The errata and typos in the first print run that were mentioned in the first thread on this topic have all be addressed in the second print run. Fortunately, the book was selling well enough that the publisher had to go to a second print run much sooner than any of us planned. That allowed me to get those corrections in.
Thanks to everyone who posted suggestions for changes!
Doug Hubbard
Mr. Hubbard! The book is very interesting and practical. I have one question: why the EVPI value in the book at page 91 is not at all in the same range as the value in the “Chapter-7-Value-of-Information-Example.xcl”? Thank’s!
First, be sure you actually put in the values as described in the example on pages 90 and 91 into the correct fields in the spreadsheet. That is, be sure you set the lower bound, upper bound, threshold and unit loss to 100,000, 1,000,000, 200,000, and $25, respecitively. This should produce a value of about half of what is in the book example. Then notice the notes in the spreadsheet that explain why the spreadsheet offers a more realistic calculation. Most importantly, it includes a “hard” lower bound. We know the demand can’t be less than zero in this example but the simple chart in the book doesn’t take that into account. The spreadsheet does, however. If you effectively remove the “hard” upper and lower constraints on the value by making them extreme, you will get a value very close to the chart in the book. For example, if you set “The absolute minimum the quantity can reach” as -10M and the absoulte maximum at +10M, then you should get $323k
I’ll make that a little clearer in the spreadsheet.
Doug
Mr. Hubbard, thank you for your prompt answer. I got it now!
Dear Mr. Hubbard,
Are there mistakes in Measure Anything 2nd edition – page 154 – measuring fish – “get a range of 3.8% to 6.2%… 25,984 fish in the lake.” ?
Sincerely,
Chaim Rotman
Chaim,
Thanks for your comment on How to Measure Anything. It looks like this example is carried over from the first edition without any change.
The second time I refer to the lower bound of the population proportion, it looks like I mistakenly use “.032” in the denominator. The population proportion could change slightly depending on whether we use the z-stat approximation of binomial distributions or if I use the the Bayesian Inversion spreadsheet (chapter 10; can be downloaded from book’s website). Neither gives exactly the answer i show in the book, but they are very close. The z-stat approximation gives a range of 3.87% to 6.13% for a 90% CI of fish population of 16,301 to 25,871. This is very close to what is in the book (16,256 to 25,984). I can’t replicate exactly what is in the book with either method, however, so I’m not sure how to explain the small difference.
On another note, you appear to be the first after 30,000 books have been sold to catch the “.032” typo. Thanks for your attention to detail.
Thanks for your input,
Doug Hubbard