| Hand in
each problem as you complete it: A is due Wednesday, Day 8
(Feb.14). The rest are due Monday Day 10 (Feb. 19) You don't need to complete them in order, but A is probably the one to start with. Use SPSS unless otherwise instructed. Print all graphs; copy down numbers onto the paper instead of printing those big tables. A. p. 33, 1.36 Rock sole . (This is file ex01-36.por; use filetype "All files" to find it.) To hand in: the SPSS stemplot, also a histogram, and find and write on your paper the mean, s.d. and 5-number summary. Also Describe as instructed in the book. Note: your histogram may not have a label on the X-axis, and it should! SPSS v. 14 on doesn't label the graphs unless the Variable View, Label column has labels in it! Go back and label your variables, now do your graph again.. B. p. 34, 1.38 Timeplot of rock sole. (the file is still the one from 1.36; files are not always under the problem number.) C. p.60, 2.36, breastfeeding and bones, comparing 2 groups. (Old saying: You lose a tooth with every child. Truth in it??) Note the form of the data file: all the change in mineral content values are in one column (bone) and a column (group) identifies each case with 1 for "other" and 2 for "breastfeeding" (check that's right by looking at the data in the book) Put in Labels, check that Measure is right. (Optional, add Value labels; like handout p. 2 bottom, to identify the 2 groups). I suggest side by side histograms, or parallel dotplots (handout p. 5) to start with. Who-Where-etc. Question about the data: Are these First-world women who are dutifully taking their calcium pills? No clue. D. Investigate the issue with data wrongly labeled Nominal or Ordinal. Use the file Studat for SPSS (Which you changed and saved after working p. 2) (SPSS for class \ Studat Completer is what you should have gotten; use that one if you don't have yours. Some computers don't have it.. Studat complete doesn't have labels.) Make a dot plot and a histogram of Pulse. Now change the Measure for Pulse to Ordinal, in the Data Editor. Try making a dot plot and a histogram of Pulse again. (The icon for Pulse should now not be a ruler. If it is, hit Reset, so the icon becomes the 3 towers. Then proceed.) Print the graph(s), compare to the first graph(s), and write on your paper what has happened, and what is wrong with the Ordinal graph(s). Hint: look at the "ruler scale" of the dotplot. Can this happen to you? Load the file for 1.37 study hours (You did this problem by hand.) What are the Measure values for these variables, and what should they be? E. Import a file: Go into the folder PC-BCSBigData (there next to PC-SPSS) and open class.dat. (The Text Import wizard messed up variable names for me, I had to fix them afterward. ) From the README file: DATA: class.dat Data provided by students (n=270) in a freshman-level course. SEX 0=Male, 1=Female HAND 0=Right-handed, 1=Left-handed HGTH Height in inches STDY Time spent studying on a typical week night (minutes) Beware outliers! [is this the source of ex1-37?] COIN How much money in coins (not bills) are you carrying? INC Guess the income of a "typical American family" Check the Measure, put in appropriate labels. Make a histogram just of Height (using Count), and side-by-side histograms of Female and Male heights. Are the single-sex histograms roughly bell-shaped? Describe the shape of the mixed sex (total) histogram. The "Living Histogram" photo passed around in class with mixed sexes showed two clear equally high humps. Does the mixed-sex (total) histogram here show the same pattern? Why not ? (Comment: if you did the Female and Male heights with Percent on the vertical axis, you wouldn't be able to see as well what causes the problem.) |
| Sievers home | Math151-Sp07/DaySp6.htm | 8pm | 2/11/07 |