This question regards methodologies used in reporting the 1900 census.

Heather Haveman, a B-school professor, is attempting to analyse the 1900 census data. A research assistant downloaded the raw data, but she is having difficulty understanding how the census number-crunchers took the data from the questions:

http://www.ipums.umn.edu/usa/voliii/90items.html

and forced them into the census categories:

http://fisher.lib.virginia.edu/cgi-local/censusbin/census/cen.pl?year=900

Her specific questions have to do with the data on literacy (she is interested in literacy data for California). There are, for example, two different census categories in the literacy section, both of which are called "No. illiterate foreign-born males 21 years of age and over" but if you click on the two separate headings and add up the data, the two numbers don't match. She is also confused by the "No. illiterate native white males 21 and over" being larger than the "No. native white illiterates 10 years of age and over" since there must be more people over the age of 10 than there are over the age of 21.

In other words, she is looking for info on how the census officials took the data gleaned from the questions and created reports in given categories. I searched CLIO for keywords like census and literacy and methodology and statistics (in varying permutations) but I came up empty.

Does anyone have any suggestions?

Thanks! -- Karen

FROM JANE WEINTROP:

The site I referenced was the ICPSR Data Archive,

http://www.columbia.edu/cu/lweb/eresources/databases/1360887.html

At the Historical Census Browser site they reference it in the statement,

"Download data. The site is not intended as a tool for downloading data

for further research or more involved manipulation. Those who require

this level of analysis should download the original data from ICPSR."

In the end when she asked directly what she was looking for, I did not

really need to go to the ICPSR site. All the info. needed was at the

Census Browser site.

This was a case of a person working with a teritary source (Census Browser

site) which was based on a secondary source (the ICPSR study) which had

been transcribed from a print source (12th Decennial Census volumes). I

wanted to check the ICPSR source only to verify that the numbers in the

study agree with the numbers on the browser site. They do. What had

been abbreviated to the point where they were confusing were the column

headers that the describe the numbers. The information to sort that out

is in the Codebook to the study and the Historic Census Browsers site

links to it directly.

We emailed excel files to each other yesterday so I know she has the right

numbers and I think I gave here the correct explanation of those numbers.

She was off on the road and I haven't received a reply. If she is not

satisfied, then the next step is to look at the print volumes to get the

original presentation for the numbers.

I hope that is all as clear as mud. My advice to folks in Butler who have

reason to recommend the ICPSR data archive is to refer them to EDS (always

use the eds@columbia.edu address).

Jane