Marriage and Literacy % using Ancestry and FamilySearch

Hello everyone,

How accurate are and when calculating from the different populations (are the populations themselves accurate when restricting the search to certain groups?) - the literacy and marriage percentages for specific groups? For instance, I would like to calculate the percentage of literacy among the different groups living in Pennsylvania in every census from 1900-1950. I have done so, just getting some interesting numbers that I would like to make sure are at least approximately accurate.   

Thank you 

  • Interesting exercise but with a lot of assumptions and variables. "Approximately accurate" is not accurate. How good a statistician are you in order to make statements that are statistically significant when you interpret your findings?  A mean plus or minus a standard deviation for starters might be better or some other statistical test.  Per your examples, literacy was not asked on the 1940 or 1950 census. And questions on a specific topic might vary as to the wording between censuses as well.  What are your "different populations" or "groups"?  If you mean racial designations, those definitions could change over time as well as what instructions are given to the enumerators on the subject.  You should check the enumerator handbooks for any question you want to compare over the years.  If you are going to use racial groups, then you will have to confront the different **undercounts** among the groups to get an "accurate" end result.  And over time, the population has aged which might influence some of the parameters asked including marriage (age of couples, incidence of single households over census years, etc.).  Why are you gong to use the two websites you cite?  Are you going to rely on transcriptions or actually use the actual census sheets?  Why not when possible use the voluminous statistics from the census bureau reports for each census?  How much are you going to drill down into the census records?  Are you going to use geographical information systems to plot census results from enumeration districts within high ethnic communities?    .You should spend quite a bit of time in designing such a study, especially if we are talking about a thesis at the Master's or PhD level.  Remember the carpenter's adage..... measure twice, cut once.    Good luck!!