In the September 2019 issue, Civil War History published an essay by Earl Hess called “The Internet and Civil War Studies.” Thanks to historians on social media, I knew to be outraged by the essay a full week before I was able to read it. (And thanks to email, I was able to read a copy scanned by a colleague with whom I have collaborated.) In any case, folks on Twitter were mad at Hess, and soon became mad at CWH, and at each other. It was not a proud moment for social media, but it did demonstrate how quickly things spread on Twitter. For every historian who commented on Twitter I probably received one email from another historian who was aware of the kerfuffle and wanted no part of it.
Earl Hess is a superb scholar, an indefatigable researcher, and just amazingly productive. This interesting essay demonstrates those research skills while also showing some limitations that perhaps deserved prepublication consideration. Hess surveys the relationship between history, historians, and the internet in a detailed study which will be a benchmark for future historians asking the same questions (which is a nice way of saying that it will be outdated soon and in some instances is already curiously behind the times). After some useful preliminaries, Hess looks at six categories of analysis. He considers research in primary sources available on the web; archives and special collections putting their materials on the web; marketing books on the web; collaborations on the web; historians and social media; and, finally, “informational websites.”
Along the way, Hess has assembled a huge amount of valuable information. I particularly like his attention to the role of archivists in this story. He gives considerable attention to my colleague Dr. James Cusick who has done wonderful things at the University of Florida’s Special Collections, including two marvelous web pages full of primary materials. There is lots of valuable evidence here.
But then there are those surveys. In 2014 Hess published a different piece on the state of Civil War studies in CWH, based largely on his 2013 survey of some Civil War historians. I have some background in quantitative methods and at the time I was on the board of CWH. At several meetings I offered to look over this piece, but I was never given that opportunity. I thought the published essay was methodologically pretty bad (for reasons that follow). I did not share those thoughts in print, thinking largely of something about horses and barn doors. It never occurred to me that we would hear from this survey again. Oh well.
Large portions of this essay are built upon that survey Hess ran in 2013 and on a subsequent survey Hess ran with CWH in 2016. The methodological problems are many. Here is a list:
- The surveys are small (2013 apparently included 33 respondents; 2016 had 95). The essay puts an awful lot of weight on very few voices.
- There is no indication of how those polled were selected, or how many people declined to participate. (I am pretty sure I received and never responded to the 2013 poll. I do not think the journal invited me to participate in the 2016 version.) Hess quotes a graduate student named Jonathan Jones who responded to the 2016 survey after finding it on Twitter. Does that mean that the respondents to this poll were quite literally self-selecting? Readers should understand these things.
- More importantly, there is no indication of the demographic breakdown of respondents. Were they mostly old white men, as Twitter historians seem to suggest? Did it include untenured scholars? Were women – older and younger – equally represented? How about people of color”? These critiques should have been anticipated and addressed.
- Beyond the obvious considerations about generation, gender, race, and so on, there are just questions of the professional life cycle. For instance, Hess asked his respondents about if they used social media to market their books. The question might make perfect sense to Earl Hess, who seems to publish a new book every six months, but shouldn’t the survey results have told us about the behavior of authors who had published a book in the last few years rather than all respondents?
- In this essay, and much more so in the previous CWH piece, Hess violates all sorts of cardinal rules about describing survey evidence (beyond all the failures in describing survey methods). He writes of the opinions of “a large percentage of Civil War scholars” (211), when he really means “a large percentage of 95 [or was it 33?] respondents several years ago.” Later, he refers to the opinions of “most Civil War historians” harboring “reservations” about social media (227), when he is really referring to “the majority of folks who responded to this survey.” The essay concludes with an observation about “the overwhelming majority of Civil War Historians” (234), when the evidence simply cannot support the assertion. The simple point is that it is professional malpractice to present a very small “survey” without any explanation of methods or the demographics of respondents, and then describe what they say as “most historians.” That simply cannot be done.
- This will sound picky, but throughout the essay Hess presents his evidence to two decimal points. That suggests an impressive level of specificity, but when your survey size is 95 it is pretty silly and not the sort of thing that I would imagine anyone familiar with quantitative methods would do. If your data says 48 out of 95, just round it to 52 percent.
There is more, but that is more than enough. Hess is a superb scholar, but he does not appear to be a quantifier. And perhaps the CWH folks should have been more attentive to this sort of thing. There are dozens of Civil War historians who are comfortable with quantitative analysis who could have reviewed the piece.
Then there is the matter of how Hess treats social media and blogs. Essentially Hess has assembled six-year-old survey quotations about how folks feel (or, really, felt) about blogs and other such things. Not surprisingly, some of them do not (or did not) like them. This leads to entertaining quotations of unclear value. Are readers of CWH really surprised that Allan Guelzo is no fan of social media (or was not six years ago)? Although there might be value in speaking critically of blogs and blogging, did CWH really need to run a lengthy, six-year-old, anonymous quotation from somebody who disliked Kevin Levin’s blog (back then)? The entire section feels unnecessary and not very useful. The punchline is that some folks like blogs and social media and others do not, or at least that is how they felt in the recent past. But the methods underlying the surveys allow for no systematic analysis that makes any sense or adds any value to the discussion. It is just hard to imagine why this discussion and these old quotes are in print.
(In putting together this essay I ran a small survey (sample size = 1; response rate 100.00%) of historians who are quoted in this essay. My respondent had no recollection of filling out the survey but said that the quoted passage surely could have been by him/her, although perhaps taken out of context. In any case, Hess had not contacted the respondent about using his/her words and name in this publication. I guess that is technically okay but it seems surprising.)
Finally, we have the social media train wreck on twitter. Kevin Levin, who was on CWH’s editorial board just before the essay appeared in print and felt he deserved some consideration given the content, shared passages from the essay on Twitter before most folks had seen it. Those passages included (old) quotations that critiqued his blog. Folks lined up to attack Earl Hess even before they had read the whole essay. Some cast the conflict in generational terms. Hess is an old guy yelling “get off my lawn” at bloggers. There is probably some empirical truth to that argument, although – like so much on social media – it seems unsupported by empirical data. [Aside, Brooks Simpson, who has nothing at all to do with this conflict is only two years younger than Hess and a pioneer in blogging.] Other responses suggested that Hess’s apparent hostility to blogging reflected unsavory tendencies concerning race and gender. It may well be true that historians who are active on social media and on blogs are less white, less male, and less old than other historians. That would not surprise me at all, although I have seen no such data. But the critique seems to require that Hess is aware of these facts, which strikes me as less likely. Has he been reading blogs enough to notice their authors and focus?
Somewhere along the line a member of the CWH board – who apparently had never actually seen a Twitter conversation before – weighed in and tried to defend Hess and the integrity of CWH. (Some folks had suggested that the whole board resign, for instance.) At some point, he made an unfortunate comment about the “tone” of the criticisms, which unleashed new torrents of critique. He responded that his critics were being “Trumpian,” elegantly giving up the high ground. I have no thoughts to offer about “tone” (because I was not born yesterday), but I will note that one early respondent wrote: “I just googled him. I saw the picture and probably have enough information.” Earl has an impressive beard, which was perhaps the information the poster needed to draw conclusions about him? It might be reasonable to conclude that that poster was appalling? Certainly the tone was not ideal.
So, in my opinion Hess’s essay included much valuable information, although it made a methodological hash of those surveys. It raises some interesting questions and CWH can reasonably argue that the essay opens a valuable conversation. Other reasonable people might wonder why CWH would ask Earl Hess – a great historian and a very nice fellow but not a guy known for either numbers or a social media presence – to be the scholar to start this particular conversation with these particular surveys. On the other hand, in my humble opinion nobody should resign because of these transgressions, even if they might rethink them.
University of Florida