Statistical Methods in Engineering and Science, STAT 390
General Items:
IMPORTANT: SYLLABUS; tentative and evolving. UPDATED: 6/2/21, last lecture canceled to make time for doing last hw set. (Remember, to play a game, you need to know the rules; and this syllabus describes the rules.
Concepts you will learn.
An evolving list of typos in our text book.
-------------------------------------------------------------------------------------------------- R_intro , R_primer ,
R_ref_card . All required data sets can be accessed at http://www.stat.uw.edu/marzban/390/spring21/xyz.txt, where xyz.txt should be replaced with the name of the data file.
If that fails for permission reasons: try http://sites.stat.washington.edu/marzban/390/spring21/xyz.txt
prelab1 (3/30); qz1, and the rubric.
prelab2 (4/6). There is no prelab2! qz2, rubric.
prelab3 (4/13); qz3, rubric.
No prelab4 (4/20)
prelab5 (4/27); qz5, rubric.
prelab6 (5/4); qz6, rubric.
prelab7 (5/11); qz7, rubric.
prelab8 (5/18); qz8, rubric.
prelab9 (5/25); qz9,
prelab10 (6/1); qz10, rubric.
-------------------------------------------------------------------------------------------------- Lecture Material:
Week 1:
Lecture 1 (Ch. 1) 3/29: Learn to deal with ambiguity. Week 2:
Lecture 4 (Ch. 1) 4/5: Distributions. Week 3:
Lecture 7 (Ch. 1) 4/12: Derivation of Binomial and Poisson. Week 4:
Lecture 9 (Ch. 2) 4/19: E of binomial, distribution variance, and qqplots. Week 5:
Lecture 12 (Ch. 3) 4/26: ANOVA formulation of regression. Week 6:
Lecture 15 (5.5, 5.6) 5/3: The sampling distribution. Week 7:
Lecture 18 (Ch. 7) 5/10: The t-distribution, t-based CIs, and 2-sample CIs (for mu1-mu2 and pi1-pi2). Week 8:
Lecture 20 (Ch. 8) 5/17: hypothesis testing with the p-value: Part I Week 9:
Lecture 23 (Ch. 8) 5/24: Chi-squared test of multiple proportions/categories in ONE population. Week 10:
Lecture 26 (Ch. 11) 6/2: CI for mean y(x), and PI for a single y*. -------------------------------------------------------------------------------------------------- Homework Material:
FIRST, consult the homework IMPORTANT RULES . SECOND, here is a master list for the problems from which some homework and test
questions will be selected:master_list.
And here are the solutions(doc) and solutions(pdf). Additional problems will be posted separately, within the lecture notes.
All required data sets can be accessed at http://www.stat.uw.edu/marzban/390/spring21/xyz.txt, where xyz.txt should be replaced with the name of the data file.
If that fails for permission reasons, try http://sites.stat.washington.edu/marzban/390/spring21/xyz.txt
Homework Assignments (and Solutions):
HW Set 1: Due 4/5, 2:20pm. (single pdf, uploaded to canvas).
Solutions and the rubric.
HW Set 2: Due 4/12, 2:20pm.
Solutions (typo in hw_lect6_2 has been corrected,
+ rubric.
HW Set 3: Due 4/19, 2:20pm.
Solutions + rubric.
HW Set 4: Due 4/26, 2:20pm.
Solutions + rubric.
HW Set 5: Due 5/3, 2:20pm.
Solutions + rubric.
HW Set 6: Due 5/10, 2:20pm.
Solutions + rubric.
HW Set 7: Due 5/17, 2:20pm.
Solutions + rubric.
HW Set 8: Due 5/24, 2:20pm.
Solutions + rubric.
HW Set 9: Due 5/31, 2:20pm.
Solutions
HW Set 10: Due 6/4 (Friday), 8pm.
Solutions
--------------------------------------------------------------------------------------------------
Grade-related:
ALL SCORES: You can see your hw, qz, and test scores in the Canvas Gradebook. But you can find also find them here HERE (UPDATED 6/12).
IT IS YOUR RESPONSIBILITY TO ASSURE THAT THESE GRADES ARE CONSISTENT WITH YOUR
OWN RECORDS OR THOSE IN CANVAS. LET ME KNOW IF THEY ARE NOT. Look for the
last 4 digits of your student ID (with 0s not shown). A "-99" is code for "no hw turned in," and it will be converted to a 0. If it's in error, let me know. A "137" is code for other arrangements made for that quiz's grade. You don't need to know about these other arrangements, because they are private for the specific student.
TEST 1: histogram. The green line is at the
mean, and the red lines show the 25th, 50th (median), and 75th percentiles,
i.e., the quartiles. Other percentiles are shown in read along the x-axis.
TEST 2: histogram.
Here is he scatterplot between the first and second test from this quarter. So, the two tests were comparable in terms of ease/difficulty; otherwise, we would have seen a vertical shift of the scatterplot relative to the diaginal line. Also, as you can see things change a lot from test to test. In other words, if you're on the top side of the histogram, congratulations! But, don't relax, and keep-up the work. If you're on the lower side of the histogram, don't despair; talk to me during the off hrs to see if I can make any suggestions. Again, keep in mind that all of these scores are tentative.
TEST 3: histogram .
The scatterplot between all three tests. It looks like test3 was easier than both test1 and test2.
SUMMARY: The histogram of hw_avg, qz_avg, test1, test2, test3, and score
are shown HERE (UPDATED 6/12). After you find your
percent score on each of these items (hw, qz, ...) in the scoresheet above,
then this figure will show you how you're doing relative to class on each of
the items (hw, qz, ...). The histogram of the "Score" (last histogram)
determines the final grade as follows.
GRADES: I determine (mostly) two cutoffs: The 4.0 cutoff on the high side
of the histogram of Score, and the 2.0 cutoff on the lower side.
The cutoffs themselves are placed where there are
big gaps in Score; this way, small changes in Score (e.g., from a poor test
grade) do not affect who "passes" at the 2.0 level and who doesn't. In placing
the bottom cutoff I also look at individual test grades in percent.
The 2.0 cut-off ended-up being around a Score of 0.50 (i.e., 50%), and
the 4.0 cutoff was around 93%.
At the end of the quarter,
I also look at similar cutoffs for all other grades, to make sure that the
final 4-point-scale is insensitive to small changes of any
individual item that goes into the Score.
Based on all the items shown above, the
summary scores are HERE (UPDATED 6/12).
The format is CL_avg (this clicker score should be ignored), hw_avg (after dropping lowest), qz_avg (after dropping lowest), test1, test2, test3, Score (weighted sum of the various items), and the corresponding percentiles,
followed by the grade.
The histogram of the grades is HERE (UPDATED 6/12).
The green line is at the mean, and the red lines show the 25th, 50th (median), and 75th percentiles.
------------------------------------------------------------------------------------------------ Student evals from my past stat/math 390 classes:
Winter15 ,
Spring15 ,
Winter16 ,
Spring16 ,
Summer16 ,
Winter17 ,
Spring17 ,
Summer17 ,
Winter18 ,
Spring18 ,
Summer18 ,
Winter19 ,
Spring19 ,
Summer19 ,
Autumn19 ,
Winter20 ,
Spring20 ,
Summer20 ,
Autumn20 ,
Winter21 .
Grades from my past stat/math 390 classes:
Winter15 ,
Spring15 ,
Winter16 ,
Spring16 ,
Summer16 ,
Winter17 ,
Spring17 ,
Summer17 ,
Winter18 ,
Spring18 ,
Summer18 ,
Winter19 ,
Spring19 ,
Summer19 ,
Autumn19 ,
Winter20 ,
Spring20 ,
Summer20 ,
Autumn20 ,
winter21 .
--------------------------------------------------------------------------------------------------
Lab/Quiz Material:
Download R, because many of the hw problems will
require R.
lab "book"; different parts of this book will be assigned weekly (called prelabs).
--------------------------------------------------------------------------------------------------
Lecture 2 (Ch. 1) 3/31: Types of data, and histograms.
Lecture 3 (Ch. 1) 4/2: Interpretation of histograms, and probability.
Lecture 5 (Ch. 1) 4/7: Named Distributions, and left-areas under standard normal.
Lecture 6 (Ch. 1,2) 4/9: Area under N(mu,sigma), percentiles, and boxplots.
Lecture 8 (Ch. 2) 4/14: Sample mean, sample variance, sample standard deviation, and distribution mean.
Test1 4/16. See zoom/recording on 4/21.
Lecture 10 (Ch. 3) 4/21: Scatterplot, correlation.
Lecture 11 (Ch. 3) 4/23: Correlation "defects", and regression.
Lecture 13 (Ch. 3) 4/28: nonlinearity, transformations, polynomial regression, and overfitting.
Lecture 14 (Ch. 3) 4/30: multiple regression (collinearity and interaction).
Lecture 16 (Ch. 7) 5/5: CI for true/pop mean mu_x.
Lecture 17 (Ch. 7) 5/7: minimum sample size, and CI for true/pop proportion pi_x.
Lecture 19 (Ch. 7) 5/12: Examples from previous lect, and paired CIs.
Test2 5/14. Wed. 10AM I'll go over the test, rubric ...
Lecture 21 (Ch. 8) 5/19: hypothesis testing with the p-value: Part II
Lecture 22 (Ch. 8) 5/21: 1-sample and 2-sample tests, for means and proportions.
Lecture 24 (Ch. 9,11) 5/24: 1-way ANOVA F-test, and the probability model of regression.
Lecture 25 (Ch. 11) 5/28: t-test of betas, and F-test of model utility.
Lecture 27 (FYI) Neural Networks, cross-validation, bootstrap, and more.
Test3 6/8. I have set-up an off hr (F. June 10, 10-11AM) to go over the test and the rubric. Stay tuned ...
--------------------------------------------------------------------------------------------------
HW1: hw_lect1_1 (see lecture 1 above), and go over prelab1 BEFORE Tuesday's lab; there will be a quiz on it. No book problems today.
HW2: hw_lect2_1 - hw_lect2_2, 1.16 (by R).
HW3: hw_lect3_1 - hw_lect3_2, 1.11(b; by R or by hand).
HW4: hw_lect4_1 - hw_lect4_3, 1.19.
HW5: hw_lect5_1 - hw_lect5_3, 1.32(e).
HW6: hw_lect6_1 - hw_lect6_4, 1.34(b)
HW7: hw_lect7_1 - hw_lect7_3, 1.55, 1.75 (Hint: This is a 2-step problem, one involving Normal, and the other involving Binomial).
HW8: hw_lect8_1 - hw_lect8_4, 2.16.
HW9: hw_lect9_1 - hw_lect9_5.
HW10: hw_lect10_1 - hw_lect10_3, 3.18.
HW11: hw_lect11_1 - hw_lect11_5.
HW12: hw_lect12_1 - hw_lect12_3.
HW13: hw_lect13_1 - hw_lect13_3, 3.31, 3.34 (this one doesn't require much more than using the Minitab output given in the problem).
HW14: hw_lect14_1 - hw_lect14_4.
HW15: 5.47, hw_lect15_1 - hw_lect15_4.
HW16: 7.7(c), 7.11, hw_lect16_1 - hw_lect16_3.
HW17: 7.4(a), hw_lect17_1 - hw_lect17_2.
HW18: 7.42, 7.48(a), hw_lect18_1 - hw_lect18_2.
HW19: hw_lect19_1, 7.58.
HW20: 8.1, hw_lect20_1 - hw_lect20_2.
HW21: 8.14(a), 8.24, hw_lect21_1 - hw_lect21_2.
HW22: hw_lect22_1, 8.30 (use df=welch=29), 8.40, hw_lect22_2.
HW23: 8.45, hw_lect23_1 - hw_lect23_2.
HW24: hw_lect24_1, 9.23, hw_lect24_2.
HW25: 11.13(part a only, but do both CI and p-value), 11.38(a,b,c), hw_lect25_1.
HW 26: hw_lect26_1 - hw_lect26_2.
--------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
Other items of possible interest:
