Econ 378 Part 3 Review

A review of the final part of Statistics for Economists - BYU course ECON 378.

Estimators and their properties, including bias, consistency, and efficiency. Samples, standard errors, confidence intervals, hypothesis tests, and intro to regression with Ordinary Least Squares (OLS).

TIMESTAMPS:
0:00 Spiritual thought
6:54 Fantasy data setup
8:38 True values vs. estimators and their distributions
14:30 Bias in estimators
16:47 Consistency in estimators
21:33 Efficiency in estimators
22:12 Confidence intervals
26:52 Proportions
34:38 Hypothesis tests
57:09 Recap of general construction of CIs and H-tests
59:30 Regression and OLS
1:06:14 Calculating regression coefficients
1:15:20 More on R squared and prediction
1:19:09 Differences in means
1:26:08 Standard errors review
1:36:12 Exam 3b #4-b
1:40:05 Exam 3b #4-a
1:43:37 Exam 3b #4-c (see corrections)
1:49:43 Exam 3b #4-d

CORRECTIONS:
27:36 I label the population distribution as p, but this is incorrect. The population distribution is for the binary random variable (say "S" for shark attack). "p" is the true population mean (comparable to mu). "p hat" is the sample mean (comparable to x bar). We generally call these the true proportion and sample proportion.
28:38 Again, I mean the variance of "shark attack" over n.
1:31:16 It can easily be seen that this is the same as the other formula, as the sample variance of x is simply the squared sum in the left formula put over (n-1); thus, multiplying by n-1 allows you to just plug in sample variance of x.
1:39:35 sxx here means the sum of (x_i - x_bar)^2. Thus it is the same as the denominator in the left formula for the standard error of B1_hat from earlier, also equivalent to the denominator in the left formula.
1:45:28 The correlation coefficients are for the x's and y's of each regression. The two regressions themselves are independent samples, hence no covariance between the estimators.
1:46:38 Again, ignore this comment because of the above.
1:47:12 Again, refer to earlier correction about sxx.
1:59:50 Since our test stat would be normally distributed here, the CV would actually be +/- 1.96.

Видео Econ 378 Part 3 Review канала Doggonomics