From Mass to Niche

We’re happy to report that we have a new article out today, “From Mass Medium to Niche Medium: Advertising in American Comic Books, 1934-2014”. This piece appears in a special issue of Comicalités edited by Jean-Paul Gabilliet and Nicolas Labarre, and it collects several articles from the conference on Bédéphilie that was held in Angoulême in June 2019 (in one of the most severe heatwaves France had ever seen - it was almost impossible to set foot outside the museum in the noon day sun!)

This contribution has been a long time coming. We’ve spent a lot of time thinking about paratexts in the WWC corpus because they were the easiest thing to count, and, thus, the first thing we counted. Some initial thoughts on the topic were presented at the Canadian Society for the Study of Comics conference at the University of British Columbia in 2019, then refined for Angoulême, and then refined again for the essay that you can find here (on the journal site) or here (on ours).

This essay really sprung from the ad pages like the one I’ve included up top. Working through a count of all of the ads in the earliest comics in our corpus was pretty straightforward as there were generally only four or five per issue, but the rise of these Yellow Pages-style collections of mini-ads in the 1970s (primarily in comics from Marvel and DC) really threw our calculations for a loop - suddenly we had books with dozens and dozens ads that had to be recorded. And, of course, not very costly ads either. What did we make of this shift? Well, we lay all that out in the article.

There are a couple of additional things we’d love to write about when it comes to advertising in comic books. We’ve moved on from a simple count of the number and kind of ads to an analysis of the content - are the ads themselves comics? Do they contain comics intellectual property? We are curious to see if graphing that type of data undermines or reinforces our observations about the sheer number of ads in comic books over time.

One other area that I’m particularly eager to explore is the presence of ads for comic book dealers in comics - the image above has three of them. The development of the direct market that Dan Gearino talks about in his book figures prominently in our current article, but I would like to come back to the way figures like Robert Bell (advertising near the top left here) shaped the field by the advertising of comics in comics.

That’s for later. Until then, Gearino has some interesting things to say about Robert Bell on his blog and you can dive down a rabbit hole of looking at Bell’s old price guides on sites like the CGC messsage boards (below is one from 1977, the same year as this ad, which is taken from Brave and the Bold #132, or, as we like to call it, Comic Book #1424).

Pre-Thinking About Credits

Over the past couple of days, I’ve been working to catch up on some data entry for the 400 or so new comic books that we added to the WWC corpus in March. One of those tasks involves recording the presence or absence of titles for stories, as well as where that title appears in the story, and then also the presence and absence of credits on the stories.

Since WWC addresses itself to a randomized corpus of works, our interest in story credits was, initially, fairly limited - a series of simple yes/no toggles indicating, for example, whether or not a particular story credits an inker. We are not especially interested in writing biographies of creative personnel so much as attempting to grapple with the broad shifts in the industry over time. Across the history of the American comic book industry the practice of crediting on stories has been - shall we say? - sporadic and we are interested in tracking those changes.

For example, as you can see in the spreadsheet image above for some books published in 2015, few books credit an inker. This is, of course, due to the shifting nature of the division of labour in the industry. We have been interested in mapping precisely those kinds of shifts.

Two weeks ago, the wonderfully astute comics scholar Rebecca Wanzo spoke to the RoCCET Lab at Carleton University. Dr. Wanzo was discussing her path-breaking 2020 book, The Content of Our Caricature, and spoke about how she conceptualizes her work as part of the scholarly process of ‘recovery’ - that is, of going back into the history of American comics to locate and shed light on the contributions of creators who have too often been marginalized by comics studies. Her project focuses specifically on Black comics creators. Her approach is well established within representation studies for equity-deserving groups, and results in fascinating studies that can totally upend the accepted history of a medium. Methodologically, it is the exact opposite of our work. Hers is a deliberate seeking of neglected works, while we are scooping up everything with a net. It is our hope that with the launch of Phase II, we can establish a supportive infrastructure for scholars like Wanzo to bring that focussed lens onto more subjects, transforming how we understand the comics industry and comic book history.

During the first phase of data collection on creator credits, we did not collect any data on individuals. We may know, as comics scholars and fans, that Carl Barks created many of the best loved Uncle Scrooge stories, or that Frank Doyle and Harry Lucey wrote and drew some of the best Archie stories, but they did so - initially at least - anonymously. When we record whether or not credits are present on a story - and what specific creative and editorial roles are credited over time - we can make observations about the role of authorship in comics generally. Indeed, the WWC corpus includes Barks work from the 1940s and 1950s where he is not credited, but it also includes reprints of Barks work from the 1990s where he is not just credited, but also some where he is celebrated (the Carl Barks Library reprints, for example).

With the new project, and the addition of Dr. Rebecca Sullivan, who specializes in recovery research, we’ll commence with systematically building a middle layer between WWC and Dr. Wanzo’s approach. We will go back through the corpus to record not only what roles were credited on stories, but specifically who was credited. Of course, this raises a whole host of intriguing rabbit holes and even a few Easter eggs. Let’s look at one.

This is the first page of the story we affectionally know as B108S006 - the sixth story of the 108th book in the WWC corpus. This is the sixth story in Wings Comics #36 (Fiction House, 1943).

In our initial coding, this is tagged as having credits (a simple yes/no), and then is tagged as containing an Author credit. By including “By F. E. Lincoln”, Fiction House means to imply that someone named F. E. Lincoln produced this entire work - wrote it, drew it, inked it, lettered it, and so on. While we might presume (correctly) that this was not the case, the story presents itself in a way that is akin to a “By Milton Caniff” or “By Charles Schulz”. This is (deliberately) misleading, but, for our purposes, misleading at least in an interesting way.

In our second round of coding, however, we are now interested in teasing out what information we can get. Who was F. E. Lincoln? I don’t know. The GCD lists ninety stories to someone with that name (mostly from Wings Comics, but not exclusively), and Who’s Who of American Comic Books has no biographical information about this person. It’s likely a pseudonym and may wind up being a dead-end.

But look more closely at this page, to the leaf below the caption at the bottom of the first panel. Here is written “L. Renée”. For our purposes in the first pass at coding this is not a credit. There is no “by”, there is no role affixed. It is an example of an artist signing their work but not being credited for that work, if we are splitting hairs.

As we move forward, however, this is the much more useful piece of information. We know that “L. Renée” was Lily Renée, one of a small number of female artists working in the comic book industry in the early-1940s, and someone who has extensive credits at Fiction House (primarily) on the Jane Martin stories. The GCD has extraordinarily complete information about her comics contributions, and Trina Robbins and Anne Timmons produced a biographical graphic novel about her in 2011. Last week she was also profiled in Newsweek by Jo Ann Toy, in a piece that is well worth reading, and yesterday she was announced as a 2021 inductee into the Will Eisner Hall of Fame.

The story that is told by Robbins and Timmons is a fascinating one, and stems from the same impulse for recovery that drives Dr. Wanzo’s scholarship. The work shines a spotlight on important contributions that might be overlooked. Indeed, were we to have left our own analysis at our first stage of inquiry, we too would have overlooked it, since our coding protocol had no space to negotiate the multiple and often conflicting strategies used to credit comic book stories - anonymity, pseudonymity, deliberate misrepresentation (“Walt Disney presents”), and selective crediting (pencils but not inks…). As we move forward over the next few years, one of our key tasks will be to close these sorts of gaps.

Here We Go Again (Round Two)

We’ve been quiet on the blog during this pandemic year, foregoing reporting on our work in favour of simply churning out more and more data. As we head into April 2021 we find ourselves fact-checking our completed Book and Story level data while research assistants finalize the collection of data at the Page level. Paratext data (ads, letter columns, and so on) still needs some additional cleaning to deal with content, but was mostly finished a few years ago. We had originally proposed to index comic book elements on five planes (Book, Story, Page, Panel, Paratext) and four of those are all but finished now, six years into our original five-year project.

“So,” asked one of our research assistants, “What are you going to do now?”

Well, the answer is “Why stop there?”

When we began this journey in April 2015, we wrote:

Our project is the foundational step in a larger program of work that seeks to reorient the study of comics (comic books, comic strips, graphic novels) through the use of large-scale, quantitative research methods. We will create the most comprehensive and accessible research tool for the study of the American comic book, and we will use the data produced by this tool to write a data-driven history of the American comic book as the development of a set of styles and techniques that existed across the industry as a whole. We believe this will enable new approaches to periodization and force us to re-evaluate many taken-for-granted truths that have long circulated among fans and scholars alike.

That foundation is fairly firmly in place. True, we still have all of our Panel level data to collect, but we have been intentionally pushing that off because we imagined that we might combine that effort with an expansion of our work that would allow us to do two things at once. For that to happen, however, we would need additional funding.

Today we are thrilled to announce that we are, in fact, moving on to What Were Comics? Phase Two

Last week, we were informed by the Social Sciences and Humanities Research Council of Canada that our application for a second round of funding for this project was successful. This means that we will be continuing our work through 2026 (at least).

Given that I just reported that we are almost finished, what does this mean going forward?

Well, we are expanding our work in two distinct directions

First, inspired by Rebecca Wanzo’s observation that comics are an art of exaggeration that “reduc[es] people to real and imagined excesses in order to represent something understood as essential about their character” we want to examine the checkered history of the comic book industry in the United States when it come to representing women, people of colour, 2LGBTQ+ people, and people with disabilities. To that end, we will be producing a systematic picture of textual representation in each of the 17,202 stories included in the What Were Comics? corpus. This analysis will proceed at the Panel level (i.e. What is depicted in every single panel of our corpus?, which is why it made sense for us to defer that effort).

Second, we are also keenly interested in the composition of the pool of professional comic book creators over time. The comic book industry in the United States finds its roots in Depression-era work-for-hire systems that afforded no ownership rights and little creative autonomy. Over the course of decades, it has emerged as a site of creator-ownership where cultural entrepreneurialism is a hallmark. Each of these eras has attracted different types of creative personnel to comic book publishing, and this project will correlate shifting representational strategies on the page to the changing face of creative labour within the industry. Our project, therefore, is the first full-scale longitudinal analysis of the dominant English-language comic book tradition that tracks the two major factors of representational diversity analysis: labour and textual representation.

Big tasks, both of these, and we are incredibly grateful to SSHRC for continuing to underwrite our efforts in this regard.

And, because the task is so large, we have grown our team. For this round of funding we are incredibly proud to have added Dr. Rebecca Sullivan as a co-investigator. Rebecca is a feminist media studies scholar with award-winning publications, substantial leadership experience, and a deep understanding of qualitative and quantitative research methodologies. We are fortunate to be able to draw on her extensive expertise.

Oh, and one final thing: While we were all in various states of lockdown, we decided to expand the corpus to bring it closer to the present day. We added just over four hundred additional comic books to the corpus, representing two per cent of the published work from 2015 to 2019. A few thoughts on that process next time.

Credits: Print vs Digital

It’s been a while since we’ve updated the blog about our progress this project. Despite the pandemic - or, in many ways, because of it - we are actually ahead of schedule at this point. Back in March we hired three dozen additional research assistants to work on counting the panels of every comic book in the corpus, and that data has now been collected (although not input into the database, so we can’t begin analyzing it yet). WIth that task completed, we moved on to counting the words in all of those panels, and we hope to have that data finished by the end of the calendar year - although it, too, will need to wait a bit before going into the database.

In the meanwhile, Bart spent most of the first half of 2020 fighting to process the RA invoices (an amazing tale, full of administrative hurdles, tax forms, and supplier IDs). When he wasn’t working on that, he was fairly mindlessly entering all of the creator credits into the database. We’re just at the point now where we can begin to do some preliminary analysis - and we’ll share some of that over the next couple of weeks.

Today, a word about credits in print and digital.

We have mentioned previously that we are fortunate that we have physical copies of every single comic book in our corpus. While there are legal scans of many of the books in our corpus on sites like ComicBookPlus, these are sometimes incomplete, so it is necessary that we have our own copy. There are scans of many of our more recent books on sites we won’t link to, but these have their own challenges.

Take, for example, this scan of our corpus book #2840 (Exiles #55 (2005)).

Flipping through a scan, we found no creator credits, but, this being a 2005 Marvel comic, that seemed highly unlikely. Turning to our physical copy, however, we found a full set of credits (albeit a fairly complex one, as artist and writer roles are not distinguished, rather there is a joint author credit signified by the “By”)

What happened to those credits? I’m not sure, because I don’t know the source of this scan (is a pirate edition? a publisher-produced edition?). You can see that the credits haven’t been dropped in yet. The scroll effect is outlined here, and it will cover part of the drawing of a ship, but this page wasn’t final at the time it hit the web.

Moreover, as you can see below, the title of the story (which we are also indexing) has been eradicated in a way that suggests an unskilled worker with a copy of Adobe:

Of course, this isn’t a one-off. There are dozens of examples of this phenomenon out there relating to our corpus books.

Our takeaway? A scan of the comic book (whether by the publisher or by pirates) is not the comic book. Always, always check the comic book.

Now Hiring

(Please note that at the present time we are not able to hire additional Research Assistants as we have had more demand than we have work. If the situation changes in the future, we will update this page. Thank you for your interest - BB)

Given the alarming circumstances surrounding academic work for graduate students and contract faculty in the face of the COVID-19 pandemic, the research team behind the What Were Comics? project is offering a limited number of contracts for research assistants. The work can be completed remotely, of course.

Before we proceed any further let us be honest: this is not the most exciting work that our project has to offer. It is the non-mentally taxing work of counting panels and, later, words in our corpus. That said, the work can be done from home and is an ideal opportunity to be paid for research work while sitting on a couch while socially isolating.

What is the work?: We are hoping to find four or five research assistants to do manual hand counts of comic book panels and, later, comic book words (in captions and word balloons). The RAs will have access to PDFs of our corpus material through a shared Dropbox folder and will be asked to label the panels on the PDFs (see example below). In our experience, this is most efficiently performed on an iPad with an Apple Pencil and a PDF editor like GoodNotes, although other options would work. Please note that we are not able to provide any computer hardware or software subscriptions. So that means you have to have the tools to do the work at hand. I know this isn’t ideal but it’s what we can offer at this time. 

When can I start?: We are ready to set you up immediately and continue until the work is complete. RAs could work full or part-time and can invoice us weekly. The total number of hours will depend on the number of RAs, but we have currently budgeted approximately 3,000 hours for this work. Those hours are co-related with work completion goals. In other words, depending on how many hours you sign up for, we will assign you a work load level which we will check against your invoices.  

Who are your priority hires?: We are particularly hoping to fund graduate students or precariously/non employed recent graduates who are experiencing financial hardship due to COVID 19. As this research is funded by the Social Sciences and Humanities Research Council of Canada, preference will be given to Canadian citizens and landed residents of Canada, but we are willing and able to hire RAs from any location. 

How much are you paying?: We are able to offer a standard rate of $22 Canadian per hour for this work. Unfortunately, during the crisis and the resulting shock to global oil prices, the Canadian dollar has recently declined against major world currencies. Please check to determine the exchange in your own currency, but today the rate would be $15.17 USD per hour. We wish we could offer more. Also, you will have to complete the University of Calgary’s Electronic Transfer Form, and it includes your bank information (https://www.ucalgary.ca/finance/files/finance/scm-eft-form.pdf). Once you submit an invoice and we check it against your submitted completed work, we have to fill out an Outside Vendor Request, and the university will direct deposit the money into your bank account. While this is normally a routine process, given the number of University of Calgary employees working remotely it will take weeks before money shows up in your bank account. I should also note that if your government has a withholding tax agreement with Canada, the money will be withheld and there’s nothing we can do about it.

Money is nice, but how about recognition?: It has been - and continues - to be our policy that our RAs are our collaborators and that the corpus is also theirs to develop. If you have ideas for your own scholarship based on this work – lay it on us, we’d love to expand the knowledge and would love to collaborate with you more fully on this project.

Ok, my life long dream has always been to label the panels in randomly selected issues of Our Army At War, how do I sign up?: Click HERE to fill out an application form

I have a question!: Feel free to contact me at beaty@ucalgary.ca

Clean hands, clear heads, open hearts.

On Tom Spurgeon

“You could do far worse than to build a lifetime of friendships with the people you meet in comics.

Far, far worse.” — Tom Spurgeon

I don’t really know how to write about the impact that Tom Spurgeon had on my life. The influence that he had on the development of my writing was immense and arrived at critical junctures. When I look back over the past quarter century I can see a number of moments when things might have gone differently for me, when things could have gone less well, and at so many of those points Tom was there pushing me in the proper direction. It is astounding to me that he was only a year older than me - he always seemed to be so much more mature and insightful. I’ve always been jealous of his writing, and his energy, and his commitment to get work done, and I’ve tried, as best I can, to follow his example.

You can go on Twitter today and read hundreds - maybe thousands - of people tweeting about @comicsreporter. They are all variations of the same theme: Tom was the kindest person in the comics field, he was the one who reached out to new people and used whatever influence he had to bring them into the spotlight, he was the type of generous spirit that we should work to become.

Maybe let me try to flesh out what it was like to work with him.

Tom was the first person who ever paid me to write, not just about comics, but generally. In 1994 I was a grad student in Montreal who spent too much time bantering on an online mailing list (the fabled comix@ list) in between working on term papers. One day, almost twenty-five years ago to this month, Tom sent me an email that told me I was “wasting my time” writing in a closed venue, I should write for him at The Comics Journal. He sent me the first few issues of Zero Zero and told me to write about them and I got paid a whopping penny per word. Suddenly, I was a professional writer.

Tom was an easy editor for me to work for. He was enthusiastic about my writing, rarely offered suggestions, and let me do basically anything I wanted. After that first piece ran he asked me to pitch subjects for review essays, and in that first year I wrote about Tom Hart and Megan Kelso and Debbie Drechsler, all artists that I thought deserved a much wider recognition. He let me follow my own muse with that early work.

In late-1996 I pitched him, with a friend, the idea of a monthly column about comics in Europe. The Journal had never really covered European work in a systematic manner. We workshopped the idea for a while, and in early-1997 I submitted my half of the first column. When my co-author was late with his contribution, Tom would call me every day to demand to know where the rest of the column was, becoming increasingly irate about it. On the fifth day he said “This is your column now, you’re a soloist, write next month’s column by Monday” (which is why the second Euro-Comics for Beginners column was about one book by Baru, atypical of the series). A few months later my pay rate doubled and I was added to the masthead as a columnist. He never brought up the fact that he almost fired me ever again.

It seems strange to me now, but writing Euro-Comics for Beginners paid my way through grad school. I was in an underfunded program and on a fellowship that wasn’t sufficient to pay my bills. People have asked me why those columns were always so long, and the truth was that I wrote them until they were worth $100 at two cents per word, because the extra $1,200 per year kept me afloat. It also me made me a faster writer, and a better writer. I dreaded the idea that a TCJ would come out and be filled with columns better than my own. Tom and I would talk every month about the column (usually about the art choices - I would have to FedEx my copies of the books to the office for the art director), but he rarely asked for corrections or changes. Clarifications, mostly.

I almost quit The Comics Journal under his editorship when TCJ #200 came out. I knew that this was to be a blockbuster issue, and I did a special column for it in the form of a quiz. I worked much harder on that piece than on anything else I had ever done for the magazine. I wanted to “win” that issue by writing the most talked about essay. A few days before it shipped Tom phoned me. He had just discovered that an art director had taken my essay home to work on and forgotten about it - it wasn’t going to be included, the magazine was already printed. He told me he would, of course, pay the kill fee. Then he told me he’d double it. When I still wasn’t talking to him he asked me what I wanted. I told him he should send me the Complete Crumb Comics. Two days later FedEx delivered a box of those hardcovers, along with the sketchbooks. The note read “Fuck it, it’s only Gary’s money”

My column outlived Tom’s run as editor, but not by that long. I liked a lot of the post-Tom editors at TCJ, but not as much as I had liked working with Tom. By the time we had both gone I had a job and didn’t need the two cents per word nearly as badly. I converted most of the work that I did for the magazine into the foundation of my book Unpopular Culture. Looking back, I realize that every relationship that I have with any cartoonist working in Europe stems from his decision to run my column. That column gave me the excuse to work with hundreds of cartoonists and publishers, to get the word out about their work. The funny thing is that it was the weirdest possible column. Almost unthinkable. Tom let me write five or six thousand words per month about comics that weren’t even in translation - comics that were completely unavailable to most of his readership. Why would any sane editor do that? Well, Tom was just really, really personally curious about what was going on in Europe, and he thought other people should be as well.

Just yesterday I saw that Yvan Alagbé’s Yellow Negroes and Other Imaginary Creatures made the AV Club’s list of best comics of the 2010s. Tom let me write on Alagbé in 1998, twenty years before his breakthrough on this side of the Atlantic.

After I quit The Comics Journal and Tom had made a success of The Comics Reporter website I became anxious to return to writing about comics in a less demanding form. I emailed him from the Angoulême Festival one year in the early-2000s to ask his advice about blogging: how hard was it? How technical is it? He said that I should just send articles to him for a while to get used to it, then, he suggested, we would work on an online feud (we were both fans of professional wrestling) and I would split off to my own site, and the feud would drive traffic to both sites. This was, I thought, an excellent plan. I ended up writing for him on and off for the next decade or so and he kept the archive of that writing as a sidebar on the site for far longer than it deserved (indeed, my name is still there). 

This was the type of person Tom was: when The Comics Reporter won its first Eisner in 2010, Tom phoned me from San Diego to thank me for winning him an Eisner. I noted, correctly, that I contributed about half of one per cent of the content on that site and deserved none of the acclaim, but he said he was going to send me the trophy. He won a few more times over the years, and when he decided to decline further nominations he checked with me first, even though I was barely a part of the site at the time. “I just don’t want to turn down your Eisner,” he said. I told him, well, you still haven’t actually sent me the trophy, so you might as well. He said he had a bunch that he had collected for other people over the years and that maybe he’d send me one of those. I think he was going to send me a Chris Ware Best Lettering Eisner, but I never got that either.

Over the past few years Tom and I would talk by phone a lot less often. He would phone me if there was a major Euro-Comics story and I would give him background. We developed some mutual enemies that way. On the morning of the Charlie Hebdo massacre he was the one who broke the news to me, and he told me that he was referring all of his interview requests to me. Tom never wanted to be known as the expert, he wanted to facilitate expertise. He knew everyone, and he liked most everyone. One of my favourite nights at TCAF was simply sitting with him for about eight hours in the Marriott bar because everyone would come sit at our table for a while to talk to him. I think he pitched three different books for me that night with the line “You should get Bart to write a book about that for you. He’s your guy”.

That’s how it was with me and Tom - I would always just be happy to stand beside him and watch him talk to people. He was so generous with his time. He wanted people to have opportunities. He was never really someone that you could get gossip out of, because he never really seemed interested in tearing other people down. He wanted to build up the whole community. And that’s what he did.

If you didn’t know Tom, and didn’t get to work with Tom, and you want to understand him, I’m going to suggest you read the Christmas Carol run of Wildwood that he wrote in 2001 with Dan Wright. Notice the way that every character in this tragically short-lived strip has a unique voice, and how, even in the confines of the gag strip, they have strong personalities. Tom was a great comic writer because he listened to people and he cared about people, both are traits that you can see in his work. He was a Pastor Bobo figure putting people together and trusting them to do their best work. I’m eternally grateful that he reached out to a mouthy grad student on a mailing list and said “you can be better”. I try.

What is a Cover?

Our new scanner arrived last week, and yesterday we spent some time experimenting with it and testing it out. One of the things that got scanned was corpus book #0001, an issue of Gulf Funny Weekly from 1934. This is the physically largest comic book in our entire corpus, so it made for a good test of the machine. 

When I sent this image to Ben and Nick, Ben asked the obvious follow-up: How many comic books in our corpus don’t have covers?

What he meant by that is that GFW has a comics story on its first page, and that is counted as a one-page story (this issue has three one page stories and an ad in its four pages). A “cover” for our intents and purposes does not contain part of a story. While it might refer to one of the stories (as is often the case), it isn’t actually a part of it - it is a distinct formal element. 

Generally speaking, of course, magazines have covers, and books have covers. Newspapers, on the other hand, have front pages. It is not terribly surprising that GFW is more akin to a newspaper than to a magazine or book. Indeed, this issue is a newsprint page fold in half to make four pages. GFW seems like an anomaly in our corpus. We have three issues, and none of them have covers.

The other series that is similar are The Spirit supplements, which, of course, ran in newspapers. We have seven Spirit supplements in the corpus, and none has a “cover” - everyone of them simply begins the story on the first page, like so:

All of this would add up to the pretty simple observation that newspaper-aligned comic books don’t have covers, while magazine-aligned comic books (which are the vast majority of them) do. Were it not for this:

This is corpus book #0078, Silver Streak Comics #21, published by Lev Gleason in 1942. At first glance, this seemed to have a cover - like most comic books of the period there is a distinction between the glossy cover stock and the interior paper that suggests a magazine tradition. 

Closer inspection, however, troubles that conclusion. The fact is, the story featuring the Saint starts, as the text says, “here”. This is page one of the lead thirteen-page story (the book has eight stories in its 68 pages). However, unlike The Spirit supplements, there are some “cover” elements here beyond the paper stock. Notably, the “In This Issue!” Portion of the page acts like a traditional cover, drawing attention to interior features. From this standpoint, it seems to be both a cover and not a cover. It’s Schrödinger’s Cat Comics #21.

Just to add to our confusion, this is the next story page: 

Note that that pagination indicates that this page, the third, is the first page of the story, contradicting the “cover”.

So, it’s anomalous and we’ll ponder it. But then, what about this:

That is corpus book #1128. Some might argue that a two-panel comic stretches the definition of “story”, but for our purposes that is a distinct story. Indeed, that 68-page comic book contains a whopping 57 stories, none longer than a single page. Many of them even have their own titles!

I think it’s interesting (but not analytically defensible) that, on first pass, the issue of Jughead’s Jokes was coded as if it has a cover but Silver Streak Comics was coded as if it does not. There are additional Archie comics in the corpus that similarly have “covers” with stories on them - although very few other publishers did the same.

All this is a reminder that coding decisions matter and that consistency will be a major consideration as we move forward. I’ve already changed the “cover” of this Jughead’s Jokes to an other story - this issue now has 58 stories in it.

OMG. A Sampling Frame Error

A few people have asked, with an ever so slightly increasing level of annoyance, “when are you publishing your Sampling Frame?” or “Hey! Didn’t you promise to publish your Sampling Frame?” or, from the co-investigators, “Weren’t we supposed to make the Sampling Frame available?”.

Yes, we said we would. And we will.

Right now it’s a bit of a victim of my perfectionism. I live in constant fear that there are errors. Indeed, last week, while coding ads in comic books from the mid-1980s I realized we had the wrong issue of Teen Titans in our collection. That sent me into a minor frenzy of checking the Sampling Frame, to see if there was an error there that had led to an error in the corpus. It turns out that there was not - the error had been in the purchasing end, driven by clicking on the wrong Teen Titans series. Sloppy, but not a big problem (the correct issue has already been purchased).

Today, though, I learned of a genuine error.

I was reading Love on the Racks: A History of American Romance Comics by Michelle Nolan (recommended, by the way). In her introduction she notes that Michael Vassallo has noted that the stories originally intended for Atlas’s Love Tales #59 were actually printed in Lovers #42 (October 1952). Love Tales was cancelled with #58. However, when the title was brought back three years later the first new issue was #60. Apparently, because the stories for #59 had been paid for, the assumption was that it had been released. But it had not, it had been shifted. So there is no Love Tales #59.

The GCD has this information correct - their listing skips #59. MyComicShop, unfortunately for us, has a listing for #59 (though, obviously, no cover art). And Overstreet lists it as having been published. So, two of our three data sources had it, and we went with the majority, but the majority was in error.

None of this really impacts us at all. It does not change the number of comics sampled from 1953 (which is where we included it) and it wasn’t randomly selected (or this would have been caught earlier), although it does mean that our essay about the Sampling Frame has a minor inaccuracy, and that drives me crazy,

So, that is why we still haven’t published the Sampling Frame - I’m still desperately, probably vainly, trying to locate all the bugs.

Digital Dilemmas

The biggest news in comics publishing for 2018 snuck in just under the wire. Last Friday, Viz Media announced that Shonen Jump would be changing formats and moving to a day-and-date release for new chapters of their manga series in English translation and uploading them free on their website. Further, subscribers to Shonen Jump would be able to pay $1.99US/month for access to more than 10,000 chapters of older works. As the father of teenager who cannot stand the wait for new instalments of My Hero Academia, I can assure you that this announcement is a game changer.

Viz’s move is widely being read as an attack on scanlators - the fans who post unauthorized translations of manga online. Both the price (free!) and the timing of the releases (identical to publication time in Japan) will undermine scanlations, but, of course, Viz is also taking a massive risk with their existing revenue streams. Will my son still want paper copies of MHA if he can read the entire series from his iPad? It’s not entirely clear.

Ironically, the day before Viz’s announcement, I was talking at length with colleagues in our Communication and Media Studies department about the What Were Comics? project. One of my colleagues (correctly) noted that I had said extremely little about digital comics all day, opting to focus on traditionally printed comics. His question lingered with me as I read Viz’s news release, because it seems to point towards something that is so difficult in the study of contemporary comics: the absence of even decent data about the state of the industry.

Let me note that comics data has always been something of a problem. Sales and readership data from the first decades of the comic book industry are either absent or unreliable - even circulation statements are at least somewhat suspect. John Jackson Miller and Brian Hibbs both do remarkable work with the current sales data that is available to us, and both of them will remind you of dozens of caveats that come with their numbers. All of this, no matter how problematic it is, is far better than the data that we have for digital readership. This profile of Comixology, for example, gives absolutely no sense of its sales or subscriber levels - it simply tells us how many employees the Amazon subsidiary has; it’s a business profile that doesn’t want to tell us anything about the business. 

A pointed example of the limited data on digital comics has been the launch this fall of DC Universe, the hybrid streaming service for DC’s superhero tv shows, films, and comic books. Though the service launched in the United States in September, there has been surprisingly little coverage in the entertainment industry trades or even comics blogs about its take up rate. This article in Sensor Tower from the end of September noted that the app had been downloaded 140,000 times, but that only 33,000 people had paid the $7.99US monthly fee. Based on the adoption rates of other niche streaming services, it is clear that the largest surge in subscribers comes at launch, and then there is the long, slow slog to find and convert additional users. 33,000 seems to be to be a terrible number for a service backed by Warner Entertainment. Even if the number is higher today (and it likely is), that is a small fraction of the number of subscribers to UFC’s Fight Pass or the WWE Network, both of which are far more niche than Superman and Batman.

I should also note that it is nearly impossible for me to assess DC Universe on my own - it is geo-blocked in Canada. Yet another problem with digital delivery, but for another post.

Whenever we talk about What Were Comics? we are asked the question about digital, and it does trouble me. The two biggest stumbling blocks to expanding our study into this domain seem to me to be:

(1) It seems almost impossible to take a step back to the point where we can get a good idea of how much material is included in a field that we might call digital comics - webcomics, web series, twitter comics, fan fiction. The field is extremely vast and extremely decentralized and there is nothing even vaguely approximating a standard reference work akin to the Overstreet Guide. Finding a way to sample it all is daunting, to say the least.

(2) We are still working with extremely limited data because of the proprietary nature of these systems. Sensor Tower’s data is the exception (although we’d love an update!). As for other companies - just like Netflix refuses to provide the public within anything akin to ratings, the actual numbers of subscribers and what they read online is a mystery enshrouded in an enigma.

No matter how the Shonen Jump decision shakes out in the long run, it seems to me that they are taking the lead in digital delivery. As we witness a massive run-up in streaming competition over the next few years with the American broadcast networks and legacy movie studios each positioning themselves to take on Netflix, it will be interesting to watch comic book publishers make similar moves. Viz is offering a significantly more comprehensive back catalogue to its readers at a fraction of the cost of DC Unlimited. It will be interesting to see if or when Warner and Disney will be forced to react in kind by expanding their offerings or lowering their prices.

Yet for scholars and historians of this material, the closed nature of the decisions being made today is troubling. When we study the history of popular music, or television, or film our discussions are driven by data to at least a certain extent. At comics studies conferences, however, I frequently hear assertions that low-selling comic book series are “popular in digital”. These assertions are typically offered with absolutely no evidence. As comics publishing - which has never been up front about circulation in the first place - increasingly shrouds its business practices our work as researchers will become ever more challenging.

Endgame

Today we reached a major milestone when a FedEx truck pulled up at my door: The 3,563rd comic book from our corpus and we officially concluded the process of acquiring all of the comic books. This was an incredibly long and involved process. We began acquiring the corpus books in earnest about fourteen months ago and we purchased comic books from more than thirty dealers. We bought comics in retail stores, at conventions, online, by auction, and by old-fashioned honest-to-god catalogues. Our final books were purchased over the past month from The Beguiling, from Lone Star Comics, from ComicConnect, and from Ebay.

When we began this process, we understood that there was going to be a LOT of work to do. Our checklist basically looked like this:

• Write our coding protocols

• Establish our Sampling Frame (list of every comic book published 1934-2014)

• Randomize the Sampling Frame to construct our corpus

• Purchase all the comics in the corpus

• Produce/Acquire digital copies of all the comics in the corpus

• Build a tool for recording data

• Code the corpus at the Book level

• Code the corpus at the Story level

• Code the corpus at the Page level

• Code the corpus at the Panel level

• Code the Paratexts (ads, letters pages, editorial pages, etc)

We are now at the point where we have finished six of these eleven tasks. At the moment, we are coding all of the corpus books at the Paratext level, which should take the remainder of the academic year (we are almost half-way done this task). We are also working on producing digital copies of all of the books, and we have a research assistant tasked with this thankless task. We hope that this will also be wrapped up by the end of this school year.

That would leave us only with coding Stories (next year), Pages (summer 2020), and Panels (the 2020-21 school year), which, not so coincidentally, will carry us exactly to the termination of our initial funding period.

Frankly, I never really anticipated that we would get to this stage in our first round of funding, and we fully anticipated working from digital scans of some of our earliest (and, therefore, most expensive books). That would have been challenging for some of the work that scanners shy away from for copyright reasons (the Mickey Mouse Magazine that was in our final box, for example). This is a major milestone for us, but I have to admit that I will miss the active process of hunting for books at cons and shops - it always gave me something to do!

The final five.

#2133: X-Men #17

We are getting spectacularly close to having a complete set of all the books in our corpus (indeed, I’m expecting a FedEx shipment with some of the final books to arrive either today or tomorrow). But there’s a big caveat: I think we’re getting close.

As I mentioned to Ben a couple of weeks ago, we have to anticipate a certain number of mistakes that will necessitate re-purchasing. Due to the constant renumbering and relaunching of series, and the migration of titles between publishers, errors have crept in to our buying. Not a ton of them, but enough to be annoying. Fortunately, they haven’t been too expensive, and I was even able to trade a number of the comics that were purchased in error for credit on one of our costliest titles.

I do, nonetheless, live with a certain kind of fear about mistakes in purchasing - or, worse, mistakes in our sampling frame itself. This morning kicked out one of the latter.

Comic book #2133 is X-Men #17. This is the second X-Men series, launched in late-1991 with Jim Lee as the artist. Lee had already left for his own projects at Image by 1993; this issue is drawn by Andy Kubert. As I move through the coding that we’re doing right now, I’m going backwards in time and also backwards alphabetically through the years, so this was one of our first books from 1993. Anyway, I looked at the indicia to make sure that I was in 1993 and this is what I saw:

That evinced a little panic.

It didn’t seem plausible to me, because I remember the history of the launch of these titles and the founding of Image, and I couldn’t imagine that Kubert had taken over the title by February 1992. When I looked at the front cover things got worse. That 1962-1992 Spider-man box seemed to confirm that I’d bought the wrong X-Men #17. Plausible, because if you type “X-Men 17” into the search bar at MyComicShop you get a couple of dozen hits. However, their database claims that this issue is from 1993 - contrary to the front cover and the indicia. I began to think that they had to have it wrong down there in Texas.

So I looked at the GCD - where there’s an even greater number of X-Men #17s due to international editions. They too say that this is from 1993. They couldn’t both be wrong, surely?

Overstreet notes that X-Men #1 came out in October 1991, which made sense to me - I can remember the store I was in the day that it shipped (I recall a business man buying an entire box. 100 copies? 200? MyComicShop shows them to be worth $3.20 today a gain of 45 cents per issue in inflation adjusted dollars - almost two cents per year gain!),

Finally, the Standard Catalogue noted that it was cover-dated February 1993 but that the indicia was wrong.

So, we’re using all four sources to determine that the comic book is wrong about the comic book.

Facts are funny things in the history of this industry.

#2416: Vampirella Dracula the Centennial (1997)

Okay, this one is genuinely troubling from a coding perspective.

Our copy of this book has 76 pages. The problem is that 48 of these appear twice. The last 24 pages of this books seem to have been included twice. The GCD reports this as 52-page comic book, and, indeed, aligns with exactly what we have here, except that we have some of it twice.

I would call this an error and disregard the second 24 pages except for the fact that this is a squarebound comic and all of the binding looks exactly aligned, as if it was supposed to be a 76-page comic (which would make sense, as it is a special edition centennial thingy). So, we’re filing this under “I don’t know”.

The writers of the stories here, by the way, are Warren Ellis, Alan Moore, and James Robinson. So little seems to have been written in the scholarship about the overlap of Vertigo and Harris Comics….

#2418: Vermillion #12 (1997)

Strange one. This is the final issue of this series, part of DC’s late-1990s abortive Helix line, their attempt to do for science-fiction what Vertigo had successfully done for fantasy earlier in the decade. There is nothing noteworthy at first glance about this comic, except that our copy has the pages printed out of order. For a second, I even considered that it might’ve been done deliberately.

I just checked around, and it doesn’t seem to be a commonly reported incident, so I’m guessing that it is just this copy that is messed up, rather than the complete print run. According to the numbers at Chomichron, this issue sold only 5,669 copies. Even with these sad sales numbers, this title was one of the more successful books in the Helix imprint - only four titles made it to a twelfth issue (the most successful, obviously, was Transmetropolitan, which was shifted to Vertigo when Helix folded).

I’m not sure that the misprinting actually affects our coding of the book, so we probably won’t bother to replace it.

#2570: Marvel Selects Fantastic Four #2 (2000)

And here’s another weird one.

Following up on yesterday’s post about Marvel Selects Spider-man, we find ourselves with Marvel Selects Fantastic Four, also a reprint of material from the 1970s. Like the Spider-man book, this one has reprints of ads for companies that no longer exist.

What is more strange, though, is that when I was counting the story pages I miscounted. I had 20 pages for the main story, 6 for the back-up story, 6 pages of the weird ad reproductions, a cover, an editorial page, two external ads, and a back cover. Great. Add it all up and it’s a 37 page comic book.

Now, I miscount pages all the time. It’s trivially easy to do that, and you just start again to find the error. So, I recounted the pages, three times. I was driving myself crazy trying to figure out how I got an odd number, particularly given that the story pages are numbered, making the count a lot simpler.

Well, it turns out that this reprint of the first story, “The Monstrous Mystery of the Nega-Man” from Fantastic Four #108, skips from page 11 to 13, reducing the page count for the lead story to 19 pages and bringing us in at a nice, logical, divisible by four total of 36.

I then wondered: “what kind of half-assed reprint series skips a page of the comic book that they’re reprinting?” and I went to look at FF #108 to find out what got cut. The electronic version that I have, however, is no help at all - because that comic also jumps from page 11 to 13. Indeed, there is no story gap at all - the dialogue between Ben and Reed continues in a manner that suggests there was never a twelfth page.

It took another moment of looking to see the answer: Page 13 here also has a panel with the page 12 numbering. Clearly, in the original publication, both pages were half-story and half-advertising or editorial content, and they have been conjoined in the reprints.

A good reminder that the role of the coder in this project is to actually count the pages, and not simply rely on Marvel editors from the 1970s to do it for us! Sometimes a twenty page story is only nineteen pages long.

#2571: Marvel Selects Spider-man #3 (2000)

So this is a weird one.

Marvel Selects was a pair of strange series published in 2000, one title dedicated to Spider-man and one to the Fantastic Four. The comics featured reprints of stories from the 1970s, which was not exactly the peak material in the runs of these characters. Both series ran only six issues, so it doesn’t seem like the idea was a success.

What is weird, though, is that this comic features ads that are not ads.

Our coding protocol addresses three types of ads: external ads that are sold to some company that is not the publisher (a film studio, a video game company, a food company); in-house ads that advertise other products from the publisher itself (subscriptions, other forthcoming titles, trade paperbacks), and public service announcements. What it doesn’t count are what this book has: reproductions of ads that ran in the original comic book that are marked at the bottom as no longer valid. That is to say, they are ads for companies that no longer exist.

So, what are they?

Not an ad.

We have a category for editorial content, and the little notices on the bottom seem like that could make them qualify, but it’s not really what we mean by that term. We have a category for activities like puzzles, but that also doesn’t seem to fit. We have one for title pages, but that’s a no.

There are two categories that seem close. The first is pin-ups, which are images meant to stand alone outside of a narrative, which these do. The other is behind-the-scenes material. If we would want to make a case that the comic is being meta-reflexive about its own publishing history (which, to be fair, it actually is!) then this category might bend enough to accommodate these pieces.

Frankly, I’m not sure where I fall on this question yet, though I lean towards pin-up. If you have an opinion, we’re all ears.

There is, of course, always the dreaded “other”, but we’d prefer not to have to do that….

#3240: X-Men #2 (2010)

As we move through the process of counting pages, stories, and ads in the comic books in our corpus, Ben asked a pretty basic question: Given that every ad that shows up here will have a unique identifier, can we link those across books? So, for example, if an ad for the Green Arrow television series appears in fifteen different DC Comics, can we note that it is the same ad?

The answer is, of course, that we can, but it is going to be a lot of extra work. Right now we are noting when ads are internal (ads in Marvel Comics for other Marvel Comics) and external (ads in Marvel Comics for a videogame) and a few other features, but we’re not specifically dealing with the content of the ad itself. Adding that line to the database is trivially easy - it would then be the matching up that would be time-consuming, particularly in cases where ads have minor differences in them over time (Charles Atlas ads, for example). Still, it is certainly worth doing and it is the type of area where we’d like to expand our efforts if we have time and funding to do so.

One place where it would be trivially easy to do this work, however, is corpus book #3240, which has the same ad in it twice! The ad is an internal ad for the first issue of the third Wolverine series (2010). There is an ad for it on the inside front cover of #3240 and also near the back of the book. At first I thought I was imagining it (these things tend to blend together a bit), but here it is:

At first I figured that must’ve been a weird editorial glitch, but then the exact same thing happened in corpus book #3239 (also an X-Men title) and I realized that, no, no, Marvel really, really wanted you, the reader, to be aware that there was a new Wolverine series launching in September 2010 and that Wolverine was going to hell.

I suppose it worked. Wolverine #1 (3rd series) was the top-selling book of September 2010, with just over 100,000 copies sold. That didn’t sustain, of course, as the title has been cancelled and relaunched twice in the eight years since (and also re-numbered at least once).

In a piece for Flow, Ben asked the question of what a comic book reader was worth to advertisers. He concluded, not an awful lot. As we move through the 2010s in order to code the advertising in comics, I find more and more evidence to support his contention. This particular comic book has thirteen pages of non-comics material (of thirty-six total) and only four have been sold to external advertisers:

• Kids Headquarters, selling Marvel branded apparel

• MadEngine.com, selling Marvel t-shirts and other clothes

• AboveTheInfluence.com, a public service campaign of the ONDCP

• Got Milk?

All of the rest of the ads are for other Marvel comics, but most especially for Wolverine. In hell.

#3462: Suicide Squad #19 (2013)

With the return of the school year, we've started back on the process of coding our corpus. Right now we are doing a second run through all of our comics, focussing on paratexts (ads, letter pages, recap pages, credit pages) with a secondary focus on story length - which will set us up nicely for our third pass, which will focus on stories themselves. Combining these two elements, of course, also gives us comic book length as well. We've started with the most recent comic books and are moving backward in time. This means that the majority of our count so far has been 32 and 36-page comic books, with occasional forays into 52 pages and a very few 28-page books (plus one 200-page Jughead Digest!). 

The way that we're doing this now is crunching all of this by hand on index cards that are then inserted into the bag along with the comic book. We will add all of the information from the cards into the database as a second set, but we've already found that it is much faster to count and then record than it is to try to do it all at once (mostly because of tabbing backwards and forwards through the computer program). Also, the index cards give us a fallback if something catastrophic happens to our database.

For the most part, at least with contemporary comic books, the process of counting ads and story pages is very fast. I can do a regular comic book in about ninety seconds as a series of shorthand notations (c for cover, in for internal or house ad, ex for external ad, LC for letters page, and so on). Most elements in contemporary comics are full-page (I think I have only encountered one comic book with half-page ads so far). The somewhat less rare double-page elements get recorded as "in+in" or "ex+ex" to show them as one element across two pages. Again, this is all very quick.

Once the notes are on the pages, I do a quick double-check by adding the number of story pages with the number of paratext pages. This gives me a number divisible by four, and I know that I'm set. If it does not, I have to go through the comic book again to figure out where I made my error.

So imagine my frustration with #3462 (Suicide Squad #19), our first 38-page comic book. 

How 38, you ask? Good question.

As you can see above, Suicide Squad #19 has a gatefold cover, with the right hand portion tucked under the front cover. The reverse of these two pages is a house ad with a double-page spread. At first I counted this as a cover and an internal ad, but that seemed wrong. I thought that it was all just one cover, no matter how large, and so shouldn't be counted twice. But that made no sense, because in reality we're counting four pages on each piece of paper, but they are, of course, only one piece of paper. Then I circled the "c" to come back to it, but that made no sense either. So then I changed it to "c+c" and "in+in" to reflect the double-page spreads and allowed the final page count to stand at a non-standard count. 

So, the first exception to our rule of four.

#0052: Bob and Betty and Santa's Wishing Whistle (1941)

See that beautiful image up there? Yeah, well our copy looks nothing like that. Nothing, at all. 

When you set out to purchase 3,563 comic books with funds provided by the federal government one of the keys words is: economize. Whenever multiple copies of the same comic book were available to us, we always took the one in the worst shape, since that would be the least costly. I can't recall if we had multiple options on this one, but we could not have gotten one in much worst shape. Ours is missing its staples and it looks a lot like a dog has chewed the cover. The paper is so brittle I am loathe to even turn the pages. Handle with care!

The comic itself is fascinating. I'm not sure if this is the only comic book in the corpus that is wider than it is tall, but if not it is one of only a very few. As you can see by the cover, this was a Sears Roebuck give-away comic book from Christmas season 1941. I don't think Bob and Betty were an on-going pair of characters (there was a Betty and Bob series in Catholic Comics, published by Charlton later in the 1940s, but I don't know if they were the same duo).

The story is almost exactly as you would picture it - Santa takes a brother and sister to the North Pole for a visit and the get to see how all the toys are made. It is twelve pages long, with no ads and no credits of any kind. There is tiny printing notice on the back. One striking element is that the panels are all numbered, despite the fact that the layouts are extremely traditional. This is a holdover practice in a lot of comics targeted towards very young children in the period.

#0069: Eat Right To Work and Win (1942)

Following up from yesterday's notes about the Sampling Frame, I got asked: what is it?

The Sampling Frame is our document that lists every comic book published in the United States between 1934 and 2014 according to the criteria that we established for this project. It is just a huge list in an Excel spreadsheet of Title, Number, Year, and Publisher that allows us to distinguish 176,275 comic books without any duplication or omission (ideally).

To establish our Corpus, we randomly selected comic books from that Sampling Frame eighty-one times, once for each of the eighty years included. When we did the randomization, what we did was enter the number range of rows that represented a single year (for 1942 that range is 2985-4064) and then generated a randomized non-repeating list of integers representing two per cent of the numbers within that range. Then we highlighted the selected rows in the spreadsheet and moved on to the process of acquiring the books.

I'll be honest. I won't tell you that I wasn't occasionally tempted to falsify the selection. As I scrolled to the next integer through the Excel spreadsheet, I'd see that we had narrowly missed I book I found personally interesting by one or two numbers and I would think "who would even know if I just changed it?". But I never did. Random had to mean random. The proof, if you need it, is that the randomization process produced back-to-back five hundred page reprint books. If I were going to rig the process, the very first thing I would have changed would have been to remove that ridiculous amount of labour, trust me!

The occasional thrill of compiling our corpus, however, was when the randomization hit something I had become genuinely curious about during the construction of the Sampling Frame. Book number 0069 is one such instance. What, I wondered, could this comic, produced by the Office of Defense Health and Welfare Services during the Second World War, even be? Imagine my joy, then, when it was picked. And, again, after we found a copy after a long search.

It turns out to be really weird. It will undoubtedly wind up as one of the most anomalous books in the corpus. It is a sixteen-page nutrition guide that recommends eating an egg per day, and drinking a pint or more of milk per day as an adult. It is mostly typeset text, but at the top of most of the pages there is an original two-panel comic strip featuring the characters and artists from King Features (Blondie, Flash Gordon, Thimble Theatre, Brining Up Father...). 

This is the type of thing that most people would not classify as a comic book and which we classify that way because it appears in all three of our data sources (which is our pre-requisite). 

I'm now wondering what would happen to me if I switched to 1942's official diet? The dairy industry was big back then...

#0226: Terry and the Pirates [Popped Wheat Giveaway] (1947)

This is one of the most confusing books that we have in the corpus.

First of all, the whole "Promotional" section of The Overstreet Price Guide is a little bit confusing to begin with. Let's just stick with the Terry and the Pirates section. There are two entries, one for comic books published by Dell (1939-1953) and one for comic books published by Libby's Radio Premium (1941). The Dell section includes nine different titles, some of which seem, at first glance, to possibly be the same. For instance, they list a "Buster Brown Shoes giveaway (1938)" and a "Shoe Store giveaway" also from 1938. My first reaction was to wonder if those aren't the same thing? But since they are listed with different page counts, so probably they were different things.

This book, which got randomly selected into our corpus, is listed by Overstreet as "Popped Wheat Giveaway (1938) - Strip reprints in full color; Caniff-a". Well, that's what this is - a sixteen-page comic book reprinting Terry and the Pirates in color by Caniff with an ad for Popped Wheat on the back cover. So, good. Right?

Not so fast. The GCD lists this comic book as having been published in 1947 - not 1938 - alongside a whole host of other Popped Wheat giveaways. This has a great deal of credence, because, well, those books are advertised on the back of this one. Further, the GCD notes that this wasn't published by Dell, but by a company called Sig Feuchtwanger. Interestingly, the entry itself indicates that there is no publisher listed, but Sig Feuchtwanger is clearly indicated in our copy.

So, not Dell. But is it 1938? Well, the copy we have has a 1938 copyright in the indicia. But that is likely because the reprinted strips are from 1938 (and 1939). Given that this comic book contains comic strips from January 1939, it seems impossible to believe that it was published in 1938.

So, we have accepted the dating of this as 1947 and agree with the GCD, but we bought our copy from MyComicShop, who list it as appearing in 1938 (they have a couple for sale here, cheap!). Thus, we disagree with two of our three data sources on this one.

The confusion extends elsewhere. AbeBooks lists two copies for sale, one dated 1938 and one dated 1947 even though they are clearly the same comic book.

Anyway, this is just a taste of the kind of hand-sorting of the data that was necessary to construct our Sampling Frame. There were literally hundreds of comics like this that needed to be reconciled across our three data sources, very few of which then made it into the final corpus. When people used to sometimes ask me why this sampling process was taking so long and why didn't I just leave it all to the RAs, my mind would flash to comics like this one.