---

	FORUM ON BEER, HOMEBREWING, AND RELATED ISSUES
		Digest Janitor: janitor@hbd.org


***************************************************************
       THIS YEAR'S HOME BREW DIGEST BROUGHT TO YOU BY: 

          Northern  Brewer, Ltd. Home Brew Supplies
        http://www.northernbrewer.com  1-800-681-2739

    Support those who support you! Visit our sponsor's site!
********** Also visit http://hbd.org/hbdsponsors.html *********


Contents:
  Brew Pot Advice (thamric)
  Re: Leffe clone (Bill Rogers)
  Re: Sealing a conical lid - Why Bother? ("Don Van")
  I stand corrected - partially ("Jon Steinhauer")
  RE: Stupid pump questions ("Lou King")
  Saccharomyces infections (Kevin Kutskill)
  Harsh bitterness in CAP ("Dan Gross")
  Wrong Address Posted... ("Eric R. Theiner")
  Asheville Beer (David Perez)
  Saccharomycoses (Jon Steinhauer)
  Homebrewing in Czech Republic (Stan Burnett)
  chi-square and triangle tests ("Frank Tutzauer")
  Re: Brewer's yeast and yeast infections ("Steve Dale-Johnson")
  Triangel Test ("A.J. deLange")
  statistics and homebrew ("Christopher T. Ivey")


*
* Show your HBD pride! Wear an HBD Badge! 
* http://hbd.org/cgi-bin/shopping
*
* The HBD Logo Store is now open!
* http://www.cafeshops.com/hbdstore
*
* Beer is our obsession and we're late for therapy!
*

Send articles for __publication_only__ to post@hbd.org

If your e-mail account is being deleted, please unsubscribe first!!

To SUBSCRIBE or UNSUBSCRIBE send an e-mail message with the word
  "subscribe" or "unsubscribe" to request@hbd.org FROM THE E-MAIL
  ACCOUNT YOU WISH TO HAVE SUBSCRIBED OR UNSUBSCRIBED!!!**
IF YOU HAVE SPAM-PROOFED your e-mail address, you cannot subscribe to
  the digest as we cannot reach you. We will not correct your address
  for the automation - that's your job.

HAVING TROUBLE posting, subscribing or unsusubscribing? See the HBD FAQ at
http://hbd.org.

The HBD is a copyrighted document. The compilation is copyright 
  HBD.ORG. Individual postings are copyright by their authors. ASK 
  before reproducing and you'll rarely have trouble. Digest content 
  cannot be reproduced by any means for sale or profit.

More information is available by sending the word "info" to 
req at hbd.org or read the HBD FAQ at http://hbd.org.

JANITOR on duty: Pat Babcock and Karl Lutzen (janitor@hbd.org)



----------------------------------------------------------------------


Date: Tue, 22 Apr 2003 22:42:00 -0600
From: thamric at attglobal.net
Subject: Brew Pot Advice

Greetings,

After a dozen years of extract brewing, I am preparing to make the jump
to full grain brewing and have an equipment question. I secured a
surplus Sankey keg and have placed a hole in the top suitable for a lid
(so far so good). I also plan on having a SST nipple welded into the
side of the keg just above the bottom seam. I have read how others
recommend running a tube arrangement in to the center of the lower dome
to effectively drain all wort. 

Given that I plan to use this primarily as a brew pot (boiling vessel),
won't this arrangement simply ensure that I effectively drain  all of
the accumulated trub and hop residue from the bottom, or am I missing
something? Would it not be better to leave the drain port a couple of
inches off the bottom and away from the center to enable drain of the
clear wort only?

I've obviously never had the luxury of a drain in my brew pot, and would
benefit greatly from some sage advice.

Tim Hamrick
Boise, ID


Return to table of contents


Date: Tue, 22 Apr 2003 22:09:19 -0700 (PDT)
From: Bill Rogers <bill6beers at yahoo.com>
Subject: Re: Leffe clone

Jay Wirsig asks about a clone for Leffe Blond.  I haven't tried this recipe
because I've just recently bought the book, but it looks like it might be
close.  I'm actually very interested in trying the Rochefort 10 clone in this
book next. mmmm.  

************************
** Leffe Blond recipe **
************************
from "Beer Captured" by Tess amd Mark Szamatulski
OG=1.067, FG=1.016, SRM=7, IBU=26, abv=6.5

Mash:
11.75# Belgian Two-row Pilsner Malt
4 OZ. (113 g) Belgian Biscuit Malt
4 oz. (113 g) Belgian Aromatic Malt
4 oz. (113 g) German Munich Malt
2 OZ. (57 g) Honey Malt
Mash at 152F (65.6C).
Bring the water to a boil, remove from the heat and add:
8 oz. (226 g) Belgian Clear Candi Sugar
2 OZ. (57 g) Malto Dextrin
1/2 OZ. (21 g) Pride of Ringwood  at  9.3AA (4.7 HBU) (bittering hop)
Boil for 45 minutes then add:
1/2 OZ. (14 g) Styrian Goldings (flavor hop)
1 tsp. (5 ml) Irish Moss
Boil for 15 minutes.
Chill and pitch.
1st choice yeast: Wyeast 1762 Belgian Abbey II; Ferment at 68-72F (20-22C)
2nd choice yeast: Wyeast 3522 Belgian Ardennes; Ferment at 68-72F (20-22C)
Primary for 7 days, then rack to secondary.
Bottle when fermentation is complete, target gravity is reached and beer has
cleared (approximately 3 weeks) with:
1/2 cup (120 ml) Corn Sugar and 1/3 cup (80 ml) Belgian Clear Candi Sugar
that has been boiled for 10 minutes in 2 cups (473 ml) of water.
Let prime at 70F (21C) for approximately 5 weeks until carbonated, then
store at cellar temperature.

Note: Scott Murman has notes (http://smurman.best.vwh.net/zymurgy/wyeast.html)
which indicate that 1762 is Rochefort yeast and 3522 is Chouffe yeast. 
Myself, I think the 1388 -- Belgian Strong Ale Yeast (Duvel) yeast might be
really good for this beer.

Bill Rogers
Madison, WI




Return to table of contents


Date: Tue, 22 Apr 2003 22:27:07 -0700
From: "Don Van" <dvanv at earthlink.net>
Subject: Re: Sealing a conical lid - Why Bother?

Regarding sealing a conical lid ---why bother?
Fermentors don't have to be 100% sealed from the outside environment.  If
you ever go on a tour of Anchor Brewery in San Francisco, you will see
fermentors that are shallow open air fermentors.  This is not the only
brewery that uses open fermentors.   I have successfully used a simple
window screen placed on my plastic bucket to test the open fermentation
concept with no problem.  As long as the conical lid overlaps over the top
of your conical, it will keep things out and work fine.  Just don't leave it
in the fermentor too long.  when krousen (sp?)  is gone, transfer to a
closed system.

Relax, have a homebrew
Don



Return to table of contents


Date: Wed, 23 Apr 2003 01:06:22 -0500
From: "Jon Steinhauer" <jsteinha at charter.net>
Subject: I stand corrected - partially

As I said, my search was brief, and had I included "vaginitis" in the
parameters, it would have yielded several recent articles in respected
publications, that I have now seen the corresponding abstracts for.  I
apologized for my impulsive, perhaps reckless response.

However, I am still not convinced that the very small number of women in
these series are not immune compromised in some fashion.  Some of you may
have these articles sitting handily on your shelves, and can readily find
out.  They may have been diabetic (I strongly suspect this is often the
case), IgA deficient, on unrelated medications, etc. Perhaps they also
participated in bizarre behaviors that would predispose to such things.  I
also suspect that the selection criteria were subject to significant bias.
It would take a clinical history of baking or brewing for a living to even
suggest to most clinicians to culture for S. cerevesiae (Candidal vaginitis
rarely confirmed by culture).  It may turn out,  if appropriately studied,
that the proportion of women in those reports who are colonized with S.
cerevesiae does not differ from those with recurrent or persistent Candidal
vaginitis or even those without symptoms.  It would be also be interesting
to know if any had celiac disease (sprue),  with which anti-saccharomyces
cerevesiae antibodies are known to be associated, although the reason for
the association is unknown.

I am still not convinced that normal, healthy adults get S. cerevesiae
infections.

Regards

J. Steinhauer



Return to table of contents


Date: Wed, 23 Apr 2003 06:18:57 -0400
From: "Lou King" <lking at pobox.com>
Subject: RE: Stupid pump questions

Michael - you'll need to place your pump at the lowest part of your
system and prime by letting the flow of your liquid enter the pump
before starting the pump.  I wouldn't worry about pumping up some
distance.  I do this all the time with a similar pump from March.

The lower the pump is in the system, the easier it is to prime.

See http://www.lousbrews.com and click on "Lou's Brewery" .  You can
barely see the pump mounted on some wood on the floor next to the
fermenter in the first picture.  I regularly pump between the kettle and
the fermenter, and between the kettle and the hot liquor tank on top.  I
put the pump on a chair to pump to the hot liquor tank, but I think it
is because my hose is too short.

You can also see a similar pump on the RIMS page, and I will tell you
that this one is harder to prime because it is so close to the bottom of
the mash tun and the wort doesn't flow into it thoroughly because of
that.

Lou King
Ijamsville, MD
============

Michael <grice at binc.net> (of indeterminate location) said:

> I just received a pump from morebeer.com as a belated birthday
> present. It's the one sold by morebeer.com as model H315, 
> corresponding to March model 809HS.
>:
> First, do I need 
> to do anything to prime the pump besides starting it up 
> slowly? With the discharge valve cracked open and the suction 
> valve open completely, it seems.
> :
> Second, how critical is pump placement here? 




Return to table of contents


Date: Wed, 23 Apr 2003 06:25:39 -0400
From: Kevin Kutskill <beer-geek at comcast.net>
Subject: Saccharomyces infections

Just a point of clarification.  I believe the original question was that if
drinking homebrew could lead to kidney infections.  I agree with Jon, that
fungal infections of the kidneys only occur in immunocomprimised individuals
(i.e., HIV patients, chemotherapy patients, etc.).  These kinds of
infections are typically very difficult to get rid of, usually requiring
hospitalization/IV anti-fungal therapy.  I also agree with Alan, that
otherwise healthy women can have Saccharomyces vaginosis (a Saccharomyces
yeast infection of the vagina).

So, in response to the original post, no, drinking the scrumpy the night
before did not lead to Alex's wife's kidney infection--pure coincidence
(unless she is immunocomprimised).

Kevin
beer-geek at comcast.net




Return to table of contents


Date: Wed, 23 Apr 2003 06:32:34 -0400
From: "Dan Gross" <degross at starpower.net>
Subject: Harsh bitterness in CAP

I brewed my first Classic American Pilsner in early March and it's finally
in the keg.  There is a harsh bitterness in the finish that is not at all
pleasant.

I suspect that a couple of things might have caused the unpleasant
bitterness and wonder if anyone can help me sort it out.  My hopping
schedule should have been:
1oz Saaz -- first wort
 .5oz Cluster -- 45 min
 .25oz Styrian Goldings -- 15 min
Total boil time was planned to be 90 min.

Due to distraction during the boil, I missed adding the Cluster hops at 45
minutes and didn't realize it until the boil had reached 90 minutes.  Soooo,
I decided to extend the boil and simply added the Cluster after the boil had
already reached 90 minutes.  The Styrian Goldings was added after the boil
reached 120 minutes bringing the total boil time to 135 minutes.   This of
course meant that the Saaz added at first wort had boiled for 135 minutes.
I added water to the boil to bring the final volume up to 4 gallons rather
than the 5 gallons that was planned.
In my mind the extended boil and lower final volume may be the two primary
causes of the harshness.

I also wonder about my water chemistry and it's effect on this style.  The
only water data I have comes from the municipal water web site and it is
about two years out of date.  Another complication is that my area could be
served by one of two very different water sources (the Patuxent and Potomac
rivers), or perhaps a combination of the two.  The best information I can
get has my water with the following mineral content:

Ca  38ppm
Mg  10ppm
Na  20ppm
SO4  33ppm
Cl  36ppm
HCO3  unknown (I cannot find a reference to bicarbonate levels in the tap
water analysis, is this called something else in the anaylsis, perhaps
alkalinity?)
Total dissolved solids  226ppm
Hardness  133ppm
Alkalinity  84ppm

Thanks,
Dan Gross
Olney, Md




Return to table of contents


Date: Wed, 23 Apr 2003 07:33:50 -0400
From: "Eric R. Theiner" <rickdude02 at earthlink.net>
Subject: Wrong Address Posted...

Hey Guys,

My last message (a shameless plug for myself and my new European 
distributor) included the incorrect address for Das Bier!

If you're interested, try www.das-bier.de

Sorry.

Rick Theiner
LOGIC, Inc.




Return to table of contents


Date: Wed, 23 Apr 2003 08:38:52 -0400
From: David Perez <perez at gator.net>
Subject: Asheville Beer

James is looking for the brew scene in Asheville. Barley's Taproom, at 
42 Biltmore Ave, is my favorite. http://www.barleystaproom.com/  They 
have really good pizza and an excellent tap list. Try the Cottonwood 

The Green Man Pub aka Jack of the Wood, on Patton St. have 3 or 4 of 
their own brews with one of them on cask. Nice atmosphere, pretty good 
(not great) beer.

The Bier Garten, (don't remember the address but right in downtown) has 
an incredible bottle list and a good draft selection, and excellent food.

There is an Irish Pub across the street from Barley's that is pretty 
good but a bit smokey for my taste. I will get their name and address 
for you this evening if you need it.

Hope this helps and bring us back some good beer. You do have to drive 
through Gainesville after all.

Dave Perez
Hogtown Brewers
Gainesville, FL




Return to table of contents


Date: Wed, 23 Apr 2003 08:40:57 -0500
From: Jon Steinhauer <jsteinha at PATH.UAB.EDU>
Subject: Saccharomycoses

While I was supposed to be doing other things, I was able to retrieve three
full articles online at the office.

Nyirjesy, et al. in Obstet Gynecol 1995;86:326-329 acknowledge that "(S.
cerevesiae vaginitis) has been suggested in few cases; to our knowledge it
has never actually been demonstrate."  While this is not evidence, I assume
these investigators from Temple and Wayne State Universities reviewed the
literature.  They identified eight women out of a symptomatic 750 with S.
cerevesiae.  They admit that six of the eight had known "predisposing"
factors.  They do not disclose any of this information in their paper, nor
do they disclose whether they sought to rule out predisposing factors in the
other two patients.

McCullough, et al. from Stanford and Duke Universities and other
institutions in J Clin Microbiol 1998;36:557-62 retrieved 19 isolates of S.
cerevesiae from 16 vaginitis patients.  When comparing the study group (with
S. cerevesiae and vaginitis) to the control group (No S. cerevesiae and with
vaginitis) the rates of exposure to S. cerevesiae in the environment was not
found to differ.  They mention in their methods that they reviewed medical
records for "related medical history," but did not elucidate and showed not
even a summarization of this data.  Thus, we have no knowledge of the health
status of these individuals.

Posteraro, et al. from Italy in J Clin Microbiol 1999;37:2230-35 included 30
women with and without symptoms who had positive S. cerevesiae vaginal
cultures.  Only 16 were symptomatic (the others were "colonized" but not
"infected").  Interestingly, 22 of the thirty were pregnant (a state of
relative immune suppression, insulin resistence and glycemia and altered
vaginal environment).  They do not mention how many of the sixteen vaginitis
patients were pregnant.  No mention of other health conditions was
mentioned.

I conclude that while there is abundant evidence that S. cerevesiae can
colonize humans, there is no conclusive evidence that it can cause infection
in healthy, immune competent adults.  There is also no evidence that I have
found or in these articles that shows those with high environmental
(occupational or otherwise) exposure have a higher rate of colonization than
the general population.

A somewhat common theme among S. cerevesiae vaginitis patients in these and
other papers is recurrent vaginitis and women who repeatedly self-treat with
antimycotic agents.  Assuming some of the recurrent vaginitis is real (I
would lay down money saying some of it is not), it is impossible to say
without experimentation how repeated or prolonged exposure of S. cerevesiae
(or other rectal or vaginal flora) alters the virulence of this organism or
the vaginal environment.

I thank those involved for stimulating my interest, as I have learned much
that I did not know two days ago.

Regards,

Steinhauer


Return to table of contents


Date: Wed, 23 Apr 2003 09:16:42 -0600
From: Stan Burnett <stanb at xmission.com>
Subject: Homebrewing in Czech Republic

Ahoy, homebrewers!

I-ll be moving to the small town of Tabor, Czech Republic this fall for an 
extended 8-12 month visit.  I'd like to set up a home brewery with my 
brother-in-law (who has never brewed before, but speaks Czech fluently).

My brother-in-law has not been able to track down any homebrewers in Czech 
Rep.  (A Czech contract, Radek Hauser, can be found on 
several "international homebrew" sites, but has out of date information.)

Does anyone have any suggestions for finding homebrewing support in the CR 
or a neighboring country (probably Germany)?

Thanks,

Stan




Return to table of contents


Date: Wed, 23 Apr 2003 11:47:11 -0400
From: "Frank Tutzauer" <comfrank at acsu.buffalo.edu>
Subject: chi-square and triangle tests

Larry from Texas has some good advice on setting up the triangle test for
statistical analysis:

"A die can be used to help randomize how the samples are to be given to the
judges, such as whether they get two samples from population A or from B,
and where the unique sample will appear in the arrangement.  It is
important
that is be completely random!  For each judge, roll the die and arrange the
three samples as follows:
     1 = AAB
     2 = ABA
     3 = ABB
     4 = BAA
     5 = BAB
     6 = BBA
The judges are merely asked to identify which of the three samples is
different from the other two.  All you record is whether they are correct
or not, and feed this information into the Chi Square analysis."

Beware.  Major geekitude follows.

If we're going to get geeky enough to run statistical tests on this, then I
would use a z-test of proportions instead of a chi-square.  The problem
with the chi-square in this application is that there is only one degree of
freedom, and the chi-square runs into continuity problems.  The issue is
that the chi-square is a continuous distribution, but the situation here
involves a discrete variable (a binomial, actually), hence the choices are
only approximately distributed as a chi-square.  If the degrees of freedom
are large enough, then then the approximation is fine, but with only one
degree of freedom, many would argue that a continuity correction needs to
be applied.  The catch is that a continuity correction drastically
decreases the ability of the chi-square to detect differences.

Instead, I would side-step the issue by conducting a z-test of proportions.
As Larry correctly notes, the judges would be correct one-third of the time
by chance alone, so you test the null that the proportion of correct
answers differs from one third.  It is true that the z-distribution is also
continuous, but the adequacy of the approximation does not depend on the
degrees of freedom.  (In fact, the z-distribution has none.)  Instead, the
adequacy of the approximation depends on the sample size and the
proportion.  As long as the proportion is not too extreme (too small or too
large) and as long as the sample size is adequate, the z- would be fine.

If you really want to do this and need the appropriate formula, let me
know.  After counting the correct answers it will take you about 30 seconds
on a pocket calculator to get your results.  If, for some reason, you think
that the z-approximation is inadequate, then you are left with an exact
binomial test of proportions, which, for most people, is too much trouble
to calculate by hand (but I could show you how to do it).

As for 8 beers instead of two, I'll need to think the design through more
before I offer advice.

Geek mode off.

 --frank in Buffalo (on my way to administer a stats final to my grad
 students!)



Return to table of contents


Date: Wed, 23 Apr 2003 11:35:20 -0700
From: "Steve Dale-Johnson" <sdalejohnson at hotmail.com>
Subject: Re: Brewer's yeast and yeast infections 

Pete Calinski asks:

>What would we get if we cultivated it and stepped it up to make a 5 gallon 
>batch?

Isn't this is what Anheuser Busch means when it says they "know of no other 
beer which costs so much to make" as Budweiser?  ;)


Steve Dale-Johnson
Brewing at (1918, 298) Miles Apparent Rennerian
Vancouver, BC, Canada.






Return to table of contents


Date: Wed, 23 Apr 2003 22:48:05 +0000
From: "A.J. deLange" <ajdel at cox.net>
Subject: Triangel Test

Paul asked about the triangle test a couple of days ago. First, with 
respect to the distribution of beers among glasses, the ASBC MOA says " 
All possible arrangements of the two beers among the three glasses 
should be distributed approximately equally among the tasters by 
appropriate random distribution, e.g. by the use of a table of random 
numbers." Some other points of significance (1) the tasting room should 
be quiet and free from extraneous odors (2) the glasses should have been 
freshly washed and drained and should preferrably be ruby red (3) 
tasting should be conducted between 10 and noon or between 2 and 4. No 
more than two sessions per day with at least an hour between tests. Only 
one set of sample should be given to panelists in a session. (4) Serving 
temperature should be between 40 and 60 F. (5) Panelists should be 
isolated from one another so that their determinations are independent.

 Panelists should be instructed to smell, swirl and smell again. The 
more odoriferous sample (or samples) should be tasted after the less. 
The beer which is different from the other two should then be selected 
and then a determination shoud be made as to whether the different one 
tastes better or worse than the others, is hoppier or less hoppier or 
has more or less of whatever quality you are looking at.

The Tables are the "Bengtsen Tables" but are not particularly hard to 
construct (for a single or a couple of panelist numbers) using Excel or 
another spreadsheet. If you want to recreate the whole table a computer 
program should be written).

 A panelist choosing at random has a chance of 1/3 of picking the odd 
(the one that is different from the other two) beer. Thus if N panelists 
pick randomly the probability of choosing correctly m times is 
(m,N)*(1/3)^m*(2/3)^(M-n) where (m,n) is the binomial coefficient. The 
Excell function Binomdist will calculate this formula for you. So use 
Binomdist to calculate the probability of correct random choice of 0 or 
more out of N (this is 1), 1 or more out of N, 2 or more out of N and so 
on up to N or more out of (N). P(m or more out of N) = 1 - 
BINOMDIST(m,N,(1/3),1) + BINOMDIST(m,N,(1/3),0). Calculate this for m = 
0 up to the number of panelists, N.  Now decide on the confidence level 
you want and go down your table until you find the n which gives you a 
probability less than this or, and here is where the spreadsheet is 
beetter than the publsihed tables, find the number who picked correctly 
and read the probability that this could have happened from random 
guessing.  For example, suppose you had 25 tasters and 15 correctly 
identify the odd beer. The table so constructed shows the probability 
that this outcome was by chance is 0.6% i.e. it is very likely one of 
the beers is noticeably different from the other.

The next step is to take all the ballots in which the odd beer was 
correctly identified and count the number which are the same with 
respect to the property of interest. For example, suppose out of the 15 
panelists in the example, 5 say the beer is hoppier. If this is less 
than half, subtract from the number of correct votes (15) and interpret 
the result as saying that 10 find the beer less hoppy. To see if this is 
significant we need the probability that 10 or more find the beer less 
hoppy. To get this we need the joint probabilities P(10 say less 
hoppy,10 correctly picked), P(10 or 11 say less hoppy,11 correctly 
picked), P (10, 11, or 12 say less hoppy, 12 correctly picked and so on 
up to the number of panelists. Correctly picked refers to correct 
identification of the odd beer. The sum of these probablilities is the 
probability P(10 or more picked out of correct guesses). We want this to 
be less than the desired confidence level. The joint P(m or more say 
lesshoppy, M correctly picked) = P(m or more say less hoppy | M 
correctly picked)*P(M correctly picked) where the | is read "given 
that". This is (m,N)(1/2)^N + (m+1,N)(1/;2)^N + other terms up to (N,N). 
Again (m,N) is the binomial coefficient and, again, the Excel function 
Binomdist comes to the rescue. Putting the requisite spreadsheet 
together is more involved than I thought it would be when I started to 
write this so I won't give the further details. Someone with familiarity 
with statistics ought to be able to do it from what I've described. For 
the example, the  probability that out of the 15 who correctly 
identified the odd beer would call it less hoppy by chance (tossing a 
coin to make the decision) is about 1%. From this we conclude with high 
confidence that the test beer is different from the control beer and 
that it is less hoppy.

The tables are found in the MOA's and in Bengtsson, K, Wallerstein Labs 
Commun. 16:231(1953) so just go back to your Wallerstein Labs Commun 
collection from '53 and there you are. As the triangle test is used 
widely in the food and beverage industry you ought to be able to find 
the tables on the net (I couldn't, however) and elsewhere.

A.J.



Return to table of contents


Date: Wed, 23 Apr 2003 22:15:43 -0400 (EDT)
From: "Christopher T. Ivey" <cti3c at unix.mail.virginia.edu>
Subject: statistics and homebrew

Now there's a subject worth drinking about.

I like Larry's randomization suggestion (using a die).  I don't see this
as a problem for a "chi-square" or other contingency table approach.  My
impression is that Paul simply wants to know, for each trial, whether the
taster successfully identifies the "unique" sample.  Thus each trial will
have an outcome of "success" or "failure."  This seems like a classic
binomial problem, and the formula for calculating the test statistic
should definitely be in that dusty old textbook.  (Don't trust your Excel
spreadsheet calculation; the National Institutes of Standards and
Technology have slammed Excel's performance on standardized statistical
calculations in every release for nearly a decade now, though Microsoft
doesn't seem to care) Paul if you actually perform this test, please
share; this would be an example that would get students' attention!

Larry points out:
A statistical curiousity about the Chi Square test is that it cannot be used
to PROVE anything; all it can do is DISPROVE something.  That something is
called the "null hypothesis", which stated simply, is that there is NO
significant difference between the two populations A and B.

Me again:
 ...the populations being the "success" and "failure" to identify
the unique beer.  The characteristic that Larry reminds us about chi
square tests is of course not unique to this test, it is also true of
binomial and all hypothesis-based statistical tests.

I'm not sure what would be gained by Larry's suggestion of testing outcome
A against B, and A vs. C, etc. (other than a few more glurgs of tasty brew
downed!).  You might consider just using pairs of samples, randomly paired
as the same or different beer, in a double blind setup.  Okay 'nuf
geekiness--back to beer.

Homebrew in my basement: a wit, a few APA bottles, ESB, barleywine,
imperial stout.  Commercial beer: some 2003 Bigfoot Barleywine, 2003
Celebration Ale, a couple Pyramid Snowcaps (2003), and some Bell's
Expedition Stout.  Next to be brewed: I think it will be another APA.

Cheers,
Chris Ivey
Champaign, IL



Return to table of contents

---