QI Tools

Positive deviance: identifying deviants with probability distributions

KUSM-W logo
 
Enter inputs
  • Simple peer comparisons can be toxic(Chun, 2023). Before using this type of plot:
    • Have consent from participants to share results among themselves, or preferably, anonymize the data (Vilendrer, 2021 PMID 34479736). Consider identifyig the positive deviances, if they consent, as this allows reciprocal learning of subtle or tactic tactics (Ivanovich, 2015 PMID 26188970). Some groups may choose not to attribute data on the first improvement cycle until trust is obtained in the process.
    • Conduct the positive deviance seminar at the time, or soon after sharing data. The rational is that feedback can generate much emotion (Payne, 2016 PMID 27412170) and is more effective if the solution to improvement is provided(Hysong et al, 2009 PMID 19194332). However, these approaches are uncommonly done(Hysong et al, 2021 PMID 33830786).
  • Replace the data below in this example probability distribution with your observations or.
  • Upload a csv file formatted as in one of thes files: example file - proportions.csv or example file - means.csv. 1st row is column names?
  • Separate columns with commas. Click here to add commas.





Value

X-axis limits: to




Example scenarios

Scenario 1: Presidential horseshoes (proportion of ringers over multiple matches

  • In this analysis, was the proportion] of ringers among the presidents better or worse than the benchmark, Lafayette?
  • In this analysis, which points (indicate the points by their percentages on the x axis) would you choose to be positive deviants?

Scenario 2: Presidential sack races - average times over multiple races)

  • In this analysis, was the mean speed among the presidents better or worse than the benchmark, Lafayette?
  • In this analysis, which points (indicate the points by their percentages on the x axis) would you choose to be positive deviants?

Scenario 3: Presidential horseshoes (number of ringers over one match)

  • In this analysis, was the mean count among the presidents better or worse than the benchmark, Lafayette?
  • In this analysis, which points (indicate the points by their percentages on the x axis) would you choose to be positive deviants?
 
Hint: if bad spacing between title/footer text and the plot, open the PNG link above the plot in a new browser window and alter the image dimensions in the URL of the window. Click here to see screen capture of an example. opens in new window

About

Technical details

This is an OpenCPU application.

Please help improve this application
by editing the source code.

Resource link
Package Info link
Function Sourcelink
Source Codelink
Help Page (html)link