*Thyne, Clayton
*Intro to Stats, Chapter 2 examples

clear
set mem 80m
cd [your directory]

*  NOTE: the above command asks you to change the directory.  The directory will depend on how you've set up
*  the folders on your computer.  For example, my command is...
*  cd F:\teaching\Fall_2008\PS572-401-baby_stats\do_file_directories\chapter2
*  This is necessary so you know where stata is saving things.  "box_trade.gph" in "Part 4", for example, 
*  will be saved in the directory you specify.

*PART 1: examining frequency using tables and graphs
*replicates Wonnacott & Wonnacott (1990: 27)
	use http://www.uky.edu/~clthyn2/intro_stats/Ch2_example1.dta
	list
	tab days
	hist days, frequency width(1) xtick(1(1)13) title(TABLE 2-1  Absenteeism in the Shipping Department) xtitle(X = Number of days absent) ytitle(Frequency)
	hist days, norm frequency width(1) xtick(1(1)13) title(TABLE 2-1  Absenteeism in the Shipping Department) xtitle(X = Number of days absent) ytitle(Frequency)
	hist days, addlabel norm frequency width(1) xtick(1(1)13) title(TABLE 2-1  Absenteeism in the Shipping Department) xtitle(X = Number of days absent) ytitle(Frequency)
	twoway (kdensity days) (hist days)
	stem days

*PART 2: group into centiles
	centile days, centile(25 50 75)

*PART 3: examine with a boxplot
	graph box days, medtype(line)

*PART 4: examine IR data
	clear
	use "http://www.uky.edu/~clthyn2/intro_stats/data_set_1.dta"
	replace ttrade=. if ttrade<0

	graph box ttrade, saving(box_ttrade, replace)
	gen lttrade=log10(ttrade)
	graph box lttrade, saving(box_lttrade, replace)
	graph combine box_ttrade.gph box_lttrade.gph

	hist ttrade, frequency saving(ttrade.gph, replace)
	hist lttrade, frequency saving(lttrade.gph, replace)
	graph combine ttrade.gph lttrade.gph

	graph box lttrade, over(region)

*PART 5: mean vs. median
	sum ttrade, detail
	**compare 50th percentile to mean
	
	sum lttrade, detail
	**compare 50th percentile to mean

*we can also look at these together with...
	gen exists=1
	table exists, c(n ttrade max ttrade min ttrade mean ttrade median ttrade)

*PART 6: finding descriptive statistics
	clear
	use "http://www.uky.edu/~clthyn2/intro_stats/Ch2_example2.dta"
	sum

*a more complex example with percentiles...
	clear
	use "http://www.uky.edu/~clthyn2/intro_stats/data_set_1.dta"
	sum cap, detail
	return list

*PART 7: examining standard deviation
	clear
	use "http://www.uky.edu/~clthyn2/intro_stats/data_set_1.dta"
	gen lttrade=log10(ttrade)
	hist lttrade, frequency width(.1) xtick(1.5(.1)6)
	hist lttrade, frequency width(.1) xtick(1.5(.1)6) norm

	sum lttrade
	hist lttrade, xline(2.18 3.88 5.96) frequency width(.1) xtick(1.5(.1)6) norm
	di 3.88+.874
	di 3.88-.874
	hist lttrade, xline(2.18 3.88 5.96 4.754 3.006) frequency width(.1) xtick(1.5(.1)6) norm

*PART 8: using relative frequencies
*replicates Table 2-8 in WW page 53
	clear
	use "http://www.uky.edu/~clthyn2/intro_stats/Ch2_example3.dta"
	sum
	expand frequency
	sort height frequency
	sum height
	**compare to values on page 53

*PART 9: assessing normality
	clear
	use "http://www.uky.edu/~clthyn2/intro_stats/data_set_1.dta"
	sum cap, detail
	**Is "cap" normal?
	**does mean=median=mode?
	**skewness stat... should be 0.  If it is +- .8 or less, it is considered normal
	**kurtosis...Normal==3.0
	tab cap

*Part 9 (continued): using figures to look for normality
	quantile cap
	hist cap
	graph box cap

*Part 9 (continued): clearly not normal, so what's the best transformation?
	gladder cap
	qladder cap

**now, transform the variable, and check for normality
	gen lcap=log10(cap)
	sum lcap, detail
	quantile lcap
	hist lcap, normal
	graph box cap