Page Tabs

Sunday, January 20, 2013

The pinball approach to genealogical research

With red and white lights flashing, the pinball machine shrilly rings "Ding, ding, ding" every time an obstacle is hit, and the metal ball ricochets off in a different direction. Just where that ball goes isn't so important since the goal is to keep hitting the obstacles to gain points and to avoid the holes and troughs that lead to the ditch at the bottom, ending the player's turn. In the frenzy, coarse words may spew, and there is always the temptation to "tilt" the machine itself to get a higher score.

Competent genealogy researchers must take pains to avoid the "pinball approach to genealogical research". There is no point in quickly bouncing from one document to another, without fully considering the info each contains. Moving too quickly means you may be barking up the wrong family tree. It also means possibly neglecting to search for other documents in the locality that may prove useful in your kinship study.

One thing I've learned during this week's course titled "Advanced Genealogical Methods" with Thomas W. JonesPh.D., CG at the Salt Lake Institute of Genealogy:

It is essential to "get outside" your usual genealogy management program by using a spreadsheet program like MS Excel, or tables in a word processing program. 

Why? You'll want to create a timeline of the documents you collect indicating events in the lives of your research subjects. Some documents may be irrelevant, and won't be attached to the ancestor in your genealogy program.
Research subjects? Yes, I know you are looking for one item, say, the father's name. Quickly you'll accumulate a few documents that mention your known ancestor. Once you've exhausted the records that directly mention your ancestor, you'll be canvassing a variety of record groups looking for information on all:
  • with the same surname
  • in the same locality
  • for a period of 100 years or more 
From information in these documents you may be able to deduce or infer kinship.

During my week's SLIG study, Ol' Myrt here realizes my broad searches are disorganized.

Don't get me wrong, I understand the theory of community in kinship determinations - the FAN club. (Friends and Neighbors)

I am describing my "sometimes" disconnect that is overcome by the physical move from say RootsMagic or Legacy, to using a spreadsheet for my analysis. Doing this gets me out of the person-centric rut and into community-oriented research when solving complex kinship challenges.

For some folks, analysis of documents works well using "mind mapping" techniques, where blocks of info can be moved around. This just doesn't work for me. Perhaps having taught Excel in the post secondary setting informs my preference for a spreadsheet workspace.

This week Dr. Jones directed us to create tables of the info we're finding in documents, then do our comparisons. When looking for a father, most of my classmates developed worksheets that look like this:

NOTE: The citation column is not shown in this view.

The idea here is to put ALL entries for ANYONE named John in the records of the locality in the John column, even if you can begin to see a pattern indicating perhaps more than one Henry lives in the area. The same goes for anyone named Ezekiel, Thomas, William, Robert, and Stephen. Consolidate where spelling of the given name or surname varies. Sort out who is who during the analysis and correlation phases of research. Be sure to make note of all individuals mentioned within the document and those witnessing each document.

Previous attempts at this type of spreadsheet analysis found me creating an additional John column for each individual I suspected was a different John. I was jumping ahead too quickly, and the spreadsheet became too wide to effectively analyze each item individually and later correlate the items together in the broader view.

Having created that worksheet for analysis, it works for my brain to print out what I've found, and then draw color-coded circles around those who are clearly one John or another.

Previously, I had also inserted a line for "estimated birth" say based on an age in the 1850 census. What happened for me was the list became too lengthy for me to analyze the info accurately. It is easier to write the estimated birth dates in pencil in the John column circling documents I believe are concerning one individual and not another. All this is part of correlating the information to identify each unique "John".

Am I silly for using a printout and colored pencils? It works for me. Would this work for any of my DearREADERS?

In my case, the mechanism of using a genealogy software program or the pedigree chart on Ancestry Member Trees prevented me from doing this sort of analysis. This is not the fault of those valuable resources. With that type of research, either a document directly mentions an ancestor or it doesn't apply and one ignores what may be relevant in the final analysis.

In the olden days we'd put an extra leaf in the dining room table to lay out the documents under consideration when attempting to answer the question, say "who is the father?" or "what is the wife's maiden name?" We're into complex research, where no single document has the answer we seek.

Without a spreadsheet, I wasn't able to process the mass of seemingly unrelated documents in a broader search of all individuals in the locality by the surname. 


Fast-paced games of speed and thrills should be left to pinball enthusiasts. Family historians find better results following a systematic approach to genealogy research.

Happy family tree climbing!
Myrt     :)
Your friend in genealogy.

Twitter: @DearMYRTLE
G+: +Pat Richley-Erickson
Second Life: Clarise Beaumont


  1. Printouts and highlighters have worked for me. Sometimes I need to see it on paper and not a monitor. Having a printer that prints on ledger size paper helps, but I have also taped letter-size sheets together for wide spreadsheets.

    I've usually used multiple columns for multiple Johns, too. I think if I switch to one column for all Johns I might be tempted to use different colors to group the items that are obviously associated. I wouldn't want to overlook, for example, all the entries associated with the same tract of land. But I'll have to try this one column per name style and see if it works for me.

    Thanks for the idea.

  2. Thanks so much for this post. One of my goals for this year is to get more organized with my research.
    Great ideas!

  3. The spreadsheet chart you showed seems terribly inefficient to me. If you have entries in the time frame (e.g., 1756-7) for all of the names, it would be very difficult to read them vertically and compare the information in the columns.

    I don't have a good solution for this. My preference is to create a table for each candidate father (in the case you showed) and list the documents found with summary results and quality evaluation, and also put them in a timeline. It's hard to see what you don't have without a list of record types, and you can't see if the person is a candidate without the timeline.

    1. I hear you Randy. As the weeks progress I'll try to better demonstrate the use of a spreadsheet as Dr. Jones suggests.

      In the end, it is whatever works for the individual researcher. The goal is to effectively analyze individual documents, and then correlate the information in the context of all documents for, say, the likely fathers in the locality in question.

    2. For me, Randy's approach would be the second stage after your approach, Pat. Once one had accumulated enough data suggesting characteristics of more than one person named John, then one needs to do more digging oriented to each one's characteristics (that is, the preliminary conclusions) to verify. There could be more by a name than readily apparent, such as similarly-aged first cousins living next to each other. I am still stuck on my Aholiab Sawyers, but have straightened out my Richard Moores thanks to land records in conjunction with tax rolls (bless the assessor who needed to keep straight who to collect from for what, promoting "Jr." after "Sr." died, moving up "3rd" at the same time, assessing one without a suffix and showing him with land bequeathed by his father's will) (these were 2 fathers, 2 sons; one of the fathers was son-in-law of the other; they bought adjacent farms at the same time). That problem required close attention to timing, immensely eased in a spreadsheet. A splendid run of surviving records, including estate proceedings, was indispensable in this case.

    3. Dear Geo,
      Yes, I think once you look at the men in an area via the surviving records, recorded as I mention in this article, THEN you can move them to a worksheet as Randy describes.

      I am glad a spreadsheet worked for you. I consider them flexible. Multiple tabs, for working off of the initial listing with links to that original list is great as well.

      I also like the color coding of text.

      One of my fellow students talked about inserting comment boxes. For me, information needs to be constantly visible, so I'd insert a comment column rather than using the Excel comment feature for cells.

  4. The pinball approach to genealogical research has been my approach over the past but after 20 years, now I'm trying to move away this style of research. Your article and all its link has help to direct, focus and motivate me. Old habits are hart to change..... but persistence will knock down the resistance.
    Thanks for your informative post.

  5. I use Excel spreadsheets all the time to abstract, summarise and analyse genealogical data. I find Excel spreadsheets more flexible than tables in Word, but what Excel does with dates drives me crazy. I make liberal use of 'fill colour' and text highlight, so rarely resort to paper versions.

    1. The problem with the dates in Excel has a fix. You select the cells, right click on them, then select "format cells". On the "Number" tab is the category "date" where you can make the adjustments.

      Once you have saved a template, whenever you start a new research project, open the template with the date settings you like, then "save as..." to rename the file, preserving your template for the next project.

    2. Excel is the bane of my existence with dates prior to 1900. Especially if you ever need to sort your entries by date. (If you don't think you'll need to sort by date, you should just make it a "text" cell and enter your dates in the standard yyyy-mmm-dd format).

      In any case, this looks like a method I could wrap my head around - and I might just try it and see what happens :)

    3. I use two columns for dates in Excel spreadsheets, especially timelines: one column I name "DateCode" using the format 2013.01.23 - the other I name "Date" using the 23 Jan 2013 format. When sorting the spreadsheet by date I just use the DateCode column, which puts the Date column info in order. Uncertain dates I code as 1895.00.00 and indicate "abt 1895" in the date column. 1900 census birth info gets coded 1833.07.00 and the date column reads "July 1833." For dates after a certain date I use 1918.06.01 to indicate "aft 31 May 1918." Has worked well for my needs.

  6. This particular spreadsheet concept might work very well for some of my stubborn clusters. Sounds like an off-shoot of a 'brain-storming' thing.... All data points and references to 'any John' in one column, for 'any Seth' in another, and don't let yourself analyze until it's all there.

    *Then* go through gradually hash things out. Print-outs, highlighters, color-coding, pencils, cool. But you haven't accidentally mixed up data by assuming anything, before it's all there in front of you.

    Ye gods, I need to make use of this! Thanks!

  7. I am The PINBALL WIZARD QUEEN! Really hard to stay focused. Too busy reading all the genealogy daily email's that I love and reading everyone's business of the day. My day dictates on WHAT my favorite blogs talk about. Unfortunately, too little is EVER said about Italy so it s mostly Ireland I attempt to work on.

    1. Ok, just for you, I'll do an Italian research blog post. Stay tuned!

  8. The viewpoints you make in your article are so well stated. This is easy to understand from the beginning. This is interesting to read. Thanks for clearing up some things I have been thinking about.
    irishfamily roots