The current issue of the New England Historical and Genealogical Register, 167(Jan 2013):5, contains a wonderful editorial by Henry B. Hoff and Helen Schatvet Ullman about what to look for in published genealogies. I would say the same holds true for any online genealogies as well.
-
"Look at the chronology. Do the dates make sense? Make sure women are not marrying at 6 or having children after 50. If there are few actual dates available, has the author provided reasonable "say" [about] dates of birth and/or marriage?"
-
"Look at the lineage. Are there many statements without documentation? Is there documentation for the mother's maiden name? Is there evidence to show children belong to their purported parents?"
-
"Look at the reasoning. Is there analysis to discuss possible or probably connections? Or is everything presented as 'fact'?"
-
"Look at the documentation itself. Do the sources cited seem likely to support the information in the text? Are wills and deeds cited, or is the entire publication based on a few secondary sources and a few published primary source records? The latter scenario may be sufficient, but often so much more could be done."