[blair.streit]

Nice job!  As usual, interesting results.  Can someone please point me to a good explanation of all this p value stuff? Basically half of the people could ID and yet it is not significant?! 

I am pretty sensitive to lactic acid and actually believe acid malt is superior to lactic acid.  I've done the research and they are not the same.  Of course, I have plenty of skeptics but that's OK.  I would love to see a side by side: acid malt vs lactic acid someday...

I'm a long way removed from stats class, but in layman's terms the p value for these experiments is the odds that the results could be merely due to random guess. The lower it is, the more likely to be a significant result. They use a p value of .05, which basically amounts to a 5% chance it could be random guess or 95% confidence that the results are significant.

Remember, this is a triangle test so you'd expect 33% to be right, on average. Above that, and it is more likely to not be due to random chance. The actual calculations are where the statistics comes in, to determine how many positive results you would need from a given sample to hit a particular p value.

Great summary. I started typing a response last night and mine became so convoluted that I just discarded it

Since this one was "in the neighborhood" of significant he did include the preferences, which I also thought were interesting. Of the 11 people that successfully selected the odd beer out, only 4 selected the phosphoric acidified sample as the beer they preferred. 5 had no preference, 1 preferred the lactic sample and 1 could no longer tell the difference (not sure how that happens).

[denny]

Nice job!  As usual, interesting results.  Can someone please point me to a good explanation of all this p value stuff? Basically half of the people could ID and yet it is not significant?! 

I am pretty sensitive to lactic acid and actually believe acid malt is superior to lactic acid.  I've done the research and they are not the same.  Of course, I have plenty of skeptics but that's OK.  I would love to see a side by side: acid malt vs lactic acid someday...

Keep in mind that p value is not the be all, end all...it's merely a sign that further investigation is warranted.  We discussed this with Marshall the last time he was on the podcast to talk about one of our experiments.  Here something to look at ...http://www.stats.org/mismeasure-scientific-significance/ and

[mabrungard]

Since this one was "in the neighborhood" of significant he did include the preferences, which I also thought were interesting. Of the 11 people that successfully selected the odd beer out, only 4 selected the phosphoric acidified sample as the beer they preferred. 5 had no preference, 1 preferred the lactic sample and 1 could no longer tell the difference (not sure how that happens).

As noted, I don't think the lactate content in those beers would have been high enough to create a notable taste difference. The resulting scatter in the preferences should not be surprising. If the tap water had been a bit higher, I do believe that the tasters would have picked out the lactic acid more reliably.

[charles1968]

I like the idea of CRS. However, the fixed proportions of sulfuric and hydrochloric acids can end up pushing the beer flavor into the "minerally" range. The main problem is that CRS adds almost as much chloride ion as it does sulfate ion. We are better off tailoring the addition of either acid to produce a better result.

I pretty much agree, but everything depends on what you're brewing and where you want to end up. CRS is just another tool to have in the toolbox. For certain beers I can use CRS and nothing else and hit mash pH and mineral targets with tap water. In other cases I still resort to dilution with soft water or using gypsum etc.

[charles1968]

Keep in mind that p value is not the be all, end all...it's merely a sign that further investigation is warranted.  We discussed this with Marshall the last time he was on the podcast to talk about one of our experiments.  Here something to look at ...http://www.stats.org/mismeasure-scientific-significance/ and

Yes, the key thing is not to confuse evidence and proof. A significance test provides evidence, not proof. Scientific experiments can't prove a theory is true. Strictly speaking none of the exbeeriments has proved anything, but there's enough evidence to change the way we brew in some cases.

A big problem in science is "p-hacking" - cherry-picking trials that return significant results and not publishing the non significant ones. Hence the importance of transparency and being open with results however disappointing/unpalatable they are.