Mini Research Project

[JJohnson]

Hi, I am a adult student in a accelerated degree completion program. I am wondering if I could get some basic help on starting a mini research project. I was asked to get local stats for battered women for the last five years. The only other instructions for the assignment was to compare that information to the national average/stats.

Do I understand that I need to get the mean, mode, median, Standard deviation for my five years and then take the same for the entire data that I gathered and then somehow do the same for the national stats?

I think I am stuck because my local stats were raw data on actual persons and most of the national stats are percentages. They just don't seem to be comparable data.

JJohnson

 

[JohnM]

Basically, what is/are the research question(s)? What is the objective of the research?

Once we answer that, we'll know how to do the comparison / what kind of data we'll need.

 

[JJohnson]

I think the question is: Is the sample of data on domestic violence against females in the local Pennsylvania area representative of the national findings on female domestic violence victims?
Because there is so much data in various areas, I chose to look at three in particular. They are number of women victims seen by the local agency, of those how many are cauc., how many are black and how many are hispanic. I also am looking at how many of those women are under 18 or over 18.

JJohnson

 

[JohnM]

OK, and the national percentages tell you these types of overall break-downs?

Do you want to compare local percentages against national percentages?

 

[JJohnson]

That is where I am confused about comparing the data. I have actual people numbers from my local sources and I have numbers in the thousands for the national averages. For instance, for my first fiscal year of comparison its this:
FY2005
Local Actual

Total Women 419
Race
White 291
Black 18
Other 12
Ethnicity
Hispanic 21
Non-Hispanic
Age
Under 18 177
Over 18 242

Then for my national average it is this (per 1000 persons)

National Per 1000


Race
White 42.40
Black 55.90
Other 35.60
Ethnicity
Hispanic 52.40
Non-Hispanic 43.00
Age
Under 18 165.20
Over 18 185.90

 

[JohnM]

Do you have a good estimate of the local population count?

Then you could transform the local counts into "people per 1000" and do a direct comparison with the national numbers.

 

[JJohnson]

Yes, my local population is 213,674. How do I convert my numbers?

 

[JohnM]

I'm thinking, for example:

White 291

convert to: 291 / 213674 = 0.00136 = 1.36 per 1000

Black 18

convert to: 18 / 213674 = 0.000084 = 0.08 per 1000

The local numbers / percentages seem to be much lower than the national numbers.

 

[JJohnson]

I see. You are right, the local numbers are very small. So, do I understand that once I convert the local numbers that I would then calculate the mean, mode, median, STDEV and STDEVP for my numbers as well as the national?

I apologize but I really don't have a good understanding of how you compare these things - even though I do understand how to make the calculations. I seem to have trouble applying them.

 

[JohnM]

Yes, the logical first step would be to do a quick comparison of summary statistics, but all you have are percentages (expressed as "number per thousand").

"number per thousand" is easier to read and optically compare than reporting low percentages like 0.0002%

I would simply make a two-column table - one column for local, one column for national, and list the number per thousand for each category.

The differences should be so large that they're obvious, and don't require a sophisticated statistical comparison.

 

[jerryb]

i was going to post this message here:

Another interesting thing i notice, and that you might want to report, is that it seems either: white women are much more likely to be abused in your area or that black women who are abused are much less likely to seek assistance in your area.

this is dramitically shown in your data, so presuming that your data is valid for your area (it includes all agencies, neighborhoods etc.) then i suspect the human srvices people need a wake up call to the situation because the clearly are not servicing the black community adequately in that regard.

until i reread your posting, i have to wonder if you have a demograhic profile for your area? perhaps you live in a area similar to where i live that has a very small proportion of black persons. tha other thing i wonder is if the national number is for 42.4 out of 1000 white women reported and 52.5 out of 1000 black women, or is it 42.4 white women out of 1000 total women? that will make a significant difference in how you move forward.

cheers
jerry