated with Hoverboard Use: A Case Series of Emergency Department Patients

Weingart

Glueckert

Cachaper

et al. 2017

WestJEM

View full text Add to dashboard Cite

show abstract

“…All four DDS17 subscale scores and the Blood Sugar Testing subscale of the SDSCA required log transformation; the DDS17, PHQ-9 and GAD-7 total scores required square root transformation. Outliers (scores more than 3 SDs above the mean) were addressed using the winsorizing technique, in which outliers are replaced with the mean value plus 3 SDs [78]. Winsorizing is recommended to reduce the impact of outliers on the Type 1 error rate [78].…”

Section: Methodsmentioning

confidence: 99%

Web-Based Benefit-Finding Writing for Adults with Type 1 or Type 2 Diabetes: Preliminary Randomized Controlled Trial

Crawford¹,

Wilhelm²,

Proudfoot³

2019

JMIR Diabetes

View full text Add to dashboard Cite

Background The high prevalence of diabetes distress and subclinical depression in adults with type 1 and type 2 diabetes mellitus (T1DM and T2DM, respectively) indicates the need for low-intensity self-help interventions that can be used in a stepped care approach to address some of their psychological needs. However, people with diabetes can be reluctant to engage in mental health care. Benefit-finding writing (BFW) is a brief intervention that involves writing about any positive thoughts and feelings concerning a stressful experience such as an illness, avoiding potential mental health stigma. It has been associated with increases in positive affect and positive growth and has demonstrated promising results in trials in other clinical populations. However, BFW has not been examined in people with diabetes. Objective This study aimed to evaluate the efficacy of a Web-based BFW intervention for reducing diabetes distress and increasing benefit finding in diabetic adults with T1DM or T2DM compared to a control writing condition. Methods Adults with T1DM or T2DM and diabetes distress were recruited online through the open access Writing for Health program. After completing baseline questionnaires, they were randomly allocated to receive online BFW or an active control condition of online writing about the use of time (CW). Both groups completed 15-minute online writing sessions, once per day, for 3 consecutive days. Online measures were administered at baseline, 1 month, and 3 months postintervention. Participants were also asked to rate their current mood immediately prior to and following each writing session. Results Seventy-two adults with T1DM or T2DM were recruited and randomly allocated to receive BFW (n=24) or CW (n=48). Participants adhered to the BFW regimen. Greater increases in positive affect immediately postwriting were found in the BFW group than in the CW group. However, there were no significant group-by-time interactions (indicating intervention effects) for benefit finding or diabetes distress at either the 1-month or 3-month follow-up. Both the BFW and CW groups demonstrated small, significant decreases in diabetes distress over time. Conclusions BFW was well tolerated by adults with diabetes in this study but did not demonstrate efficacy in improving diabetes distress or benefit finding compared to an active control writing condition. However, due to recruitment difficulties, the study was underpowered and the sample was skewed to individuals with minimal diabetes distress and none to minimal depression and anxiety at baseline. Future research should continue to investigate the efficacy of variants of therapeutic writing for adults with T1DM or T2DM, using larger samples of participants with elevated diabetes distress. Trial Registration Australiand New Zealand Clinical Trials Registry ACTRN12615000241538; https://www.anzctr.org.au/Tria...

show abstract

“…On the other hand, we observe that recurrence -free survival times of the patients with tumor grade II and III have few outliers and positively-skewed distributed. Liao et al (2016) suggests that Kruskal-Wallis test should be used in such cases. Therefore, this test is also utilized along with ANOVA to compare groups.…”

Section: One-way Analysis Of Variancementioning

confidence: 99%

onewaytests: An R Package for One-Way Tests in Independent Groups Designs

Dağ¹,

Dolgun²,

Konar³

2018

The R Journal

106

View full text Add to dashboard Cite

One-way tests in independent groups designs are the most commonly utilized statistical methods with applications on the experiments in medical sciences, pharmaceutical research, agriculture, biology, engineering, social sciences and so on. In this paper, we present the onewaytests package to investigate treatment effects on the dependent variable. The package offers the one-way tests in independent groups designs, which include ANOVA, Welch's heteroscedastic F test, Welch's heteroscedastic F test with trimmed means and Winsorized variances, Brown-Forsythe test, Alexander-Govern test, James second order test and Kruskal-Wallis test. The package also provides pairwise comparisons, graphical approaches, and assesses variance homogeneity and normality of data in each group via tests and plots. A simulation study is also conducted to give recommendations for applied researchers on the selection of appropriate one-way tests under assumption violations. Furthermore, especially for non-R users, a user-friendly web application of the package is provided. This application is available at

show abstract

Outlier Impact and Accommodation Methods: Multiple Comparisons of Type I Error Rates

Cited by 48 publications

References 34 publications

Injuries Associated with Hoverboard Use: A Case Series of Emergency Department Patients

Injuries Associated with Hoverboard Use: A Case Series of Emergency Department Patients

Web-Based Benefit-Finding Writing for Adults with Type 1 or Type 2 Diabetes: Preliminary Randomized Controlled Trial

onewaytests: An R Package for One-Way Tests in Independent Groups Designs

Contact Info

Product

Resources

About