Keying Tool Update – Version is now live!

In this version of the tool we have made a few very noticeable changes that we hope you are as excited about as we are.

The first change you’ll notice is the addition of your accuracy rating!  We have received a lot of feedback from you asking how you are doing and wondering if your keying is accurate.  The number you see is calculated based on the number of image sets you’ve keyed that have been arbitrated so as your accuracy fluctuates this indicates that more of your image sets have been arbitrated.  Don’t be discouraged if your accuracy rating is lower than expected.  As we continue to improve the tool and our instructions, based largely on your feedback, we expect to see everyone’s ratings improve.  We’re in this together, and we’re all getting better with practice. 

Another change you will likely notice is when you open the Download Image Sets screen you will now see a question, “Not finding the project you are looking for?” followed by a “Learn why.” link.  This link leads to an article that outlines the process each project follows and provides an understanding for why you see projects appear and disappear from the listing.   

And I know that a few people have been waiting to hear that we have added guides as an alternative to following along with the highlighted fields.  You can now add both vertical and horizontal guides to help mark your place on an image.

Additional enhancements were made to the arbitration process, error tracking, login process, and the international language tools.

We appreciate your feedback and assistance as we continue to work through the Beta testing and make improvements.  If there is something you think we could do to make your keying experience better please click on the “Provide Feedbacklink in the top right hand corner of the tool. 

Information and Links

Join the fray by commenting, tracking what others have to say, or linking to it from your blog.

Other Posts

Write a Comment

Take a moment to comment and tell us what you think. Some basic HTML is allowed for formatting.

Reader Comments

I am keying the NYC Naturalization records. With the new update, there is now a slight hesitation as I move from ‘page’ to ‘page’ that wasn’t there before.
More importantly, when I click on the Surname Box to enter the name, the highlight is no longer on that name. I have to scroll up to read it. All the other boxes are fine. I hope this glitch can be repaired – it really interupts the flow.

Sorry – I explained that problem wrong. The Highlight does go to the surname, but the card does not ‘drop down’ in the screen, causing me to have to scroll to read the name.

I have just seen my accuracy rating of 15% after over 3000 records keyed – if this is a true reflection of my accuracy (85% wrongly keyed? I don’t think so!!!), I am tempted to pack the whole thing in! If it reflects only the records that have been arbitrated against the total keyed then you’ve got things wrong – afechter says ‘don’t be discouraged if your accuracy rating is lower than expected’, well I’m telling you that many, many keyers will be discouraged if these are the sort of accuracy figures they’re getting. You may have a revolution on your hands!

Thank you for the accuracy tool it does at least give us some feedback on how we are doing, but going on from Joes comments would it be an improvement to include the number of how many records have been arbitrated? Has people are entering a lot of records their records keyed total & those records arbitrated could be quite different couldn’t they? what do others think?

How do I know how accurate I am?

If you rate accuracy as the number of records that go to arbitration shoudn’t you subtract those records that are correct? Only those records that are keyed wrong should be used to rate accuracy.

To Jeanette
Your accuracy figure is given under your ‘My Stats’ – if you have had no arbitration yet on any of your keyed records, there will be no accuracy figure.
To Tom & Emery
You do not say what your accuracy rating is – is it so much better than my 15% that you don’t like to say, or is it so much worse and ditto?

I have no accuracy rating as yet but anticipate a very low number. I have been keying the “England, Newspaper Index Cards (Andrews)” and have found that the information, which is sometimes complex, can be keyed in several different ways. Although I have done my best, another contributer may key this information differently but just as comprehensively, if that makes sense. I would not wish to be the arbitrator assigned to this particular project. I feel many of the records will need review. However, I find the project interesting and will not be put off by a low accuracy score.

Type your comment here.

I have no accuracy shown after 1500 plus entrys. Maybe I have 100% NOT

I have around 18000 keyed with a 88% accuracy. I don’t think we should worry too much about the accuracy as when you get a census line with the first letter and then a bunch of scribbled loops, I don’t see how anyone can have a better idea what it actually says. Also depends on the records being keyed. For example, the notes say the Nebraska census is easier than the UK criminal registers. For me, it’s the opposite. I have a tough time decrypting the handwriting on the census and for me the UK registers are easy to read.

I had to submit a completed image set before getting my stats. Didn’t appear until then.

My accuracy is in the 90′s on just under 20,000 records – a real surprise but like the rest of you, it would be extremely helpful to know what we are doing wrong. After all – everyone wants to do the best job they can. If someone has a low accuracy it may be a single very easy error to correct so don’t give up.

I understand that if you pick the the wrong form type it goes to arbitration so there may not be anything wrong with the actual keying. I key on the all the projects except the new Italian one -not quite that confident.

I would also recomend you take an occasional glance at the arbitrators message board where some common errors are discussed.

So, from what I read…

Accuracy rating may be 0 or higher if you and the 2nd keyer agree and you have had few records sent to the Arbitrator.

Which may also logically mean that a high Accuracy Rating may mean more of your records were sent to an artibrator, AND you were found to be keying accurately to that percentage.

Is that right?

And so, the accuracy rating is not a percent of all the records you keyed, but just the number they reviewed. Is that right?

I agree that it would help us improve if the Arbitrators would create a list of common errors.

Also, where is the arbitrator message board?


In my opinion, this accuracy “grade” is in poor taste. We are volunteers, not school children. In NO way does this help anyone, because there is nothing showing you what you are doing wrong. You cant fix, what you dont know is wrong.

I am new here, but I have been indexing/arbitrating for some other place. They shouldnt be judging this by what the arbitrators choose, because even THEY are not always correct.

I work on foreign records, and I am not even fluent in the language, I just have years of experience in viewing and deciphering these type records…so of which “nobody” could translate, have to take “your” best guess, which might not be someone elses.

It is also my opinion, that when you choose a record to work on, you take that record out of the “pile”, you do your best, and the record now moves on to someone else. Had you not taken that record, it would still be sitting there. If people become discouraged at their accuracy “score”, they wont be taking anymore records.

If that score must show, then they should add an on/off switch to those of us who dont wish to see it.

I have no rating even though i keyed 5200 records. Is this normal? I know others don’t either, but is this not unusual?


I think the accuracy is based on

1. Only those of your records that are arbitrated.
2. Of those on the number of your fields that the arbitrator accepts as correct.

So theoretically, Someone could have keyed say 100 records, 10 of which are arbitrated and of those say 5 records are not accepted. This would give a 50% accuracy rating but this does not reflect the 90 records which have not been arbitrated. In fact the total accuracy would be 95% not 50% – in other words – statistics can say anything and can’t be trusted.

The arbitrators message board is with the other new project specific message boards – either use the link on the keying tool message dated 16 Dec or click into a message on the old main message board on the dashboard. Look above it and the pathway will say

You are here: Message Boards > World Archives Project > World Archives Project

Click on the 1st Word Archives Project written above – and it will take you into the new message boards.


How I agree with Jan (comment 14) – I was going merrily along thinking I was doing a good job when up pops this ‘grade’ of 15% accuracy! Not even a proper explanation of how the figure is calculated and, as yet, no response from afechter. I am seriously considering packing it in unless something is forthcoming from the powers that be – surely the technology is there to give direct feedback to the keyers from the arbitrators. And who says the arbitrators are always right?
Afechter – are you there? Do you actually read the comments after you have posted an article? Some response from someone, please.

I’m trying to access the “how can I improve” page, it won’t accept my email address

Stop! Stop! Stop! — Please. Do not worry so much about your ratings. Joe, keep keying merrily along. Everything you key is one more record for a researcher to be able to find. I don’t care if every one of your entries has to go to arbitration – you’re still performing a valuable service. If you’re filling in all the empty boxes to the best of your ability, and you’re scanning yourself for those unintended errors, you are doing a great job. It’s the arbitrators job to keep the entries as error-free as possible. Let them worry about it.
Keep keying – keep keying – keep keying, everyone!

Thanks for the encouragement, Barbara, but I still feel that the powers that be have shot themselves in the foot with this one – I still would like to hear from afechter (or even a lesser mortal) to answer these questions:
Does the arbitrator have the name of the keyer for the records they are checking?
If not, why not?
If so, why can’t they make contact with them to tell them where they are going wrong?
Is the arbitrator the ultimate authority in these matters – if there is a dispute about translating a name, etc., is the arbitrator always right?

I will try to address all of the comments in one post but chances are I will miss something. I key and arbitrate myself so I know they are both difficult tasks!

We decided to include the accuracy rating based on feedback we received from a large number of contributors who wanted an indicator of how well they were doing. Unfortunately with thousands of keyers it is difficult to give individual feedback on how to improve your accuracy. The accuracy rating you see is currently based on only the image sets you’ve keyed that have arbitrated in the past 90 days.

The accuracy ratings are based on the number of records you have correctly keyed compared to the number of records that should have been keyed. This makes it important to fill in all of the fields available and to make sure you are completing the entire image set. If you leave fields blank that another keyer fills in this will negatively impact your accuracy rating.

We are working to improve our help articles and Field Helps in the tool so that everyone has a good understanding of how to key the records. I know that there are certain projects, the postcards and the England Newspaper Collection, where it is a little more difficult to discern what should be keyed. Going to the help articles is a good place to learn more about how these records should be keyed. The article for the newspaper clippings can be found here, Next week we will be adding more project specific help articles.

If you have additional questions please post them here or send an email to worldarchivesproject

Working on the Nebraska Census and the Keying Tool has crashed twice with this error message:
at System.Windows.Forms.ToolStripButton.OnClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleClick(EventArgs e)
at System.Windows.Forms.ToolStripItem.HandleMouseUp(MouseEventArgs e)
at System.Windows.Forms.ToolStripItem.FireEventInteractive(EventArgs e, ToolStripItemEventType met)
at System.Windows.Forms.ToolStripItem.FireEvent(EventArgs e, ToolStripItemEventType met)
at System.Windows.Forms.ToolStrip.OnMouseUp(MouseEventArgs mea)
at System.Windows.Forms.Control.WmMouseUp(Message& m, MouseButtons button, Int32 clicks)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.ScrollableControl.WndProc(Message& m)
at System.Windows.Forms.ToolStrip.WndProc(Message& m)
at System.Windows.Forms.Control.ControlNativeWindow.OnMessage(Message& m)
at System.Windows.Forms.Control.ControlNativeWindow.WndProc(Message& m)
at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)

Since the update, the tool jumps when a new image enters the viewing area; it also drags instead of a smooth flow between keying fields. If the keyer types fast, the tool hangs at the previous field, then “jumps” and returns to normal.

This is to Joe who is upset because his accuracy rating is 15% and cannot get any response from the project. I agree that he should be upset because obviously he has put a lot of time and effort into the project and no one has told him what he has been doing wrong!! With the sparse instructions that came with the request for volunteers the project is getting just what it asked for – volunteers trying to do their thing by making up rules on their own which they feel fit the situation. Then later someone else figures out something different and the work does not match and has to be done over. For example, did Joe know that he was to key certain lines in a set and not leave them blank, or skip around the data and enter what he could find a place for it making arbitration impossible? I for one worked on hundreds of sets of images keying lots of information only to discover by accident that I was actually skipping the next view because I thought I was finished? That made two out of three sets I worked on that was not finished and someone else had to key it all over again. Or just selecting the wrong type of form for entry – one little decision – can throw all your work out the window and you would never know what you did wrong. The way I look at it, you can never be charged with an error if you failed to get proper instruction to begin with. As an arbiter I have had to reject many sets because the keyer had made an error that threw all the data off, but I had no choice.

Which is exactly why there shouldnt be a score card.

I find this a definite turn off.

I index and arbitrate records, elsewhere. I see the innocent mistakes of many people who are trying to help. These are people volunteering their time, when they could be doing something else. A score card makes their effort seem not to be appreciated.

The records go thru several people, before they are put out there. There are bound to be many mistakes made along the way.

An arbitrator is no more “right” than anyone else. They are doing the exact same thing as keyers..entering things, the way THEY see it…and that doesnt mean that they are seeing it right…just possibly differently.

I dont like the idea of this accuracy score. I think its in very poor taste. I think a better approach would be for people to be able to ask a question and get an answer. Maybe then, it would cut down on the ammount of errors.

I think we are the (formerly) Silent Majority. I don’t blame them for putting up this (silly, in my humble opinion) accuracy percentage. I have seen many posts with people worried about their ‘score’ – and the site simply responded to those requests.
Now, they’ll hear from the rest of us.
I do agree with the multitude of requests for clear Blank by Blank instructions for each type of record. If these instructions were posted at the same time as the images, it would make the majority of our problems go away.
Having someone post the instructions has to be easier then enacting these types of updates, no?

First time keying since the Keying Tool Update and it has crashed 3 times. I am working on the California Naturalization Indexes and get this error message after keying in the first record. “Object reference not set to an instances of an object.” I tried 3 different sets and this happened each time. How can this be fixed?

A wild guess.

Some of these people who wanted to see their accuracy, will be very unhappy, once they see their score.

Everybody makes mistakes. We think once we are finished and have checked over our work, it is 100% correct.

And it is…according to how our eyes are seeing it. But, many times, someone else doesnt see it quite the same.

I have been using various record indexes for over 16 years. I cant say, any one set of indexes is any better than another. Even when the work passes through several people, there ARE mistakes. This is why it sometimes takes great effort to find your ancestor’s records for ship passenger lists, censuses, etc. No matter how many people view a record, some just wont make it out there correctly.

Anyway, if they want better results, then better information and assistance should be provided. And those looking for a score..have it already. You get 100% for giving up your free time to help with a project like this.

I too was dissapointed to have a score of 70% but acknowledge at the beginning I was making mistakes got to the end of one record and realised there were ages in a different area to previous records and I had plain missed it.

There is a BUT:- What does bother me are the suggested names. We should be keying what we see, there is a temptation to chose the nearest and I don’t like that. If every record has to go for arbitration then so be it but lets have accurancy
I have said this before so hope I an not boring. but according to Ancestry my Grandmother didn’t exist someone with a different name was born on the same day to that family, Though a friend who has seen the original record says her name is quite plain. So how come two people read Emma instead of Annie? or was that suggestion?
It hurts hurts hurts
Have a lovely Christmas and Carry on Keying

We were asked to participate in a BETA program keying records to help other people find information on their families. The program was offered (to me) in August. As with any BETA program, the problems are not realized until those testing report them. That is called “feedback” and it is constructive.

As for the rating, why let it bother you? Okay, I did get a higher rating than 15% but perhaps more of my keying has been arbitrated. In the beginning, it was a bit confusing but became clearer with practice and in checking the boards or emailing to ask for help.

If a field is required, something must be entered into the field. You cannot merely leave it blank, you must mark it blank. Not doing so, may decrease your rating. Although it is great to see a high accuracy rating, in the grand scheme of things it doesn’t really matter. Three people are keying from the same data. Majority rules.

As for not being instructed as to what to do, I think if people take the “time” to read, they will find more instructions than realized. For example, this was posted on the blog for The New England Naturalization.

“The New England Naturalization Indexes is a series of index cards, there are three on each image, that provide excellent vital information. Since people were coming from all over the world you get to read a lot of unique names.”

Had the person who stated that he submitted one card and didn’t realize there were more records, had read the article, his error would not have occurred.

We all make mistakes and I know I have made many, especially when a new project is released. It’s okay. It’s forgiven. If you are in this for self gratification or monetary rewards, then perhaps you should look for a different pastime.

This project is a way of preserving history for future generations. Whether you intend to keep on keying or hang it up it your decision. Whichever you choose, please just stop complaining and do it.

I got NO score as of yet, as my work wasnt arbitrated.

I just dont think this is the way to go for a volunteer program, no matter what type work is involved. I am sure hospitals dont score their candy stripers.

As far as people making comments against this score card, if you read above the posting states that they really do appreciate the feedback. This means both the for and against. They already heard from those who suggested this. Now, its time for them to hear from the other side. I also didnt see anyone say they were about to “hang it up”.

Mine crashes too since the ‘update’ gives me this box says I must quit, so I quit, it DOES save my last record, but then when I re load the software I am completely locked out and have to end it with ‘TASK MANAGER’ very frustrating to say the least.

System.NullReferenceException: Object reference not set to an instance of an object.
at EnterData.DataEntryPane.ឋ(DataGridViewCell ᙂ)
at EnterData.DataEntryPane.ᝋ(Object ᙂ, DataGridViewCellEventArgs )
at System.Windows.Forms.DataGridView.OnCellValueChanged(DataGridViewCellEventArgs e)
at AdvancedControls.AdvancedDataGridView.OnCellValueChanged(DataGridViewCellEventArgs e)
at System.Windows.Forms.DataGridView.OnCellValueChangedInternal(DataGridViewCellEventArgs e)
at System.Windows.Forms.DataGridViewCell.SetValue(Int32 rowIndex, Object value)
at System.Windows.Forms.DataGridViewCell.set_Value(Object value)
at EnterData.DataEntryPane.ᝉ(Object ᙂ, EventArgs )
at System.Windows.Forms.DataGridView.OnCurrentCellChanged(EventArgs e)
at AdvancedControls.AdvancedDataGridView.OnCurrentCellChanged(EventArgs e)
at System.Windows.Forms.DataGridView.SetCurrentCellAddressCore(Int32 columnIndex, Int32 rowIndex, Boolean setAnchorCellAddress, Boolean validateCurrentCell, Boolean throughMouseClick)
at System.Windows.Forms.DataGridView.SetAndSelectCurrentCellAddress(Int32 columnIndex, Int32 rowIndex, Boolean setAnchorCellAddress, Boolean validateCurrentCell, Boolean throughMouseClick, Boolean clearSelection, Boolean forceCurrentCellSelection)
at System.Windows.Forms.DataGridView.MakeFirstDisplayedCellCurrentCell(Boolean includeNewRow)
at System.Windows.Forms.DataGridView.OnRowCollectionChanged_PostNotification(Boolean recreateNewRow, Boolean allowSettingCurrentCell, CollectionChangeAction cca, DataGridViewRow dataGridViewRow, Int32 rowIndex)
at System.Windows.Forms.DataGridViewRowCollection.OnCollectionChanged_PostNotification(CollectionChangeAction cca, Int32 rowIndex, Int32 rowCount, DataGridViewRow dataGridViewRow, Boolean changeIsDeletion, Boolean changeIsInsertion, Boolean recreateNewRow, Point newCurrentCell)
at System.Windows.Forms.DataGridViewRowCollection.OnCollectionChanged(CollectionChangeEventArgs e, Int32 rowIndex, Int32 rowCount)
at System.Windows.Forms.DataGridViewRowCollection.AddInternal(Boolean newRow, Object[] values)
at System.Windows.Forms.DataGridViewRowCollection.Add()
at EnterData.DataEntryPane.ᝫ()
at EnterData.DataEntryPane.ᝩ()
at EnterData.DataEntryPane.LoadData()
at EnterData.EnterFlexDataControl.LoadForm(Boolean& bCancel)
at KeyingTool.MainForm.ᙉ(PreviousNextControl ᙂ)
at KeyingTool.MainForm.ᙈ(Boolean ᙂ)
at KeyingTool.MainForm.ᙕ(Object ᙂ, EventArgs )
at KeyingToolUtilities.PreviousNextControl.RaiseGotoNextStep()
at SetFormPosition.SetFormPositionControl.ᙍ(Object ᙂ, EventArgs )
at System.Windows.Forms.Control.OnClick(EventArgs e)
at System.Windows.Forms.Button.OnClick(EventArgs e)
at System.Windows.Forms.Button.OnMouseUp(MouseEventArgs mevent)
at System.Windows.Forms.Control.WmMouseUp(Message& m, MouseButtons button, Int32 clicks)
at System.Windows.Forms.Control.WndProc(Message& m)
at System.Windows.Forms.ButtonBase.WndProc(Message& m)
at System.Windows.Forms.Button.WndProc(Message& m)
at System.Windows.Forms.Control.ControlNativeWindow.OnMessage(Message& m)
at System.Windows.Forms.Control.ControlNativeWindow.WndProc(Message& m)
at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)

Is it possible to add a field to key in middle names? This would be especially helpful when more than one person has the same given name.

A final word from me:
I have received an e-mail from afechter which has convinced me to carry on keying – apparently the 15% rating was on the first 6 images that I keyed, when I had little idea of what I was doing! Obviously my keying has much improved since then and my rating should rise accordingly as more records are arbitrated.
If this information had accompanied the rating, there would have been no problem, i.e. “your rating of 15% is based on the first 6 images you keyed” or words to that effect. If this information is available (afechter found it!) why can’t it be transmitted to the keyers?
And just a word to Joan Scribner (comment 30):
I resent very much your comment “… please just stop complaining and do it”. If it wasn’t for people complaining and pointing out errors in procedures/processes etc., then beta testing would be a waste of time! The whole idea of beta testing is to find the bugs and eradicate them so that future generations of keyers will have it easier – if it was left to the sheep (like you) we would be stumbling along for years to come, making the same mistakes over and over again.

I apologize if I caused anyone resentment. That was not my intent. It’s just the way many of us state our feedback comes across as a complaint rather than a constructive suggestion/comment to make things run more smoothly. I guess the quote of “hang it up” should have been “I am tempted to pack the whole thing in!” Sorry.

Cindy, thanks for reporting the crashing since the update. I have had the same problem with the same message but notice that it only occurs on the Nebraska census. Have been doing the N.E. Naturalization and have not had to TASK MGR out of the program.

When keying the N.E. Naturalization and a woman’s Given Name is obviously her maiden name because it appears in the Alias Surname field as well, should both names be entered as Given Name? This causes the field to show as a “review” field.

I’ve also been having the same problem with the Nebraska census. I’ve found that for me it occurs when it goes from the 1st page to the 2nd page. I have to log off then back on to do the 2nd page and can then load another set to key and do the whole process again.
I’m also finding the rotation on the entries more difficult to follow as the page scrolls much further down the page that prior and by the time I’m to about the last 6/7 entries I have to scroll up to see the entry and by the time I do the 50th on the page it’s scrolled back to the top of the page.
I don’t have a rating but perhaps a general note of what the common errors are with each project and how to avoid them might be appropriate.
Spelling errors are probably going to be very common especially where more foreign names are involved which could account for accuarcy ratings.

I just finally got the software to download and install on my computer. I then downloaded my first batch to key and went to open it. At that point, I got an error that prevented me from opening the batch to view and key. How can I resolve this issue?

This is the text of the error I encountered:

System.BadImageFormatException: Could not load file or assembly ‘IULight, Version=, Culture=neutral, PublicKeyToken=0b51e950028a20a6′ or one of its dependencies. An attempt was made to load a program with an incorrect format.
File name: ‘IULight, Version=, Culture=neutral, PublicKeyToken=0b51e950028a20a6′
at TgnImaging.ImageOps.GetImageUsingLeadTools(String fullName)
at TgnImaging.ImageOps.GetBitmap(FileInfo imageFile)
at TgnImaging.ImageViewerControl.set_ImageFile(FileInfo value)
at KeyingTool.MainForm.ᙀ()
at KeyingTool.MainForm.StartNewBatch()
at SelectBatch.SelectBatchControl.ᙆ(Object ᙂ, EventArgs )
at SelectBatch.BatchList.ᙆ()
at SelectBatch.BatchList.ᙡ(Object ᙂ, EventArgs )

WRN: Assembly binding logging is turned OFF.
To enable assembly bind failure logging, set the registry value [HKLM\Software\Microsoft\Fusion!EnableLog] (DWORD) to 1.
Note: There is some performance penalty associated with assembly bind failure logging.
To turn this feature off, remove the registry value [HKLM\Software\Microsoft\Fusion!EnableLog].

I find the percentage rating rather offputting and I feel that it could be counterproductive. I, for instance, have done over 2000 and only about 300 are correct. Is it worth my time and effort doing it.
I appreciate that my percentage will go up when they have all been arbitrated but I would have thought that it would have been better to give a percentage of only those that have been arbitrated. Your figures tell us nothing.

Thanks for the accuracy tool..
My accuracy rating is 82% of those going to arbitation. Is that poor, good, or excellent?
Could you explain what happens in the arbitration process, and is there a way to see which records were arbitrated, and what the decision was on records we have keyed?

I find that the updated tool takes a lot longer to move from field to field. This is a bit off putting for fast typists…. (Whoops, where’s the cursor!)

I’ve been keying for 35 days, but still have no Accuracy calculation. I also notice that although I key 10 records at a time in the NYC Naturalization project, I get credit for 19 or 20 records. Obviously I must be doing something wrong in my data entry, but I have no clue what. I find this very discouraging and have decided to quit. I’ve got better things to do with my time then enter useless data.

I would like to add some questions and comments on the accuracy rating. First, it would be helpful if you included the number of records the rating is based on. Even though someone has entered 1,000 records, if they get a low score because it is based only on 20 records, they are going to feel upset unless they know what it is based on.

Second, there are two different definitions on how the rating is calculated. Clicking on “How is accuracy calculated” says”.. The number of your fields that are accepted determines your accuracy rating.” However, if I click on the link “How can I improve?”, the answer includes the paragraph “How Accuracy is Calculated.” This states “Your accuracy rate is based on the percentage of correctly keyed image sets you have indexed.” So, which is it based on? An image set (which contains many records .. if one entry in one record is wrong is the entire set thrown out?), or is it really based on the number of fields or records?

Here is a suggestion for improving your accuracy. After keying in records of several different types, I would suggest you find one type you like and download 5 or more image sets at a time. Why? Because multiple sets download at one time or most likely to have been written by one person and it allows you to better learn that persons hand writing. In addition, it can be benificial to stick to one project rather than jumping around a lot. This allows you to “learn” that project. Also, don’t be afraid to use the “Provide Feedback” tool to get answers about things you aren’t sure about.

Finally, a comment about the accuracy and its meaning. There have been a lot of posts on this, some pro and some con. In the introduction to this project, we are encouraged to enter data and not worry about making mistakes. We are encourage to make our best guess rather than leave something blank. However, if a lot of my data is thrown out, I am not going to be compelled to continue. Thus, along with the rating, the number of records it is based on should be given. It would, indeed, be helpful to be able to see where each of us is making the most mistakes. I think everyone will see their accuracy improve dramatically as they complete 10,000 or more records. The more you do, the better you will get. If you enjoy doing this kind of thing, stick with it and don’t worry about your rating so much. If it is low, ask specific questions through “Provide Feedback.” The answers will help you improve! Finally, if you are overly worried, dissapointed, or upset over your score, perhaps you shouldn’t be doing this. It is not healthy to obsess about something like this. You should only do those things that give you satisfaction and/or pleasure!

It’s clear that everyone particiipating in this effort supports the concept and wants desparately to contribute the effort. We all recognize how it could benefit the community at large and ourselves individually. But I think we are all struggling with the implementation. Currently it’s a “black box” that seems to produce unexpected results. We are all trying to figure it out, but so far there is little sense to it. And so far the folks at Ancestry seem to be slow in explaining it. I just hope they can resolve this issue before too many from this community give up on it. Because to lose the interest of the population that most wants to help would be very unfortunate.

This is just a thought ….

I wonder if you could post some overall statistics?

How about something for all keying done by everyone? How many records have we collectively keyed and what is our accuracy? I like seeing my personal stats, but what really would motivate me is to see that we collectively have keyed a million records and were 85% accurate (or whatever). That feels to me like we are all a group of committed people working on the same goal because we LOVE this genealogy hobby. I love the little bar graphs showing progress on the projects in process. More of these kinds of things that show our collective efforts might help us all feel good about what we are doing.

Think about the United Way’s Thermometor for their annual campaign, or the numbers flashing on the telethon screens. Those kinds of things are what build momentum for the community.

One last idea — How about overall accuracy by project. Let’s face it — some are easier than others. If I have chosen to work on a more difficult project my ratings will be lower, but so are everyone’s. At least you could see that everyone was struggling with the handwriting on those marriage bahns or the Nebraska census, etc.

I appreciate everyones comments. There are a lot of things we are still working on so we can move on from Beta…but while we are still working on these items it is extremely helpful to receive feedback, positive or negative.

I am sorry that a few of you are experiencing difficulty with the program. Our developers are working to isolate the issue so it can be resolved.

In regard to the accuracy rating I know it can be frustrating. I key under a few different usernames in order to test upcoming projects and for other work reasons, on one username my accuracy is almost 20% lower than the other! I am excited to go into meetings on Monday with your comments so we can discuss possible changes – knowing that changes will have to wait until the new year.

I personally feel it is important to participate in this proram and make my contribtutions so that other researchers benefit by having these indexes free online. It is a great feeling to know I am able to help some one else with just minutes, well sometimes hours, of time spent keying in names, dates and places, etc. It is that thought that I want to leave with you, think of records you are searching for and the benefit that these records we are keying will have for yourself and others. We try our best and hope we are getting the names right and we are all in this together.

Thank you so much for your contributions of time, talent and willingness to participate in a great effort!

As previously noted, high accuracy scores are more difficult to achieve on certain projects – like transcribing handwritten records rather than typed cards. Does WAP take this into consideration when looking at accuracy scores for selecting arbitrators? Seems like individuals who show dedication by tackling especially challenging records should receive some consideration other than just score value. I’d like to know if this is figured into the selection determination. Thanks for all your efforts.

Has anyone else experienced this problem when keying data from the New England Naturalization Index?

This has only been an issue since I updated to the latest version of the keying tool. In this Index there are three cards on each image. When entering data, the first line from each image loses the ‘Res State’ field when you move to the next image. I noticed this when I had to go back and check some information that I had entered and found that on every image this field was missing.

You can go back and retype this but this is annoying and the best work around that I’ve found is to re-type the field before proceeding onto the next image. Re-typing it makes the data stick.

This problem may be contributing to some of the accuracy issues some people are experiencing.

Update to my problem – the work around of re-typing the ‘Res State’ field before moving to the next image does not work 100% of the time. Many of the records I still need to go back after moving to the next image and add the field back into the record a second time before the data sticks in the field.

I am having trouble with the highlight for the Nebraska census jumping up and down on the page. Before the software update the highlight seemed to gradually get lower and lower, compared to the field I would be keying. Now by halfway down the page, it might be three lines too low on one column, and then a line or two too high on another column. It’s very hard to keep track of where I am.

I have a follow-on question from the following comment included in post # 30 by Joan:
“If a field is required, something must be entered into the field. You cannot merely leave it blank, you must mark it blank”
I have keyed a number of different indexes and don’t recall any of them specifying that a blank field must be marked as such In fact, some of the instructions say to skip a field if it is blank. On the other hand, the keying tool allows you to specify a blank, so I assume it must be used under certain circumstances.
Any advice from the experts ?

Hi Tim. The following excerpt was copied from keying instructions. Although I’m not an expert, I hope this helps.

“What fields should be keyed?

We have required fields, these are the highlighted fields, that are based on information that all of the records should contain. If this information isn’t available the fields should be marked “blank”. There are other fields included on form types that some records may contain, but since not all of the records have this information it would be time consuming to mark all of them blank if the information isn’t present. If there is a field available and there is information on the record that is applicable this information should be collected.

There is often information on the records that we do not have fields for. Although it would be nice to capture more information, the main reason we don’t ask for all of the information available on a record is that we are creating an index. The information we gather is based on information that will be searchable.”

Hi Jane. (Item 45)

When you are in the keying area, on the bottom right there is a link under where it says My role: Keyer

That link (See community leaders and discuss online) will take you to a dashboard that has all the projects and the collective results. Hope this helps. The link is below.

Sorry. I find no indication of my accuracy rating on my stats. Seems I am missing something.

Accuracy as I understand it is only calculated when there is a referral to arbitration. That means that two keyers disagree upon a record or records in a group or individual card data. Then each keyer gets an accuracy rating. That means that is the arbitrator disagrees with your input you get a lower rating and the one with which he agrees gets the higher rating. If you records are arbitrated only once you can have a very high rating or a very low rating depending on how many errors you were judged to have made. If you differ by only one record you keyed it wrong you have a 0% accuracy! Conversely if you were right you have 100%!
You just have to wait until other keyers key the same records with different entries then you will be rated. Until then you don’t see any accuracy for your work.
You shouldn’t be dismayed though. Over time your skills will develop and your rating will be more representative of your work.
I have keyed nearly 7000 records so far and have not received a rating yet. If the second keying agrees with the second keyer or there have been no keying of the records you have completed you won’t see any thing in the accuracy slot.

I was one of the people that wanted guidance on how accurate my keying was. Having read some of the feedback – maybe the answer is to be able to choose as part of the set up system whether or not to get this feedback.
I really did want it and am happy to have it as a percentage of the arbitrated items – this actually is a good guide for me. I know that I have issues with spelling (dyslexia) but this can be positive. I type all of the names (don’t use the scroll bar unless absolutely out of ideas). What’s the positive of dyslexia; well actually I can read names others can’t decipher because I tend to read word shapes rather than individual letters, which sometimes allows me to see the word below a very strange bit of handwriting.
I don’t know how I will feel if my percentage slips, but I suspect anything over 50% is at least acceptable. Professional “keying” accepts an inaccuracy level of around 95%, but they are not trying to read old documents with very individual handwritten entries.
It would be a real shame if any one dropped out because of this. Hold onto the fact that some of these records are really difficult (even when assessed as average). If two keyers differ, both could be incorrect, even when arbitrated the answer could still be wrong – the nearest match is the best that we can achieve in some cases. What really matters is keeping these records for others to use in a useful format – not how accurate you are as an individual keyer. Just keep keying and Happy Holidays!

When there is an obvious mistake by the censustaker, such as entering “M” for the sex of “Eunice” who is plainly shown as “Daughter”, should we enter the “M” anyway, or should we enter the correct “F”?
How about clear, unambiguous, misspellings of names, counties and states? Should they be entered as spelled, or as they were probably intended?
Are you more interested in an accurate transcription, or an accurate interpretation?

I am having the same trouble as Nancy #50. The highlighter is jumping all over and it’s really difficult to keep track of where you are on the line both horizontally and vertically. I have keyed over 7800 records and have no accuracy rating. Also goes against my grain to put Annie as a male daughter!

I’ve just had my stats updated from 70% to 50% accuracy. I’m so sorry to have wasted your time. Unless someone can tell me where I’m going wrong, I give up

I was eager to help keying in data but am unable to do so as the form onto which data must be entered is largely off the bottom of my screen and there is no way of moving it into view.

My first accuracy rating was 81%. Since then, every time I submit, it goes down. I submitted a Nebraska Census set this morning, and the number of entries did not go up by 100, and the accuracy rating went down from 78 to 77?????

I have found that the easiest way to do that Census is to use 50% size of the page. That way, I can see the name on the left, and the state/country on the right at the same time. (The highlight doesn’t stay on the correct line)

Before I had to reinstall my operating system, and redownload the keying tool, I was able to change the color of the highlight. Now I can’t find that choice, and I like the highlight to be yellow, not red.

Still enjoying!