• Change for Fall 2014

·         More applets converted to javascript and instructions/solutions updated, and errata corrected

·         The context of Investigation 2.9 was replaced.  The goals are for the baseline proportions to be small so the difference in the two conditional proportions is difficult (or at least awkward) to interpret.  We also want to make sure there is visible skewness in the relative risk distribution to justify the transformation.  This comparison may be clearest by examining probability plots.  The term “efficacy” was also introduce but could be glossed over quickly.

·         Investigation 3.3 was split into two investigations and the discussion reordered. The first investigation now focuses on motivating the use of the t distribution for both p-value and confidence interval calculations.  The investigation makes heavier use of two applets rather than R or Minitab simulations (though this makes it more difficult to look at probability plots).  The second investigation gives students practice applying the t-procedures to the data and drawing conclusions, and then extending to prediction intervals.

·         Reworked Investigation 1.1 to help streamline the discussion slightly.

·         Added a third classic smoking vs. lung cancer study in Investigation 2.10.

·         A few practice problems were added and a few homework problems (Ch. 2 #86, Ch. 3 #84, Ch. 4 #66)

• Changes for Winter 2014
• New Investigation 1.15 using hypergeometric distribution as probability model for sampling without replacement (optional)
• Reordered investigations in chapter 3, more focus on t-distribution with one-sample and then sampling and then randomization
• Changes for Fall 2013
• Errata from Winter 2013 have been corrected
• New (optional) material at end of Investigation 1.1 providing an introduction to standard deviation
• Slightly fuller treatment of one-sample t-procedures with Investigation 3.3 (earlier)
• Replaced and reordered some contexts in Section 4.4 (more on confidence intervals, transformations moved to end)
• More integration of newer (javascript) applets, though still not taking full advantage of possible capabilities
• Included some R instructions for handling missing values, workspace functions now work with missing vlaues

• (1/22/13) iscambinomtest function has been updated to allow entry of either decimal or integer confidence levels (e.g., .95 or 95) and to reset graph window after showing confidence interval output
• Errata from Fall 2011 have been corrected
• Investigations 1.1 and 1.2 have been reworked a bit
• Optional technology instructions for importing data and producing bar graphs are now at the end of Investigation 1.1
• Investigation 1.2 does not assume technology until calculation of p-value
• Confidence intervals now refer to an interval of plausible values rather than a range of plausible values
• Investigation 1.6 has been reworked a bit and can be split into smaller chunks more easily
• Optional instructions for normal probability plots have been added to Investigation 2.1, Investigation 2.9, Investigation 3.6 (see also section of ch. 3 exercises)
• More summary of one-sample t-procedures has been added to Investigation 3.14
• Example 3.4 has been incorporated into the text
• A new function has been added to the workspace to generate boxplots using quartiles instead of hinges
• Summary function and dotplot function in workspace now allows more than 2 groups
• New javascript versions of the Test of Significance Calculator applet, which now also allows pasting in of raw data, and the Normal Probability Calculator applet, have been linked to from the data files page

From Fall 2012

• Continue to check for updates on errata page and posting new errata
• ISCAM Workshop updated 9/5 (text now uses "95" as the confidence level instead of ".95" and ISCAM Workshop functions expect "95" as the input value instead of ".95". Standard R functions still use ".95"
• Randomization Test for Quantitative Response applet now has separate boxes for counting above, below, and "beyond" (two-sided) inputted value.
• If you get a "disabled"/SSL message when trying to download HW solutions, you should be able to say no a couple of times and then the file will open.
• Investigation by Investigation comments for Chapter 2 updated (9/16) and small changes made to exercise files