TOPICS:

1. OUTLIERS and the FALSE DISCOVERY RATE Procedure
2. Time Series Data, Autocorrelation, and the Durbin Watson Test

Coal Mining Data

Variables: 26 observations from 1940-1965

Problem:Andrews and Christenson performed an exonomic analysis of mine saftey for underground bituminous coal mining. Of particular interest to the authors was the impact of the 1952 Federal Coal Mine Safety act on deaths and injuries within these mines. THey hypothesized that the injury rate from accidental causes was a function of three factors: 1) the level of technology, 2) the size of the mines, 3) the regulatory environment. As the data are measured over time, autocorrelation can be a serious problem.

1. FATAL: fatal injuries/million man-hours worked
2. NonFatal: nonfatal injuries/million man-hours worked
3. NFPerm: nonfatal permament disability injuries/million man-hours worked
4. MECHLOAD: Percent output mechanically loaded
5. TOPM: Output per man-hour
6. AvgMen: Average number of men/mine
7. AvgOP: Average output/mine (tons)
8. invMine: 1/number of Mines
9. FedReg: dummy for federal regulation (1 for 1953-1965)
10. War: dummy for WWII (1 for 1940-1944)

Plots/output:

1. Scatter plots
2. Regression outputDurbin Watson
3. Regression for NONFATAL
4. residual plot of NONFATAL vs YEAR
5. Regression for NFPERM
6. residual plot of NFPERM vs YEAR