7 replies to this topic

[denny]

httpss://www.experimentalbrew.com/podcast/episode-39-tanning-st-somewhere

In the pub we congratulate our new members of the  Governing Committee; explore what we think of the Brewer's Association recent announcement regarding sexist beer names and celebrate the sort of beery fool hardiness that lead one "intrepid" soul to attempt a beer half-marathon.

In the lab, we take a piece of feedback from a listener about what we did wrong with our BrewTan announcement and then, like a dog with a bone to worry, we decide to revist the results thanks to some new IGOR reports and watch how the numbers change. We'll discuss new directions that we want to explore as well.

Then in the Lounge we visit one of Drew's favorite brewers - Bob Sylvester of St. Somewhere Brewing in Tarpon Springs, FL. Learn just why we have a photo of a pump and a skateboard in the show notes and just how Bob produces "Florida terrior" and what measures he took to preserve it when the brewery moved.

Finally we tackle questions we can answer. Drop a quick tip on ya and explore Drew's new found home companion.

[HVB]

Drew's pronunciation of my name was very impressive!

[denny]

I'll let him know.  We actually got one more result after yours that makes the aggregate even more positive.

[HVB]

I'll let him know.  We actually got one more result after yours that makes the aggregate even more positive.

That is good to know, I also think the long term age test will prove better results.

[denny]

Yeah, I'm really looking forward to that one.

[Brauer]

Do you have the method details for the individual experiments somewhere? In particular, I would be interested in knowing which brewers experienced haze issues and which fermented in plastic.

I also wanted to caution you about the way you are performing the statistics on your mega-analysis, combining the results from multiple separate experiments, and suggest you discuss this data with a proper statistician. This is an approach that can be difficult, statistically, and has some restrictions on the groups that can be combined. Two of those restrictions are that the groups be of similar size and variance. In particular, I would be concerned that groups with very high variance also have large size. This can also be a way to introduce bias by increasing the size of outlier groups.

[denny]

Do you have the method details for the individual experiments somewhere? In particular, I would be interested in knowing which brewers experienced haze issues and which fermented in plastic.

I also wanted to caution you about the way you are performing the statistics on your mega-analysis, combining the results from multiple separate experiments, and suggest you discuss this data with a proper statistician. This is an approach that can be difficult, statistically, and has some restrictions on the groups that can be combined. Two of those restrictions are that the groups be of similar size and variance. In particular, I would be concerned that groups with very high variance also have large size. This can also be a way to introduce bias by increasing the size of outlier groups.

 

I don't have the fermenter issue, but I can tell you I always ferment in buckets.  What is your supposition?

 

We're only too aware of the difficulties in combining results.  A couple of our IGORs are PhD stats people and we're working with them to see if there's a better way to aggregate data.  In the meantime, ignore p values and just use the data as you wish.  In other words, take it or leave it! 

[Brauer]

I don't have the fermenter issue, but I can tell you I always ferment in buckets. What is your supposition?

We're only too aware of the difficulties in combining results. A couple of our IGORs are PhD stats people and we're working with them to see if there's a better way to aggregate data. In the meantime, ignore p values and just use the data as you wish. In other words, take it or leave it!

Brülosophy seems to repeatedly find a difference in beers fermented in plastic, so oxygen permeability is a potential explanation. This might be ameliorated by Brewtan B. Haze removal seems the most direct link between reduced oxidation and Brewtan B, so it would be interesting to see which brewers had persistent haze or required additional finings to achieve clarity.

Mostly, I'm just trying to think of any common variable that could explain why some brewers might see an effect and others not, since there doesn't appear to be a consistent trend. Without materials and methods it is difficult to evaluate an experiment and fully interpret the results.

I had thoughts on the statistical method, but it sounds like you have much more qualified people than me on that case!