Understanding statistical power in the context of applied research

Baguley, Thom

doi:10.1016/j.apergo.2004.01.002

Cited by 111 publications

(82 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Thus, 408 days of observation time per group (ie, 102 persons per group, with mean stay of 4 days) was required to detect a 50% reduction in this rate with 80% power and significance (P) less than .05. 19,20 …”

Section: Setting and Samplementioning

confidence: 99%

Reducing Pressure Injuries in Critically Ill Patients by Using a Patient Skin Integrity Care Bundle (Inspire)

Coyer¹,

Gardner²,

Doubrovsky³

et al. 2015

American Journal of Critical Care

122

View full text Add to dashboard Cite

Section: Setting and Samplementioning

confidence: 99%

Reducing Pressure Injuries in Critically Ill Patients by Using a Patient Skin Integrity Care Bundle (Inspire)

Coyer¹,

Gardner²,

Doubrovsky³

et al. 2015

American Journal of Critical Care

122

View full text Add to dashboard Cite

“…Although statistical differences were found in the results of the experiment, we acknowledge that the study did not have strong statistical power and was particularly limited by the small sample size as well as other aspects of statistical power (Baguley, 2004). Rather than focus on statistical power, the study reported here used participants with domain expertise in a naturalistic setting (Farrington-Darby & Wilson, 2006).…”

Section: Discussionmentioning

confidence: 99%

Impact of automation: Measurement of performance, workload and behaviour in a complex control environment

2015

View full text Add to dashboard Cite

This paper describes an experiment that was undertaken to compare three levels of automation in rail signalling; a high level in which an automated agent set routes for trains using timetable information, a medium level in which trains were routed along pre-defined paths, and a low level where the operator (signaller) was responsible for the movement of all trains. These levels are described in terms of a rail automation model based on previous automation theory (Parasuraman, Sheridan, & Wickens, 2000). Performance, subjective workload, and signaller activity were measured for each level of automation running under both normal operating conditions and abnormal, or disrupted, conditions. The results indicate that perceived workload, during both normal and disrupted phases of the experiment, decreased as the level of automation increased and performance was most consistent (i.e. showed the least variation between participants) with the highest level of automation. The results give a strong case in favour of automation, particularly in terms of demonstrating the potential for automation to reduce workload, but also suggest much benefit can achieved from a mid-level of automation potentially at a lower cost and complexity. Impact StatementResearch in the area of automation, and in particular in the examination of human interaction with different levels of automation, has normally been undertaken in Balfe et al, Impact of Automation: Measurement of Performance, Workload and Behaviour in a Complex Control Environment 2 laboratory settings using simple tasks and naïve participants where the level of automation can be easily manipulated. This research was undertaken with expert participants using complex simulation of three ecologically valid levels of automation and provides empirical field validation of some of the results found in laboratory studies.

show abstract

“…The lowess curves correspond to the 50th percentile (the median) and the 20th percentile. The 20th percentile was chosen because it corresponds to the conventional rule of thumb that power should be approximately 80% (Baguley, 2004).…”

Section: Evaluating the Function Mud When The Distribution Is Not A Mmentioning

confidence: 99%

Examining mixed unknown distributions (mud)

Wright

Skagerberg

2008

Behav Res

View full text Add to dashboard Cite

A function, written in R, for testing whether the distribution of responses in one condition can be considered a combination of the distributions from two other conditions is described. The important aspect of this function is that it does not make any assumptions about the shape of the distributions. It is based on the KolmogorovSmirnov D statistic. The function also allows the user to test more specific and, hence, more statistically powerful hypotheses. One hypothesis, that the mixture does not capture the middle third of the distribution, is included as a built-in option, and code is provided so that other alternatives can easily be run. A power analysis reveals that the function is most likely to detect a difference between the combined conditions' distribution and the other distribution when the center of the other distribution is near the midpoint of the two original distributions. Critical p values are estimated for each set of distributions, using bootstrap methods. An example from human memory research, exploring the blending hypothesis of the misinformation effect, is used for illustrative purposes. With stimuli like those in Figure 1, it is difficult to test for the blend explanation, because it is unclear what the blend of a stop sign and a yield sign should be. Because of this, some researchers (e.g., Belli, 1988; Loftus, 1977) have used original and postevent information that vary along a scale, so that it is easier to conceive of a blend of them. In our own research (Skagerberg & Wright, in press), we have used colors (where participants choose from a color chart), the height and build of a person, and count variables. In the example we use for illustration, we showed people several videos, including one of a train station. We consider three groups: (1) a control group who saw 18 people at the station, (2) a control group who saw 6 people at the station, and (3) an experimental group who saw 18 people at the station but then received postevent information from somebody who saw only 6 people. Problem. Let Dist 1 andWhen the participants were given a memory test, the means for these three groups were 16.71, 6.10, and 12.08 people, respectively. So, the means for the two control groups were near their true values, and t tests showed that the experimental group was significantly different from each of these. However, both explanations shown in Fig-D

show abstract

Understanding statistical power in the context of applied research

Cited by 111 publications

References 20 publications

Reducing Pressure Injuries in Critically Ill Patients by Using a Patient Skin Integrity Care Bundle (Inspire)

Reducing Pressure Injuries in Critically Ill Patients by Using a Patient Skin Integrity Care Bundle (Inspire)

Impact of automation: Measurement of performance, workload and behaviour in a complex control environment

Examining mixed unknown distributions (mud)

Contact Info

Product

Resources

About