rnormal()#

Twoway plot of grades from 2021/2022

You can reproduce this plot

#delimit ;
twoway 
    scatter  
	     final degree , 
	         col(blue) 
			 mcolor(%20) 
			 jitter(3) || 
	rcap 
	     y1lb y1ub degree, 
		     col(blue) ||   
	dot  
	     y1m degree, 
		 col(blue) 
		 msize(2)  
			legend(off ) 
			xlab(1 "MHS" 
			     2 "MSPH" 
				 3 "MPH" 
				 4 "PhD" 
				 5 "SCM") 
		yline(90, lcol(red) lp(dash)) 
		xti(" ") ;
#delimit cr

See also

Earn up to 1.5 bonus points for rendering this in Python, with interactive graphics!

De-identified grades

Conclusion?

The odds of an A is 1/4. Basically, an A in this class is nothing to write home about. You’d likely lose money with these odds because the market place expects an A from you. Our teaching team will do its best to keep the odds this way or even make them shorter :)

See also

The following code introduces a random disturbance to the data in the screenshot above: grade_34060071_general.dta. Use the outputed data file to reproduce the plot of 2021/2022 grades. Up to 1.5 bonus points

#

#delimit cr
quietly {
	
	if 1 {
		
		//reverse engineering?
		//set seed 340600
		
		global filename grade_34060071_general
		global vars final y1m y1lb y1ub
			
	}
	
	if 2 {
		
		use ${filename}.dta, clear 		
		
	}
	
	if 3 {
		
        g disturbance=rnormal(0,3)	
		sort disturbance //further de-identifying..
		 
	}
	
	if 4 { 
		
		forvalues i=1/4 {
			
			local vars: di word("$vars",`i')
			replace `vars' = `vars' + disturbance 
			
		}
		
		
	}
	
	if 5 {
		
		drop disturbance 
		export delimited using "${filename}.txt", replace
		
	}

}