ECN 102 Lecture Notes - Lecture 4: 1880 United States Census, Simple Random Sample, Random Variable
Analysis of Economics Data
Lecture 4
Diana Moreira, UC Davis
April 11, 2018
1 / 47
Last Class
2 / 47
Description of data using graphs.
1.Box plot
2. Histogram:
particular case: smoothed histogram.
0 10 20 30 40 50
Frequency
0 50000 100000 150000 200000
Annual earnings (in dollars)
Earnings: Histogram
0 5.000e-06 .00001 .000015
Density
0 50000 100000 150000 200000
Annual earnings (in dollars)
kernel = epanechnikov, bandwidth = 1.5e+04
Earnings: Smoothed Histogram
3 / 47
Document Summary
1. box plot: histogram: particular case: smoothed histogram. 3 / 47: smoothed histogram density is pulled up if occurrence in side bins are more frequent, positive density even when no observation is within the interval right below the relevant density. Data is often not ready for a meaningful description. For any economic data are often important transformations you have to do in the data to be able to analyze: z-score, natural logarithm. Present the transformation, talk about the properties, and then show when and why it is used. Usages of z-score: symmetry and kurtosis formula can be written in z-scores, interpretation of the magnitude of z-score is easy. 1 unit increase in z equals 1 standard deviation increase in the original score x. If the distribution looks like a normal, it is equivalent from going from the median (50th percentile) up to the 84th percentile. 10 / 47: used to interpret categorical variables.