Discovering Statistics Using SPSS: Errata

What do you get if you let an idiot write a statistics textbook? Mistakes, and this page is dedicated to them.

Known Problems:

Some of these have probably been corrected in recent prints of the book.

Flash Movies: some of the movies haven't been done yet. This is because SPSS Statistics has changed a lot. In version 18 you can do things like bootstrap, which makes flash movies about the R-plugin a bit pointless. Rather than do everything twice I decided to wait a bit. I'll be rolling out the movies as and when I get access to SPSS 18 so that I can make them as up to date as possible.
Page 8: As a former rugby player this one is a bit embarrassing ... the text says a Hooker wears a number 1 shirt; a hooker actually wears the number 2 shirt, shirts 1 and 3 are props and I am an idiot.
Page 14-15, 17: As someone who did their PhD on conditioning this is even more embarrassing ... References to "negative reinforcement" are incorrect. What is described here is punishment - negative stimuli (verbal) to decrease the likelihood the behavior will happen again. I am even more of an idiot. (Too many people to mention spotted this mistake!).
Page 19: Reference to Jane Superbrain Box 2.2, should be 2.3. (I spotted this myself)
Page 24: In the last paragraph, the question ‘how likely is it that a 70 year old committed suicide at Beachy Head?’ suggests that we look at all people of age 70 and estimate the probability they will commit suicide (at Beachy Head). It should probably say ‘how likely is it that a person who committed suicide at Beach Head is 70 year old?’
Page 38 (Figure 2.5): The variable on the x-axis of both graphs should be ‘Lecture’ and it would be clearer if the titles of the graphs were ‘Lecturer 1’ and ‘Lecturer 2’.
Page 43: The last sentence starting ‘ The first thing to note is that most of the sample means are different from the true mean …’ should read: ‘The first thing to note is that the sample means are different from the true mean …’. (i.e. delete ‘most of’).
Page 51: Egron Pearson should be Egon Pearson.
Page 94: Reference to the open air festival "Waken" should be "Wacken".
Page 143 (8 lines up): “For numeracy scores, the distribution is slightly positively skewed in the Duncetown group (there is a larger concentration at the lower end of scores) whereas Sussex students are fairly normally distributed around a mean of 7. Therefore, the overall positive skew observed before is due to the mixture of universities (the Duncetown students contaminate Sussex’s normally distributed scores!)” Replace with “For numeracy scores, the distribution is slightly positively skewed (there is a larger concentration at the lower end of scores) in both the Duncetown and Sussex groups. Therefore, the overall positive skew observed before is due to the mixture of universities”. (Sandeep Deo)
Page 154: The sentence starting "Therefore if you only have one variable that has a skewed distribution..." should read: "Therefore, if you are looking at relationships between variables (e.g., regression) it is alright just to transform the problematic variable, but if you are looking at differences between variables (e.g., change in a variable over time) then you need to transform all of those variables".
Page 210 (Formula 7.9): The index i is missing on the last X of the first equation. The last X should read X_ni This was the typesetters fault :-)
Page 219: “Leverage values can lie between 0 (indicating that the case has no influence whatsoever) and 1 (indicating that the case has complete influence over prediction).” Should be changed to “The maximum value for leverage is (N-1)/N; however, SPSS calculates a version of the leverage that takes a maximum value of 1 (indicating that the case has complete influence over prediction” (Jeremy Miles)
Page 349 (section 10.2.2.): The equations illustrating the various ways in which the means can differ are incorrect. To see the correct version of this paragraph click here.
Page 380: The additional material for the Welch F went missing. Not sure why. Anyway, you can get a PDF of it here. (Spiros Mexas)
Page 385: “In summary, is an overall effect of Viagra on libido. Furthermore, the planned contrasts revealed that having Viagra significantly increased libido compared to a control group, t(12) = 2.47, p < .05, and having a high dose significantly increased libido compared to a low dose, t(12) = 2.03, p < .05.” Just to be 100% clear please add (one-tailed) to the end: “In summary, is an overall effect of Viagra on libido. Furthermore, the planned contrasts revealed that having Viagra significantly increased libido compared to a control group, t(12) = 2.47, p < .05, and having a high dose significantly increased libido compared to a low dose, t(12) = 2.03, p < .05 (one-tailed).” (Sandeep Deo)
Page 390: Reference to SPSS Output 8.7 should be SPSS output 10.7 and can be found on page 385.
Page 397: Under "Elimination of confounds", the text in brackets that reads "(i.e. variables that vary systematically with the experimental manipulation)" should read "(i.e. variables other than the experimental manipulation that affect the outcome variable)."
Page 442: If in your copy of the book the SPSS Tip 12.1 looks like this then the equations at the bottom of page 448 (and the paragraph above them) might be incorrect and should look like this.
Page 460: last column of Table 13.1, value 5 (first row) should be 4. (Karin Cleef)
Page 476: about halfway down it refers to lecturers marking. It should refer to differences in the times to retch after eating animals. I'm starting to wonder whether Jeremy Miles really read the proofs at all:-) (Karin Cleef)
Page 481: The first r (for stick insect v. kangaroo testicle) is incorrect. The answer should be 87 and the equation should be: sqrt(22.80/(22.80+7).
Page 558: r for alcohol group. In equation, -199 should be -1.99 (note decimal point) (Sandeep Deo)
Page 560 (formula 15.1): The Kruskal-Wallis formula should be "N(N+1)" not "N(N-1)". On the following page, (p561) the worked example uses the correct version of the formula (giving 80 581). Again, blame the typesetters for this one.
Page 615: Replace “Using Pillai’s trace, there was a significant effect of therapy on the number of obsessive thoughts and behaviours, V = 0.32, F(4, 54) = 2.56, p < .05. However, separate univariate ANOVAs on the outcome variables revealed non-significant treatment effects on obsessive thoughts, F(2, 27) = 9.73, p > .05, and behaviours, F(2, 27) = 5.23, p > .05” With (note 2 changes in yellow): Using Pillai’s trace, there was a significant effect of therapy on the number of obsessive thoughts and behaviours, V = 0.32, F(4, 54) = 2.56, p < .05. However, separate univariate ANOVAs on the outcome variables revealed non-significant treatment effects on obsessive thoughts, F(2, 27) = 2.15, p > .05, and behaviours, F(2, 27) = 2.77, p > .05. (Jeremy Miles)
Page 650: Section 17.6: In SPSS 17 and onwards PCA is under the 'Dimension reduction' menu. (Muhammad Mizanur Rashid Shuvra)
Page 672: Table 17.2, shows pre-rotation eigenvalues, not post-rotation, which would make more sense. (Factor 1 pre-rotation isn't necessarily factor 1 post rotation, so you can't give it the eigenvalue.) Replace the third from last row (currently 7.29, 1.74, 1.32, 1.23) with: 3.73 3.34 2.55 1.95 Replace second from last row (currently 31.70, 7.56, 5.73, 5.34) with: 16.22 14.52 11.10 8.48. (Jeremy Miles)
Page 678: The last sentence of the page: "Therefore, any items that result in..." should read "Therefore any items that have values of α in the column labeled Cronbach’s Alpha if Item is Deleted that are substantially greater than the overall α may need to be deleted from the scale to improve its reliability.
Page 730: The last sentence of the page should read "The variables in the data file, Cosmetic Surgery.sav, are:"
Pages 731, 758 and 760: The 'Reason for Surgery' variables are coded incorrectly (i.e. 'physical reason' should be coded as 0 and 'change appearance' should be coded as 1). Early versions of the data file "Cosmetic Surgery.sav" were also coded incorrectly and as such, did not match SPSS output 19.9 (p760). This data file has been corrected, therefore, you may need to re-download the newer version.
Page 738: The typesetters messed up the matrices for the diagonal and AR(1) covariance structures. To see what they should look like click here.
Page 752: In the sentence "All we're interested in at this stage is estimating the effect of including the variance in intercepts". Replace 'intercepts' with 'slopes'.

Thanks to the various people who have spotted these and alerted me to them.