Re: st: ST: Calculating first differenced correlations

[Steven Samuels <sjsamuels@gmail.com>]

Nicole White wrote: "I am trying to figure out how to calculate the  
correlation of first differences for two variables, poverty and  
violence rates, for the years 1993 to 2004. "

--
You can do this by hand or by using time series commands; see the  
example below.  I'm not expert in this area, but analyzing log crime  
and log poverty might be worthwhile.

Steve


**************************CODE BEGINS**************************
clear
set obs 12
gen year = _n + 1992
list
set seed 4568
gen crime = uniform()*2500
gen poverty = crime[_n-1] +uniform()

// First by hand
gen d1crime= crime-crime[_n-1]
gen d1pov = poverty-poverty[_n-1]
corr d1pov d1crime

// With Time series operators d. (= d1.)
tsset year

list d1pov d.poverty d1crime d.crime d1.crime
corr d.poverty d.crime
scatter d.poverty d.crime
***************************CODE ENDS***************************


On Dec 7, 2010, at 5:59 PM, White, Nicole (UMSL-Student) wrote:

Hello--
I am trying to figure out how to calculate the correlation of first  
differences for two variables, poverty and violence rates, for the  
years 1993 to 2004.  I assume that I need to identify my dataset as a  
time series, so I entered the tsset command and identified the time  
variable as years, 1993 to 2004.  However, I do not know what command  
or formula to use to generate the first difference for these rates.   
Can anyone offer some guidance?  Thanks
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/