local host not responding

[ChrisCharlton]

I have now investigated the issue that you were seeing with scrollbars. It turns out that this is a problem with Chrome introduced in version 32 (see http://stackoverflow.com/questions/2116 ... -chrome-32). It appears that Google have now fixed this in the most recent update and the scrollbars now work again for me if I update to this (to manually update Chrome see https://support.google.com/chrome/answer/95414?hl=en-GB).

[wshadish]

I updated Google Chrome and now the scroll bars work as intended. With the other changes, I have now replicated up to p. 69 of the beginner's manual.

[wshadish]

I did upgrade from openbugs 3.21 to 3.22, and changed the settings tab accordingly. I am on page 70 of the beginner's manual. At this point, however, I get a somewhat different screen than shown by the manual. See attached snapshot of the screen I get. Specifically, I get a field called Explanatory Variables that does not appear on p. 70 of the manual. That page shows a field called "Explanatory Variables Random at District Classification". If I try to click Next without completing the "explanatory variables" box, it will not go to the next screen. If I pick a variable (e.g., cons) to fill in the box, it will go to the next screen, but then shows me the "Explanatory Variables Random at District Classification" and I have to pick cons again. Then what I get is a screen with both "explanatory variables" and "Explanatory Variables Random at District Classification" both filled in with cons. p. 70 says clicking next will lead to the run screen, but for me it leads to a screen with six more options ranging from number of chains to use default algorithms setting, and things go further astray from there. Ideas? Will Shadish

[ChrisCharlton]

Thank you for letting us know about this. The screenshot on page 70 is incorrect (it appears to be related to the next section). The correct screenshot should resemble that on page 66, except that rather than choosing eStat for the question Choose estimation engine you need to select OpenBUGS. After doing so the following questions should be the same as for eStat except that it won't include a Generate prediction dataset question.

When you get to page 73 you will also notice that the ordering of the questions in the screenshot does not exactly match what you see in the software. This is also a mistake in the screenshot, and the question Explanatory variables should be before the question Explanatory variables random at district classification. You will also need to provide an answer for the Name of output results question before you can select Next.

[wshadish]

Dear Chris, I'm through the TREE manual now. Here are final observations:

I'm not sure if I should expect to get absolutely identical results to the results you report for the comparison of different software packages. On p. 69, The means and sd and Model fit statistics are all identical, but ESS for sigma2_u is 807 rather than 806 and for beta_4 is 169 instead of 170.

Same for p. 71 Openbugs results. Identical means and sds, slightly different ESS for some coefficients. E.g., beta_3 is 5291 not 5296.

For the table comparing programs. I get identical results from eStat. My WinBUGS 1.4 gives generally similar but not identical results, though ESS is different often by 10%. JAGS and MLWin-MCMC give identical mean and sd but ESS is very slightly different (0-5 points). eStat Orthogonal gives identical means and sd’s but ESS again very slightly different. However, is the figure at the bottom of p 74 supposed to be giving the same results as the table comparing programs? It doesn’t seem to be the same. .

P 71. Should this be “Click Save”, not “Click Set”?

P 77 should “which clearly showing” be “which clearly shows”?

Same page, I didn’t ever find the browse button. However, Upload is an option on the Dataset dropdown menu, and I had no trouble uploading a dataset.

I've appreciated the ability to go through TREE and learn, and hope this has been helpful for you.

Will Shadish