Before 8 September 2026

Download:

Put all files into the same directory on your computer.

 

8 September 2026

I will give 3 talks discussing statistical thinking in support of the Biofilm Regulatory Toolbox. The data analyses discussed in each talk were gnerated by the R code in the R Modules below.   Run the R code in the Modules simply by copy and pasting code into R from the Word files or by running code chunks in RStudio from the Rmd file.

1000 - 1100 Basic Stats, Resemblance, Repeatabilty & Responsiveness

  • R Module 1: Resemblance/Repeatability of Legionella ISBR: data, Rmd, docx
  • R Module 2: ReRepeatability of Legionella ISBR: data entered manually, Rmd, docx
  • R Module 3: Responsiveness of Pseudomonas MBEC: data, Rmd, docx

1230 - 1330a Ruggedness and Reproducibility

  • R Module 4: Ruggedness of Pseudomonas STM: data, Rmd, docx
  • R Module 5: Reproducibility of Pseudomonas STM: data, Rmd, docx

1500 - 1600 Regulating Biofilm Claims with Statistical Confidence and Power

  • R Module 6: Performance Standard Assessment: data entered manually, Rmd, docx

 

9 September 2026

Hands On Options

Difficulty level: 1: easier <--> 3: harder

Exercise 1 on Modules 1 and 2 (level 1) 

  • Enter in your own control data into R Module 1 or your own log reductions (LR) into the R Module 2.
  • Is the mean log density or mean log reduction in an acceptable range at 95% confidence?
  • Is the repeatability acceptable?
  • Are there any outliers?  If there are any outliers, do they exert influence on your results?

Exercise 2 on Module 3 (level 1).  Module 3 analyzes an Oxidizer applied by MBEC.  In this exercise you will analyze either Phenol or Quat available in the same data set.

  • Quantify responsiveness.
  • Is there responsiveness at a 95% confidence level?
  • Visually inspect the outputted graph from plot.gam() to estimate a 95% confidence interval for the LR when log2(concentration) =6 (ie when the concentration is 2^6 higher than the base concentration).
  • Are there any outliers?  If there are any outliers, do they exert influence on your results?

Exercise 3 on Module 4 (level 3).  Fit a smoother model (GAM) instead of an LMM to assess ruggedness. 

  • Are the conclusions different?
  • Does the model fit the data better?
  • Are there any outliers?

Exercise 4 on Module 5 (level 1).  Module 5 analyzed a HIGH level of Chlorine applied by the STM.  In this lab you will analyze a different treatment available in the same data set.: LOW Chlorine, LOW or HIGH Phenol, LOW or HIGH Quat=alcohol.

  • Is the mean log density or mean log reduction in an acceptable range at 95% confidence?
  • Is reproducibilty acceptable?
  • Is repeatability acceptable? (pooled across labs)
  • Are there any outliers?  If there are any outliers, do they exert influence on your results?

Exercise 5 on Module 6 (level 2).  The current performance standard for the STM is to perform 3 tests and for a product to attain a LR > 6 in all 3 tests.  Consider a different performance standard and record the pass-error and fail-error rates by changing:

  • Required LR per test (maybe try 5).
  • Definition of ineffective products to keep out of the market (maybe try mean LR = 4).
  • Definition of highly effective products to allow into the market (maybe try mean LR = 7 or 8).
  • The number of required labs to 2 or 3
  • The number of required tests to 1 or 6.
  • What performance standard would you recommend?