One resulting mild irritant is that I have a number of "Fact date is earlier than the earliest estimated child birth date for the individual's father" messages that seem to be false positives. So far as I can see at first glance, all the false positives refer to children who are step-children in a second marriage, where the new spouse is somewhat younger.
For instance, Elizabeth Stretch is b1878 to James Stretch (b1854) and Sarah Pickstock (b1860). Elizabeth's 1881 census is being warned as being "...earlier than the earliest estimated child birth date..." Both parents' ages at Childbirth are set in Preferences / Estimates to 16, so there shouldn't be any issue with her birth parents. However, James dies in 1892 and Sarah remarries in 1893 - to William Hulse b abt 1868. Since Elizabeth remains with her mother, I have set her as a step-child in the family for Sarah's 2nd marriage.
I suspect that Elizabeth's 1881 census is being checked against both sets of parents - it'll be OK against her birth parents but her step-father William was only 13 at Elizabeth's 1881 census. I therefore assume that it's this step-father who is being warned about. Does anyone know if this is a correct diagnosis?
Now, I'm actually a bit stuck for a suggestion about how to get around the message. I can tell that she doesn't join that (step-)family until 1893, but I know that the message is also appearing for a totally different child where there isn't a marriage at all for the "step-parent" who just appears out of nowhere, so there's no explicit "joining date". I'm disinclined to suggest not looking at step-parents for these checks because surely in some cases one should look at them if they are part of the family at that time. Any bright ideas?
If I do check each alert and decide that it's a false positive - what's the best idea for saving my "OK it's a false positive" conclusion to carry forward against future runs of the "Show Project Statistics" plug-in? I know I can save the output from a specific run but that would need manual comparison to a future run. Is that the only option? (I get concerned that there are seriously big cans of worms here!
