Discussing negative searches

When writing a proof argument, how do all of you decide when and where to draw the line on discussing “negative” searches (that is, searches that yielded no relevant findings)? Providing extensive lists of sources checked is not practical, particularly in a work product intended for possible publication, but failing to acknowledge the examination of important sources would suggest the search was less than thorough.

What criteria do you use to decide which negative searches are worth mentioning?


Submitted byhhendersonon Tue, 03/27/2012 - 17:58

I would include them all and let the editor decide. ;-)

But I would love to see some top-ranked genealogy periodical make this dilemma unnecessary, by placing supplementary material of this kind in an on-line supplement to each published article. (The New York Review of Books does something similar.) IMO space-related editing of this kind reduces the value of such an article to novices, who may not appreciate the extent of the negative searches involved.

Submitted bybvlittleon Thu, 03/29/2012 - 12:33

While listing individual records searched is not always practical, I find that they can often be succinctly summarized. For example,

John Frederick Dorman, a respected and knowledgable Virginia researcher, published abstracts of all Orange County deed and will books from the beginning of the county in 1734 through 1749 when the Culpeper County was formed. No reference to the xyz name was found.

One can often lump other records, as well, for example,

The indexes microfilmed with the extant order and minute books for Orange County from 1734 through 1802 were searched. No reference to the xyz name was found. For a list of these volumes and their indexes (if any) see the Orange County list at "County and City Records at the Library of Virginia," Library of Virginia (http://www.lva.virginia.gov/whatwehave/local/county_formation/index.htm : accessed 29 March 2012).

Submitted bymhaiton Fri, 03/30/2012 - 09:31

I think the first issue is relevance to the specific argument. If you are discussing land records or land ownership, for example, you might include a statement that "no additional deeds were located," etc. In the citation to that statement, you can be as specific as necessary in citing the negative searches relevant to that statement. Other negative searches in the same argument could be similarly inserted where appropriate, with the specific searches detailed in the reference notes.

Submitted byahareon Fri, 03/30/2012 - 12:03

To me, it's a judgment call that has to be made in each case. Some negative searches might be more important than others. If I was researching an individual's parentage, I think it would be wise to show I had searched the sources most likely to provide evidence of that parentage. But if I found a convincing body of evidence in a variety of records, I don't think it should be necessary to describe everything I didn't find. For example, if I can build a good case but cannot find my research subject in a particular census, do I need to detail everything I did to find him or her in that one census? If my case is strong enough, I  personally don't think I need to deal with that census at all. Similarly, I don't see the need to exhaustively detail a search for a burial record in a time period where such records are rare or generally uninformative.

Because it's impractical to cite every negative search in a published work, I do think we have to choose and I think careful selection of what is presented as well as the scope of research and attention to detail overall will convince me of a conclusion as well as an everything-including-the-kitchen-sink approach.



Submitted byACProctoron Fri, 07/31/2015 - 16:31

Some good suggestions here, but my feelings are rather more straightforward: If you would *expect to find* a reference in a partticular source but none was found then list it.


Submitted byEEon Sat, 08/01/2015 - 10:43

Tony, to complicate a simple approach, may I ask for a definition of what constitutes 'expectation'? In research on the colonial and early federal Southern frontier, if I study court records to analyze the prevalence of certain types of civil and criminal offenses, I can expect to find cases of murder, assault and battery, larceny, and bastardy. If I'm searching those records for information on a specific person, I almost never expect to find that person because I know the odds are against it. If I am searching for a particular person in that society and I use a register of deeds or wills, I never expect to find the person there, because the odds are against it. &c. Yet, it is important to management of the research project that I record each negative search, despite my expectation.

Isn't "expectation" a manifestation of the experience of the researcher, rather than a methodological guideline?

I understand about the difference between a negative search and then using the result as negative evidence -- the records may not be complete, for instance. However, I admit to being at a loss to answer your exact question. If I knew that someone was in court then I would "expect" there to be a record (in an ideal) world, as opposed to when doing a search for some random relative whom I have no knowledge of being in court.

The absence of information in a given source must be assessed according to the completeness of the source, whether there were acknowledged cases of relevant people not appearing (e.g. early civil birth registrations), and ideally to compare with related sources.



I totally agree, Tony, on your negative-evidence perspective; and your published work demonstrates how solid your own expectations are.  The issue I see with using that term "expect" as a guideline is that a legion of researchers who do not yet have your experience—and the judgment that comes from it—have much different "expectations" and would, therefore, fall short of meeting a reasonable standard if the standard is set to "expectation." No?

(Odd, I keep missing the responses here because I'm not getting notifications. Probably my fault somewhere)

The more I think about what I thought I was doing, the more interesting it gets. If every source in the whole world was accessible then we'd exercise some filtering based on the relevance of each source. I was suggesting that my "expectation" of finding the relevant information might involve a consideration of the balance between "would it be reasonable to find a reference" and "would it be unreasonable to find a reference".

Once a source is selected then its usefulness as evidence -- negative or otherwise -- might involve a new assessment between "would it be reasonable to find a reference" and "would it be reasonable to not find a reference".

Of course, I'm not entirely serious about this  I've probably had too much coffee and felt a bit mischievous. :-)


Tony, coffee and mischievousness do sometimes keep company. :-)   Thinking through these issues more deeply helps us all. Thanks for sharing your thoughts with us.

Incidentally, when you create a message, look for an option below the message that says :

Notify me when new comments are posted  If you have the box checked, then you should get that notification.

Submitted byACProctoron Sat, 08/15/2015 - 05:04

In reply to by EE

That box is always pre-ticked for me but I'm still not receiving notifications. I even changed "Replies to my comments" to "All comments" but no change. I certainly used to from this site.



Submitted byJadeon Sat, 08/01/2015 - 15:01

Very interesting discussion.

If I have found that a US Census enumerator omitted ages, I will cite the particular enumeration as an instance of negative evidence.

But if US Census enumerations for a particular person did not indicate that the person was illiterate, that is a negative finding and relevant to a question at hand.  But do I need to cite each of those enumerations when no other features of those enumerations were relevant to the issue at hand, or were not in question (say, place of residence or age)?

Would omitting the citations just be laziness on my part?


Alas, a mind-glitch on my part.  The enumerator's not checking off the boxes indicating inability to read/write is not a negative finding.  It is a positive indication on the enumerator's part (assuming that the NARA copies were consonant with the enumerator's intent), as well as relevant to a question at hand!


Submitted byGaryCollarinion Tue, 06/28/2022 - 10:08

   After reading through all of the suggestions here, I couldn't find a basic template for citation of negative findings. Is there a specific formula to follow? As in, if I were to want to cite a negative finding from an online database, would I follow citation form as if I were to have found something, but in place of listing a certain page number or line number, replace that with a description of the negative finding? Looking to clear this up on a structural level in order to be able to translate it into use for all source types. Thanks in advance.

Submitted byEEon Sat, 07/09/2022 - 08:46

No, Gary. There is no basic template for citation of negative findings. Every search differs. For example:

  • Your negative search may have used just one source--in which case you cite it as you would cite a finding from that source and then, instead of citing a particular page/whatever on which certain material appeared, you state that the search yielded no result.


  • Your negative search may have included every record book and loose file in a county court—in which case you say so and give a general description of all the sources used and the methodology applied.