Chapter 11 Analysis

11.1 Logical Fallacies

Read about the [common fallacies](https://medium.com/@pnhoward/12-common-fallacies-used-in-social-research-9713e4d9bf48) in social research. Summary below:

- fallacies of authority
- fallacies of logic
- fallacies of emotion

11.2 Biases

- Sampling bias
    + e.g., 1948 U.S. presidential election (see this [case study](https://www.math.upenn.edu/~deturck/m170/wk4/lecture/case2.html))
    + even very large samples could have sampling biases if sampling methods are poor and unrepresentative of the population (e.g., [1936 Literary Digest Poll](https://www.math.upenn.edu/~deturck/m170/wk4/lecture/case1.html))
- Omitted variable bias
- Nonresponse bias
- Selection bias
- Survivorship bias
    + e.g., when bankrupt companies are removed from a stock index and replaced with profitable companies, the index would experience an upward bias. Business failures would not be accounted for in time series data.
- Recall bias

11.3 Model Selection

  • Model Selection in R

  • Linear Model Selection

  • Forward/Backward Selection
  • AIC/BIC
  • Nested F-tests
    • model comparisons for linear regressions
  • Likelihood ratio tests
    • model comparisons for generalized linear models

  • Parsimony