ISCAM2 Errata : Sheet1

1 | Date | Paragraph or letter | Problem | Solution | Institution | Last Name | First Name | |
---|---|---|---|---|---|---|---|---|

2 | 1/30/2015 | Inv. 1.3 (g) | The wording of this problem is a bit too vague for students to fully understand what is being asked for. Most of my students said "Calculuate a p-value." as one way but then didn't know what a second way was. They didn't see that the problem was asking for HOW to calculate a p-value. | Perhaps this problem could be re-stated as, "Describe two ways that we could get the p-value in order to assess whether this result is statistically significant." I think this would make it clear that we are trying to find the p-value and just need to think about how we could do it. | Central College | Mills | Mark | |

3 | 1/13/2015 | PP 1.6A | seems like don't need (d) and (e) here, also links to old version of power applet | drop d and e and follow link to new version of power applet | Cal Poly | Chance | Beth | |

4 | 2/15/2015 | Inv. 1.14(m) | Sample size is not specified for std. dev. calculation. | Have students do it for both n = 5 and n = 20, since those are the two sample sizes in the investigation. | Central College | Mills | Mark | |

5 | 2/15/2015 | Inv. 1.14(n) | Sample size is not specified for the distribution that students should be viewing. | Have students do it for both n = 5 and n = 20, since those are the two sample sizes in the investigation. | Central College | Mills | Mark | |

6 | 2/15/2015 | Inv. 1.15(f) | It isn't until the paragraph following this question that it is specified that we will be counting the number of people voting for Kerry. | Perhaps it can be indicated earlier that we want to count Kerry voters so that students know how to define the population proportion in this problem. | Central College | Mills | Mark | |

7 | 2/24/2015 | Inv. 2.1 | First paragraph of investigation references teen hearing loss data from "Investigation 1.15" | This should actually reference Investigation 1.16 | Central College | Mills | Mark | |

8 | 2/24/2015 | Inv. 2.1 (Minitab version) | No room left after (g) for students to write in an answer | Central College | Mills | Mark | ||

9 | 2/24/2015 | Inv. 2.1 | In part (h), it would be helpful to emphasize that our NEW parameter of interest is the difference of the individual parameters for each population | Modify a couple of the sentences in this part: "Define the *new* parameter of interest to be ..." and "State appropriate null and alternative hypotheses about this *new* parameter..." | Central College | Mills | Mark | |

10 | 2/3/2015 | Inv 2.1 study conclusions | missing the negative sign in front of 0.024 (to match "or smaller") | add the negative sign | Cal Poly | Chance | Beth | |

11 | 2/3/2015 | p. 128 R version | Didn't leave enough room in (aa) for interpretation | Can use space in (bb) | Cal Poly | Chance | Beth | |

12 | 2/4/2015 | p. 166 R version | No room left after (w) | Answer on previous page or p. 167 | ||||

13 | 2/25/2015 | Inv. 2.1(o) | It might be good to help students understand that "more extreme" actually means that the difference is less than -0.024. | Maybe add some parenthetical statement reminding students that they have a one-sided alternative that defines what "more extreme" is. | Central College | Mills | Mark | |

14 | 2/25/2015 | Inv. 2.1(q) | Would it make more sense to have all three histograms showing up as in separate panels of the same graph, rather than as individual graphs? Maybe then it wouldn't work to fit a normal distribution to them like what is done later in (t)? | This is just an idea to help students have all three histograms available to them at one time. | Central College | Mills | Mark | |

15 | 2/4/2015 | Inv 2.2 before (f) | Before second diagram refers to "across the 3 groups" | Should be just 2 groups | Cal Poly | Chance | Beth | |

16 | 3/6/2015 | Inv. 2.6(l) | The applet's default setting is to look at the difference of proportions, not the "Cell 1 Count". | Probably should add an instruction here to change the Statistic to be "Cell 1 Count". | Central College | Mills | Mark | I will change default link from the iscam page to add &count=1 to the URL to default to cell 1 count when you use the dolphin specific version |

17 | 3/6/2015 | Inv. 2.6(q) | If you use 0.467 in the "Count Samples" field, then the p-value will not match when you use 10 successes. | Maybe instructions should tell students to use 0.466, because this difference will give the same p-value as with 10 successes. (Or maybe the applet needs to not be as sensitive to the thousandths position.) | Central College | Mills | Mark | technically .466666 is not larger than .467. The applet will now round simulated statistic to three decimal places before comparing to input level. |

18 | 2/8/2015 | Inv 2.6 (n) | Discussion references "Animate" box | Delete first part of sentence, start with Enter 995 | Cal Poly | Chance | Beth | |

19 | 3/9/2015 | Inv. 2.7(k) | In this investigation, there is no real reason given for why the p-value should be P(X >= 10) in part (k). It is not obvious from the work up to this point whether 10 is higher than expected or lower than expected. | PerhapsĀ a question right before part (k) that asks the students what proportion or count would be expected in the "seeded" group, expecting that they would use the overall proportion from the sample to determine an expected count. Then from that overall proportion, students could see that 10 successes was high and so the p-value is a right-tail probability. | Central College | Mills | Mark | At this point. also hoping they will determine direction from Ha from part (b) |

20 | 2/11/2015 | Inv 2.9, step 1 of simulation | input values are incorrect | N = 5168, M = 210, n = 2584, so use rhyper(10000, 210, 4958, 2584). Also in Step 3 use 210-VacInfCount | Cal Poly | Chance | Beth | |

21 | 2/7/2015 | Inv 2.9(w) | R version does not have enough space | fix page break | Cal Poly | Chance | Beth | |

22 | 3/4/2015 | pg 169, Inv 2.10 | Wynder and Cronfield data, table doesn't add up | other cancer row total should be 43 and table total should be 106 | Pomona College | Hardin | Jo | |

23 | 3/10/2015 | PP 2.10A | The p-value givenĀ in the problem (Fisher's exact test) is two-sided. | the one-sided p-value is 0.018. | Pomona College | Hardin | Jo | |

24 | 3/11/2015 | Answer key for Exercise #71(g) in Ch. 2 | Answer key for part (g) discusses error ratio, not the confidence interval requested in the statement of the problem. | Central College | Mills | Mark | ||

25 | 2/23/2015 | Inv 3.1 | R commands on p. 195 and 198 | Can now use iscamdotplot( ... names = c("times", "year")) and for iscamboxplot | Cal Poly | Chance | Beth | |

26 | 3/11/2015 | Inv. 3.1(j) | This question seems unclear. Do you mean the next shortest intereruption time after the minimum value or the next shortest time after the outliers? | the smallest non-outlier observation | Central College | Mills | Mark | |

27 | 2/23/2015 | Inv 3.1(m) R version | Need more space for comparison | Cal Poly | Chance | Beth | ||

28 | 2/23/2105 | Inv 3.2 CLT | typo in Key Result box | line 3 should read "or approximately normal if the population shape is NOT normal but the sample size is large" | Cal Poly | Chance | Beth | |

29 | 2/26/2015 | Inv 3.2 after (u) | typo in "In general" comment | should say "does NOT depend..." | Cal Poly | Chance | Beth | |

30 | 3/20/2015 | Inv. 3.3 | Is the sample size for the Shigella trials missing from the first paragraph of this investigation? The answer key for this investigation says that n = 130, but nowhere in the first paragraph is that given. | In this version of the investigation, students aren't told the actual sample size for a while but can still define the symbol n. (The simulations use 13 just to have the heavier tail in the t distrbution and to later compare the impact of the sample size on the analysis) | Central College | Mills | Mark | |

31 | 3/20/2015 | Inv. 3.3(c) | File name for body temp data has wrong extension. | both an xls and a txt version exist and can be used. The xls version can be helpful if you just want to take individual columns. | Central College | Mills | Mark | |

32 | 4/6/2015 | Inv 3.3 | Note that BodyTemps.txt appears to be the 130 "faked" body temperatures that have a mean of 98.25, so should NOT be used for the investigation. | Use BodyTempPop.xls or BodyTempPop.txt | Mills College | Erickson | Tim | |

33 | 2/24/2015 | Inv 3.3, paragraph before (f) | says the two-sided p-value is around 0.14 | probably closer to 0.10 | Cal Poly | Chance | Beth | |

34 | 3/4/2015 | Inv 3.5 tech detour | Below the box refers to "both simulations above" | in R or Minitab only version of text will only be one simulation code displayed | Cal Poly | Chance | Beth | |

35 | 4/20/2015 | PP 3.9 | Typo? (e) "... if both sample means are increased by 2 years..." I think it's 2 ounces, right? | correct | Mills College | Erickson | Tim | |

36 | 4/14/2015 | Inv. 3.13 | I see that the question asking students to calculate the proportion of observed differences falling in the 90% confidence interval has been removed in this version of the book. | I really liked this question, because it got students to understand that the confidence interval cannot/doesn't predict individual observations. Can this question be put back into this investigation, perhaps as a follow up to part (j)? | Central College | Mills | ||

37 | 9/16/2014 | Ch. 3, Exercise #2 | typo - The Pew Internet and American Life Project actually 12 regions of the United States. | It isn't clear how the sampling was done (cohort? cross classificiation?) so it is hard to tell whether a test of independence or homogeneity of proportions is appropriate. | Pomona College | Hardin | Jo | |

38 | 4/15/2015 | Ch. 3, Exercise #6 | There is a discrepancy between the number of mice cited in the problem and the number of data points in the CoolMice.txt file. The problem says there were 10 mice, but there are 19 mice in the file. | HW file should now say 19 as well | Central College | Mills | Mark | |

39 | 4/10/2015 | Ch.3, Exercise #37 | The formula given in this problem has a typo in it and doesn't fit the data given in BodyMassIndex.txt. | Formula should have multiplication by 10,000 instead of 1,000. | Central College | Mills | Mark | |

40 | 4/16/2015 | Inv 4.1 | Question (l) instructions for applet | Don't check the 2x2 box. Instead, check the Show Table box and edit the table to show the table of counts for the Spock data. Then copy and paste this table into the Sample Data window and press Use Table. I have also put a datafile of the raw data on the iscam2 files page | csbsju | G | J | |

41 | 4/20/2105 | Inv 4.5 | Typo, I think: "The researchers asked a restaurant to alternate, silence population music..." :) | popular? | Mills College | Erickson | Tim | |

42 | 4/16/2015 | Inv. 4.7(m) | It seems that this investigation overlooks the possibility of getting a correlation of 0 when the data falls along a perfectly horizontal line (i.e., constant y-coordinate). While this is a rather special (and unusual) case, it might be worth mentioning as an aside or comment at this point in the investigation. | (Posted 4/20/15) My mistake!! I forgot that the standard deviation of the y-coordinates would be 0, making the correlation undefined. | Central College | Mills | Mark | |

43 | 3/2/2015 | Inv 4.14 Technology Exploration | link to Masters13.txt is not updated, refers to 2013 Bristish Open and 2000 British Open | Don't use embedded link, go to http://www.rossmanchance.com/iscam2/data/Masters13.txt, change references to 2013 Masters | Cal Poly | Chance |