sampling question
Terence Finnegan (finnegan@ncsa.uiuc.edu)
Fri, 10 Dec 1993 17:40:42 ECT
This was originally posted to H-South. The author is also on H-CivWar,
however, and you may feel free to respond to him on either list.
Robert Alan Harris
bb05196@bingvmb.bitnet
----------------------------Original message----------------------------
I have a student studying women in the antebellum South. She took a sample
from the census for 1860 and 1870 and she discovered that in 1860 the
youngest white female person who had an occupation was about 14 (in 1870 it
was about 13). Then she started through the manuscript census for 1860
counting every single white woman who was 14 or over. At each fifth white
woman, 14 or older, she stopped and looked at the occupation column. If
that fifth white female 14 or older had an occupation, she took down her
age, occupation, and had a code for whether she was "married," "widowed,"
or "single." If the fifth white female 14> did not have an occupation,
however, she did not take down any information from the census, but instead
counted another five white females 14> and if that one had an occupation,
she copied down the information, and if not, counted another 5 etc. She
never recorded the next person with an occupation in her sample; the person
always had to be a fifth white female 14> and have an occupation.
She ended up with a sample of about 200 women. My question is: Do people
think this is a valid sampling technique and if not why?
Vernon Buton
University of Illinois