Source for "consumption of drinks" data?

[dfrankow]

Thanks for your work on this project.

The data in https://github.com/PhilipGriffith/AHPy#relative-consumption-of-drinks-in-the-united-states is a pretty bold assertion: using relative comparisons one can recreate a real-world distribution.

A source on this data would be very helpful.

I saw this same example in chapter 5 of "Decision Making for Leaders" (Saaty). In that chapter, it is unsourced. Did Saaty run that experiment? What were the circumstances? In fact, the data is described as from a single person. Here is the quote from that book:

"It will be seen that even though a person may not have an idea of the final numerical value, by making the comparisons according to his or her everyday knowledge a very reasonable derived answer can be obtained. In the illustration below the individual was asked to provide judgments as to which of the seven drinks listed in the matrix is consumed more in the United States, and how strongly more. What is desired is the relative proportion that each of the seven drinks has of the total. We have given both the derived estimate and actual results taken from statistical sources. An important point here is that the judgments were provided by a person who had no idea of the true answer."

P.S. There's a little typo: the (Tea, Wine) cell should be 3, not 2. Otherwise this data agrees with the book.

[PhilipGriffith]

Hi Dan,

Thank you for your comment! The source for the drinks data in the README is Saaty, T.L. (2008) ‘Decision making with the analytic hierarchy process’, Int. J. Services Sciences, Vol. 1, No. 1, pp. 83–98. In that article, the (Tea, Wine) cell is 2, not 3, hence the discrepancy:

Regarding a source for the data, on page 87 Saaty writes:

The information about actual consumption was obtained from the US Statistical Abstracts. We see the answers are very close and pair-wise comparison judgements of someone who knows can lead to very accurate results of drink consumption. ... Table 2 shows how an audience of about 30 people, using consensus to arrive at each judgement, provided judgements to estimate the dominance of the consumption of drinks in the USA (which drink is consumed more in the USA and how much more than
another drink?). The derived vector of relative consumption and the actual vector, obtained by normalising the consumption given in official statistical data sources, are at the bottom of the table.

There is no footnote in the article, nor is there a reference to which US Statistical Abstracts were used. It's possible that a reference exists in one of the many other works he published, though I can't say for sure. 🤔

If anyone else has an answer, please let us know!

-Philip

[PhilipGriffith]

I've also updated the README with the source, for future reference.

[dfrankow]

Thanks!

You give the source, so we might close this.

However, the matrix in this example is still ill-formed. The cell with 2 has 1/3 across the diagonal. These matrices have to be N and 1/N across the diagonal (same N), so it's either 3 and 1/3 or 2 and 1/2.

It's likely a typo in the original article.

[dfrankow]

Also, the text of how they came up with the matrix is interesting: "Table 2 shows how an audience of about 30 people, using consensus to arrive at each judgement, provided judgements to estimate the dominance of the consumption of drinks in the USA"...

So this source says 30 people as you reported. However, according to this report those 30 people combined their judgments with consensus, which is not what I would've expected.

[PhilipGriffith]

🤦 This is what I get for blindly entering numbers! I was using the properly formed matrix (and an earlier source) when testing the code, so the error is thankfully only cosmetic. The README has been duly corrected.

From what I can tell, the number of ways to arrive at the final judgments to be synthesized in a hierarchy is legion, so there's really no telling how they achieved consensus.

[dfrankow]

Thanks Phillip!