Course schedule
Math 410 meets Monday and Wednesday 12.20-2.15pm.
Expected workload
A full time semester is 12 credits; this means that a 4 credit course is about 33% of a fulltime.
Out of 40 hours, the 33% makes out 13h20m in a week. We meet 4h in class. You should be spending about 9h every week out of class working with our material.
It has been shown in research that “Effective Time on Task” is the most important predictor of how much you learn. I will assign homework to train the most recent material, but I strongly encourage you to work through all the problems in the book.
Topic sequence
Week | 2016 dates | Topic | Chapter | Homework |
---|---|---|---|---|
1 | Aug 29, 31 | Welcome, intro, structure, background | ||
Bias and MSE of estimators | 8.1-2 | 1-5, 8, 16 | ||
2 | Sep 7 | Common estimators, evaluating goodness | 8.3-4 | 28, 29-30, 34-37 |
3 | Sep 12, 14 | Confidence intervals (and large scale) | 8.5-6 | 40-43, 57 (use previous week’s poll data), 63, 64 (use current data), 68. Pick a research paper. |
Sample size | 8.7 | 70-72, 74-75 | ||
4 | Sep 19, 21 | Small sample CI for means | 8.8 | 93-94, read http://www.jstor.org/stable/2331554, verify computations in Section IX |
CI for population variance | 8.9 | 97-99 | ||
Efficiency, Consistency | 9.1-3 | 1, 4, 6, 17-18 | ||
5 | Sep 26, 28 | Sufficiency | 9.4 | 38-39 |
Rao-Blackwell Theorem and MVUE | 9.5 | 57, 64 | ||
The method of moments, moment gen. func. | 9.6 | 70-72 | ||
Maximum likelihood estimators, properties | 9.7-8 | 80, 87, 96 | ||
6 | Oct 5, 6 | Elements of a statistical test | 10.1-2 | 2-3, 5-7 |
Common large-sample tests | 10.3 | 19-20, 25, 33-35 | ||
8 | Oct 17, 19 | Type II probabilities, Z test sample size | 10.4-5 | 37-38, 40-41 |
p-values | 10.6-7 | 50-55 | ||
9 | Oct 24, 26 | Small-sample testing for means | 10.8 | 61-62, 68, 72, 75 |
Variance hypothesis testing | 10.9 | 80-81, 83 | ||
Power of tests, the Neyman-Pearson lemma | 10.10 | 88-89, 93-94, 99-100 | ||
10 | Oct 31, Nov 2 | Likelihood ratio test | 10.11 | 105, 107-108, 111 |
11.1-2 | ||||
11 | Nov 7, 9 | 11.3-5 | 1, 3, 10, 15, 17-18, 20-22, 27-29 | |
12 | Nov 14, 16 | 11.6-9 | 35-36, 38, 42, 51-52, 54-55, 57, 61 | |
13 | Nov 21, 23 | Designing experiments, matched pairs | 12.1-3 | 1-4, 9, 11-12, 17 |
Elementary experiment designs | 12.4-5 | 18-19, 21-25, 27, 28-29, 31-32, 36 | ||
15 | Nov 28, 30 | Categorical data, chi-2, goodness of fit | 14.1-3 | 3-10 |
Contingency tables, r x c-tables | 14.4-5 | 14-16, 20-22, 29-30 | ||
15 | Dec 5,7 | ANOVA | 13.1-5 | 1-3, 11-18. Use of R encouraged. |
16 | Dec 12 | Project feedback | Read your assigned opponent’s report, give feedback | |
Project presentation | Adapt your report to given feedback. Prepare a short oral presentation. |
Report
The source of your grades in this course will be on a written report, evaluating the statistical methods of a research paper. You will pick a paper to review by the third meeting of the course, on September 12.
Your report should
- Comprehensively describe the statistical analysis methods and the data collection methods used by the authors of the paper.
- Critically evaluate the choices made by the authors: data collection designs, analysis methods.
- Recalculation of the statistical analyses of the authors, alternatively a comprehensive explanation of why the research paper does not provide enough information to replicate the analyses.
Each of these tasks is graded on a scale of A/C/D:
- A is awarded for C plus an additional requirement of suggestions and evaluations of alternative choices the authors could have made.
- C is awarded for comprehensive coverage without any formal flaws.
- D is awarded for up to 4 minor flaws:
- language errors to the point of ambiguous text
- arithmetic errors that do not essentially change the qualitative argument.
Final grade, assuming a passing grade both on each task in the report, and on the final exam, is an arithmetic mean of the three tasks, using A=4.0, B=3.0, C=2.0, D=1.0 and rounded:
- Up for decimal part exceeding 0.6.
- Down otherwise.
Final Examination
The final will be in two parts: one sit-down exam focusing on theory and interpretation of statistical measures, and one peer-grading review of the exam.
The final is pass/fail. Failing the final will fail the course. There will be an opportunity in the final that allow you to raise your grade by up to two minor grade steps: for instance, raise from a B to a B+ or an A-.