Mistakes and statistics

[Jeremy Wilkes]

Yesterday's newsletter has prompted me to look at some figures relating to my entries. I have only 168 ticks, and deducting this from the total number, kindly supplied by Peter's computer (if President Trump has a resolute desk, I am sure that Peter can have a kind computer), of my entries indicates that 3,452 entries are unticked. I undoubtedly make genealogical errors (and correct them when I find them), but sadly I do not feel inclined to check that many names, even after allowing for sharing of households. I have made contact with 40 people who have matching entries, but one of them does not seem to have a connection. I have tried, between January 2017 and December 2018, to initiate contact with three people who have not responded, and have not sought contact with one, who does not appear to be related (and he or she has not made contact with me). As far as I know my ancestors are British, but my father's paternal line is missing, as I have not entered his likely relations in that line, for want of sufficient evidence. On the basis of Peter's figure of a likely 200 members who are sixth cousins or closer, I should have contact with over 150 (I write "over" because I have included some relations by adoption). I therefore share his frustration with those who join but do not enter all the relations whom they have found in the relevant censuses. I feel a similar irritation with those who do not answer enquiries and those who have taken DNA tests with Ancestry or with 23 and Me but do not put their results on other websites. We are not obliged to do any of these things, but it seems rather pointless to buy an umbrella and then avoid taking it out when it rains.

 

[peter]

Sometimes experienced members like yourself do make mistakes, but it's largely a problem that I find with users who have entered fewer than 100 relatives.

Often they've been members for 10 years or more - but, of course, until they check and correct their entries they're never going to get any matches. A random sample of entries from the 1881 Census suggests that there are upwards of 7000 entries with the wrong references, and that translates to around 500 missed matches (though some will be with the same cousin).

Yet most of those errors could be picked up and corrected in less time than it takes to read my newsletter. In a week or two I'll check how many of the errors in my sample have been fixed - that will provide an estimate of the response rate.

 

[canadianbeth]

I read your newsletter last evening and immediately started checking my entries. I did the 1911 and 1841 ones before shutting down for the night and found one error in the 1911 census and one spelling error. No new ticks from them. I will start on the 1881 entries next.

I looked at my summary; have just 453 entries and have entered pretty much everyone I have found. Each time I think I have found someone new I discover that family is already entered. I counted the number of actual families; there are 42 in the 1911 census, 81 in 1841, 47 in 1881, 1 in the U.S. 1840 and 1 for 1880. None for Canada because I have found no ancestors who emigrated before 1881. (a great many who came afterwards, including my grandparents and my Dad) I have 9 ticks from those families, 3 known cousins and one relative, with whom I have been in touch. Of course, a good part of my Dad's paternal line is missing since I have no name for his father and have the same irritation as Jeremy with those who do not reply to messages.

 

[At home in NZ]

most of those errors could be picked up and corrected in less time than it takes to read my newsletter

Not so, I used FTA to load some of my Ancestors a few months ago, there were over 400 and I was not inclined to check all those so I hadn't done until now.
Checking a whole page takes far longer than reading a newsletter. I am only half way through and finding it quite laborious.

 

[canadianbeth]

I used FTA this afternoon and it added 5 new entries to my ancestors' page. However, two could not be found when I clicked on the little arrow. She was all alone in her entry but was right underneath with her whole family. Everything was the same except the page #. I was not sure just what to do with her so I left it there but will probably delete it. The second one could not be found on Ancestry either and I do not have an ancestor with that birth date. A third was added, also with his family right below him, but it was in the U.S. census and I had made an error when entering the page number, so that is fixed. And now I find yet another listed twice, once on page 5 and again on page 6, with the same list of family names although he is alone in one entry.

However, I did find that I had missed a whole family. I had entered his brother but missed him, which is strange because they are both 2nd great-grandfathers. (their children married - first cousins) so I should have remembered him. And when I checked discovered I missed a sister as well. So they are now entered. And I got a new contact from the brother - bonus. Instead of 9 sets of checkmarks, I now have ten.

 

[peter]

Not so, I used FTA to load some of my Ancestors a few months ago, there were over 400 and I was not inclined to check all those so I hadn't done until now.
Checking a whole page takes far longer than reading a newsletter. I am only half way through and finding it quite laborious.

How big is your page? If you're talking about a screenful of entries that's typically no more than 10 households, so only 10 clicks and 10 visual checks. How long can that possibly take?

Of course, if you use FTA there are likely to be lots of things that need altering - names and ages - and that's the time-consuming part. But checking should be very quick, and if it isn't, you're probably doing something differently (like checking individually rather than by household).

 

[At home in NZ]

How big is your page? If you're talking about a screenful of entries that's typically no more than 10 households, so only 10 clicks and 10 visual checks. How long can that possibly take?

I did say there were about 400 loaded by FTA, that was in addition to some I'd entered a few years ago.
My page is so long I cannot count the number of screens it is taking up because my eyes can't take it!
I was hoping there would be something that tells me how many Ancestors or Households I have entered but I can't find it.
I have spent about 6 hours checking them today and maybe an hour yesterday I think I'm about three quarters done but I'm not doing any more, it's tedious and taking too much time up.

if you use FTA there are likely to be lots of things that need altering - names and ages -

I thought FTA was supposed to be the best way to do it. I haven't even been looking at ages, all I do is click on the arrow, check FMP to see if my household is there. If i don't find it I check for errors.

 

[Bryman]

I was hoping there would be something that tells me how many Ancestors or Households I have entered but I can't find it.

I think that FTA reports the number of additions when they are submitted.
The difficulty encountered by FTA is that most people try to keep 'correct' information in their tree whereas the censuses often record something different, such as name abbreviations and tongue-in-cheek ages. FTA can be very quick but every entry needs to be verified as correct.

 

[Helen7]

However, two could not be found when I clicked on the little arrow. She was all alone in her entry but was right underneath with her whole family. Everything was the same except the page #. I was not sure just what to do with her so I left it there but will probably delete it.

If she was on the next page, she would probably not appear in the FMP list shown. But you can check by clicking on the transcript for the head of household, as discussed in another thread. The reference for head of household is the correct one to enter if the whole household is split across pages (as clearly stated in the instructions on the Add Ancestor and Edit Ancestor pages, as Peter points out).

 

[peter]

I haven't even been looking at ages

I should have said 'birth years', as that's what you see on your My Ancestors page and in the Findmypast search results. But if the name or age information appears in italics I'm afraid you will need to click the person's name (on your My Ancestors page) to check what appears in the first half of the form (which is all that matters so far as LostCousins matching is concerned).

Entries ought to be checked whether you input them manually or using FT Analyzer. But when you input them manually you can check as you go - if you click the arrow after entering the first person in the household you can confirm that the census references and surname match the census, so that if there's a discrepancy you only need to alter one record.

 

[At home in NZ]

But if the name or age information appears in italics

Where does this happen, in My Ancestors or in FMP?

 

[canadianbeth]

If she was on the next page, she would probably not appear in the FMP list shown. But you can check by clicking on the transcript for the head of household,

She was not on the next page, she was on page 4 and the whole family, including her, is on page 14. Her husband is the head of the household. When I click on her alone, on page 4, I get zero results found. I do not know why FTA put her there. Should I just delete her?

 

[Bryman]

I think that Peter is referring to your My Ancestors page where corrected information is shown in italics, ie you have supplied information to correct/enhance the data given on the census form. An example might be that the original census form shows a forename of "Wm" and you have expanded that to be "William". William would then appear on your My Ancestors page but in italics. Any such data submitted by FTA will not match the contents of the census form exactly and hence would not lead to a correct match with an entry submitted by another LC member.

 

[Bryman]

I do not know why FTA put her there. Should I just delete her?

FTA submits an entry based on the information that you have in your tree, assuming that such information matches the data on the census form EXACTLY.

I suggest that you do not 'just delete her' but rather investigate your tree data and compare against the original census form.

FTA can save a lot of time when census data has not been modified by you in your tree. However, any name or age data corrected by you will need to be uncorrected within LC to enable subsequent matches with other LC members.

 

[canadianbeth]

I have two Mary Ann Riches in my tree. One was born in 1845 and I can find no further information about her. The second was born in 1849 and married James Barrett in 1873 (no apparent relation to the Barratt's in my tree) She is in the 1881 census with her husband and family. She was my first cousin 3x removed, and while looking for more information on her than her name, birth date and parents, discovered the marriage and an 1881 census record, which will be added to LC forthwith.

The Mary Ann Riches in the 1881 census was born Mary Ann Moss in 1828 and married William Riches in 1854. She is in the 1881 census as Mary Ann Riches, with her husband and one daughter (600/124/14. Her ancestor number at LC is 25. The Mary Ann Riches who appears alone, is also listed as born in 1828 and her numbers are 600/124/4. Her ancestor number is also 25. She was my 2nd great-grandmother. I found another Mary Ann Riches, also married to another William Riches, but her age is different as are the census numbers.

 

[canadianbeth]

The other "orphan" entry that I have is an 1841 entry for John Clements, born 1806. (658/12/45/10. Zero records with the little arrow. His ancestor number is 46 - how far back is that? I have no one in my tree born on that date. I do have one born 1801; he is listed directly below with his family. (658/12/46/10) He is my 3rd great-grandfather and his daughter, also listed is my 2nd great-grandmother. Their numbers are 20, 21 and 11. (They are listed as Clemonts in the census, but the spelling is incorrect; it should be Clements)

 

[At home in NZ]

I think that FTA reports the number of additions when they are submitted.

It does but that number gets lost in the mists of time. I have found what I wanted, it's on the My Summary page.

I think that Peter is referring to your My Ancestors page where corrected information is shown in italics, ie you have supplied information to correct/enhance the data given on the census form. An example might be that the original census form shows a forename of "Wm" and you have expanded that to be "William". William would then appear on your My Ancestors page but in italics. Any such data submitted by FTA will not match the contents of the census form exactly and hence would not lead to a correct match with an entry submitted by another LC member.

I have been under the impression that the census reference is the only key to matching and I have only been checking to make sure the ref in My Ancestor leads me to the relevant census.

Italicised years/names are those that appear in the fields under the following:
Finally enter any additional information you have. If you enter a corrected surname, or a maiden surname, then the name will be added to the relevant index - but you will not be identified in any way.
Given that, I fail to see how they can be a barrier to matching.

No wonder I am confused.

 

[Tim]

The other "orphan" entry that I have is an 1841 entry for John Clements, born 1806. (658/12/45/10. Zero records with the little arrow. His ancestor number is 46 - how far back is that? I have no one in my tree born on that date. I do have one born 1801; he is listed directly below with his family. (658/12/46/10) He is my 3rd great-grandfather and his daughter, also listed is my 2nd great-grandmother. Their numbers are 20, 21 and 11. (They are listed as Clemonts in the census, but the spelling is incorrect; it should be Clements)

Hi Beth,

FTA also analyses your Notes/Comments in case census data has been added there, is this possible in this case?

 

[Bryman]

I have been under the impression that the census reference is the only key to matching and I have only been checking to make sure the ref in My Ancestor leads me to the relevant census.

Italicised years/names are those that appear in the fields under the following:
Finally enter any additional information you have. If you enter a corrected surname, or a maiden surname, then the name will be added to the relevant index - but you will not be identified in any way.
Given that, I fail to see how they can be a barrier to matching.

No wonder I am confused.

Matching at LC is performed on the following field values on the census form (or transcription if for UK in year 1881) . . .
Census Reference
Surname
Forename(s)
Age (and hence calculated DoB)

Any correction(s) that you may make to Surname, Forename(s) and Age will be displayed in italic form on your My Ancestors page BUT NOT USED FOR MATCHING.

Entries submitted by FTA will use the data from your tree which may contain corrected information. In that case they will not match entries from other members which have not been modified. That is why entries submitted by FTA and shown on your My Ancestors page in italics should be checked and adjusted appropriately.

Please reply if still confused as this is very important.

 

[Bryman]

(They are listed as Clemonts in the census, but the spelling is incorrect; it should be Clements)

In that case, LC needs to match on the census value regardless of any correction that you have made in your tree. If FTA submits an entry from your tree with the corrected name of "Clements" then that will need to be changed to "Clemonts" in LC on your My Ancestors page.