Anders A
03-14-2008, 07:10 AM
Hi!
I am working with crime data, evaluating treatment programs for innates.
Now, I have thos rather large file of about 6000 people having participated in four different treatment programs. Rows are clients and columns are variables in my data editor. I have a SPSS-file where I among others have one row in the SPSS data editor for treatment,three columns for year of inprisonement, 2000 columns about crimes comitted during a 10 year period and about 1500 columns/date-variables. The date variables was extracted to years.
Now I want to be able to figure out the number of relapses for three distinct categories of crimes (DUI, Other Traffic Incidents and Other types of Crime)
in relation to the year of inprisonement. I e. I want to calculate "relative years" (RY) - 2 and 1 year before year of inprisonement and 1 and two years after year of inprisonement. (in that way I can both fugure uyt what subpuopulation (treatment population) that was most "charged" with crimes before inprisonement and what subpopulation (treatment program) that relapsed the most.
I have constructed binary select variables identifying type of crime comitted in a certain year (I only have three trreatment years) by using "IF" and "Count" in combination.( If year "X" and crime "Y" then Selector Z = 1. And I have written syntax to figure out my "relative years". However I get funny results. (IF treatmen 97 AND Selector Y = 1) THEN RY = -1 (i.e two years before treatment).
It is true that some individuals have comitted crimes year after year despite prior inprisonement and it is true that some people stick to certain criminal behaviour year after year. When looking at the end result type and amount of crimes comitted two and one year ebfore inprisonement and one and two years after inprisonement I observe that some clients have comitted EXACTLY the same number of the same crime all years. And that is rediculus - it can not ve true thar one guy i comitting exatlt 36 or 72 (or whatever) DYI crime in five years time.
Is there anyone out there having an idea of how to come around this annoying results ? I will be happy to know - because I have a deadline in two days - and when I say deadline I mean it - DEAD .... ;).
Thanks: Anders A.
I am working with crime data, evaluating treatment programs for innates.
Now, I have thos rather large file of about 6000 people having participated in four different treatment programs. Rows are clients and columns are variables in my data editor. I have a SPSS-file where I among others have one row in the SPSS data editor for treatment,three columns for year of inprisonement, 2000 columns about crimes comitted during a 10 year period and about 1500 columns/date-variables. The date variables was extracted to years.
Now I want to be able to figure out the number of relapses for three distinct categories of crimes (DUI, Other Traffic Incidents and Other types of Crime)
in relation to the year of inprisonement. I e. I want to calculate "relative years" (RY) - 2 and 1 year before year of inprisonement and 1 and two years after year of inprisonement. (in that way I can both fugure uyt what subpuopulation (treatment population) that was most "charged" with crimes before inprisonement and what subpopulation (treatment program) that relapsed the most.
I have constructed binary select variables identifying type of crime comitted in a certain year (I only have three trreatment years) by using "IF" and "Count" in combination.( If year "X" and crime "Y" then Selector Z = 1. And I have written syntax to figure out my "relative years". However I get funny results. (IF treatmen 97 AND Selector Y = 1) THEN RY = -1 (i.e two years before treatment).
It is true that some individuals have comitted crimes year after year despite prior inprisonement and it is true that some people stick to certain criminal behaviour year after year. When looking at the end result type and amount of crimes comitted two and one year ebfore inprisonement and one and two years after inprisonement I observe that some clients have comitted EXACTLY the same number of the same crime all years. And that is rediculus - it can not ve true thar one guy i comitting exatlt 36 or 72 (or whatever) DYI crime in five years time.
Is there anyone out there having an idea of how to come around this annoying results ? I will be happy to know - because I have a deadline in two days - and when I say deadline I mean it - DEAD .... ;).
Thanks: Anders A.