Correlation: When and How

I am presently teaching a course for Applied Genealogy Institute on “Write As You Research!” a methodology that instills an efficient discipline of writing that assures us that we have extracted all relevant information from a source. A research report for a client is a necessary element of the Board for Certification of Genealogists‘ portfolio. While we each have our methods of presenting information relative to our research question, students often seem puzzled about when to correlate and what is the best correlation tool to use.

[Author’s Note: only a portion of the reference notes are included. form dictates that they are always associated with the table as shown.]

Thomas W. Jones, author of Mastering Genealogical Proof, outlines a number of tools to correlate information, including a bullets or a list, a table or timeline, and a map.[1] Each type has a role and each type is important. I must admit that I have never met a table I didn’t like. I tend to gravitate to using tables to display my information to the reader. I “see” the best method and I “see” the best information to correlate without any issue, but some students struggled with this. This post focuses on when and how to best correlate.

The above image is from my portfolio.[2] It compares three record sets: church book entries of birth in Germany; a passenger manifest of the candidate family; and the 1860 census of the same candidate family. The table responds to the research question of are these the same family? Relationships are noted in two of the records, gender in the third. Each included birth dates or age by which a birth year could be calculated. the spelling of the names varied widely and some family members were out of the house by 1860. This is a classic table comparing the same family (hypothesized) in three different records. It could just as easily been comparing the 1900, 1910 and the 1920 censuses.

Correlation is a method by which the adage of “a picture is worth a thousand words” is paramount. It focuses on the important information needed to come to a conclusion; it eliminates that which is not. This might result in a final conclusion or it might be one in a series of sub-conclusions of a genealogical problem. While one might not recognize that the passenger manifest is the same family as the other two, by placing them in this table and comparing the three records (fully cited, of course, in the submission), one can more easily come to the conclusion that they are all members of the same family just in three different sources–parish records in Germany, passenger manifest and a census.

Here are some other simple examples:

Scenario 1, Narrative: If the logic follows easily from A to B to C, a paragraph might suffice to correlate the information. The students sometimes defaulted to this method but crammed too much extraneous information into the writing. This resulted in a long paragraph which was hard to discern the wheat from the chaff.

Scenario 2, List: If the comparison is easily comprehended, a list might be the tool to use. This usually starts with a sentence of what is being proven; followed by a numbered or bulleted list of the multiple pieces of evidence that support the assertion; and finally a concluding sentence.

Scenario 3, Table: If you were trying to determine if Fred Smith in Ohio is the same Fred Smith in VA—you would need to research both, list the identifiers of each, and compare them in a table of the two men. One should stop having events in one location and start having them in another at the same time with no gaps or overlaps. The identifiers would be any facts you can document about both Freds.

Scenario 4, Timeline: If you are trying to identify the parents of Fred Smith and his birth date. You might develop a timeline of Fred’s lif starting with his death and working towards his birth, the desired information. If you lead with the birth (let’s say you found it), the reader will wonder how you got there.

Scenario 5, Map: If you are trying to visualize how difficult or easy it was to meet the spouse of your ancestor, you might map the residences of both. If your ancestor moved frequently; or bought land in multiple locations–consider a map.

ADAGE: If one tool doesn’t work, try another.

Let’s go back to the beginning for just a minute. I am going to assume you have a good research question with a person of interest. Where do you start?

First, look at your person of interest (POI) in your research question:

• What are you trying to discover? Birth? Death? Parents? Span a migration gap?
• What data points are unique to to your POI and would inform your conclusion? list them.
• Where can you find answers to any of the items on the list?
• Do a napkin sketch.

The list might be the items you compare in your Column A of your table, e.g. name, age, location of birth, spouse, location of birth of parents, etc. The horizontal axis might be the various record sets, e.g. censuses, draft registrations, naturalization records.

No matter what your research question is, it is necessary to identify your person first before you can prove a relationship. If you are trying to discover the parents of your ancestor, realize that you have to build the person’s life from death back to birth. This timeline of events can have no gaps or overlaps which would indicate that you might have conflated one or more individuals.

ADAGE: Obtain evidence from a variety of sources as close to the event as possible. The death record may name the birth date and parents, but it is not closest to the event and the informant was probably not at the birth event. Marriage records may state age, and parents. Can we get even closer to the event? draft registration, any event where one had to be of a certain age, e.g. military service, etc.

I use Evidence Blocks to present my information, which might be evidence. Each Block is composed of the citation; an image (snip) of the source; a transcription, extract or abstract; and my analysis of the information provided in the source that relates to the research question.

ADAGE: Build your case one source at a time. Assess each source for quality and independence. You do not need to justify the use of a good source, but you may have to explain why you are using an index or a book that lacks source information.

Once you have presented the sources containing evidence to the reader, you can move towards building the table.

.ADAGE: you cannot correlate what you haven’t yet presented. 

ADAGE: At first, err on the side of gathering everything. You don’t know what you don’t know.

ADAGE: Not all that you correlate will end up in the final product.

In my portfolio and my renewal, I built tables that never showed up in the final paper. They were my way of seeing the data differently.  Some made the final report, some were modified greatly and others were kicked out. Sometimes you think of a better way to convey the information.

ADAGE: As soon a reader asks themselves, “I wonder if…..” then it’s time to put in the correlation.  Corrleations may happen multiple times through the document.

WHEN TO CORRELATE

• when you have data that is comparable
• when you or the reader recognize that you have multiple answers to the same question that may not agree with one another
• when you have presented the data and you or the reader ask the question, “I wonder how it changed over time.”

WHAT TOOL TO USE

• the best correlation tool is the one that presents the information in the best way possible for the reader
• If you try a narrative or bullets and it doesn’t work so well, switch to a table or a timeline
• if you don’t like the way the data is arranged in a table, switch the x and y axes.

I hope this has been helpful. I love seeing the data in different ways. It forces me to ask questions I might never have asked.

Happy hunting!

Jill

What I have done since the last post; Mostly related to teaching the “Write As You Research Course!–prepping for class, reviewing homework, adding new content based on issues I identify in the homework, getting the evaluation and the certificates ready for distribution.

[1] Thomas W. Jones, Mastering Genealogical Proof (Arlington, VA: National Genealogical Society, 2013), Chapter 5.
[2] Jill Morelli, Element 7: Kinship Determination Project,” to fulfill the requirements of certification by the Board for Certification of Genealogists, submitted 2016, approved 2017. And yes this table was fully cited with original records.No informants were identified. It was likely that the informant of the birth records of Jan and Antje was probably the father of each. Jan/John was probably the informant for the passenger manifest and the 1860 census.

Advertisement