Substantive Changes to Texts
One change that authors might want to make to their published work would be to add new or supplemental material, possibly because they wish to share data gleaned from new research, enhance their original presentation with graphics, or even explain an argument or train of thought more fully. Question #7 asked editors to consider whether this practice should be allowed.
Editors overwhelmingly rejected this possibility, though a few felt that that there could, conceivably, be reasons to permit such additions under exceptional circumstances. “This would occur only if there were a technical error for a file or image that represented the data,” said one, while another remarked “The author(s) would have to have a REALLY good reason to make these changes,” adding “I think this would be a headache to do.”
In question 8, editors also rejected the idea of deleting or replacing existing text, and to an even higher degree (93.3%) than they did the addition of material. Only one journal indicated it would permit these changes, and even then only with errata. One reason for the stronger degree of rejection here, I think, has to do with a concern for editorial transparency. While adding material might alter some aspects of a document, none of the original content would be excised; it would still be available for readers to view. Deleting or replacing text, on the other hand, would remove content from the manuscript and might therefore be perceived as a more significant (and slightly suspicious) change.
Responses to question 9, regarding revisions to an article’s central claims, arguments, and/or conclusions, were identical to those in question 8. As a group, the majority of journal editors indicated they would not allow authors to revise texts in ways that substantively affected or changed the original meanings. Even so, one respondent (in a comment added to question 10) suggested he had at least some openness to changes that significantly altered a text’s claims and/or conclusions, saying “If there were a truly substantive revision, I would probably replace the original text with one labeled "version 2.0" or similar (with a revision date included).”
Finally, question 10 asked editors to provide a context and/or justifications for their earlier responses, identifying one or more reasons that informed their perspective – and consequent editorial policy – on post-publication revisions.
This question served multiple purposes: (1) it asked editors about general beliefs on the issue that could be correlated with responses to earlier questions in the survey, (2) it asked about their sense of responsibility to the discipline, to archival integrity, and to the journal itself, and (3) it asked whether digital journals should follow the practices of print journals or find ways to develop their own frameworks for publication and revision.
Four of the choices (#s 1, 8, 9, and 10 in the list) asked respondents about their general beliefs regarding edits to published manuscripts. Did they think that no edits should be allowed, that all edits should be allowed, that they should only be allowed in some situations, and/or that trivial edits could be made without documentation? Because these choices reflected perspectives that probably shaped many of the editors’ responses to earlier questions, it was reassuring to see that the number of people who agreed with each of these statements roughly corresponded with the positions and numbers that were collected earlier in the survey (e.g., 6 people said here that no changes of any sort should be allowed, and the answers to questions 3-6 concerning even relatively minor text revisions showed, similarly, that 5-8 people would not allow them ). In the comments section, several editors elaborated on their positions in ways that were relevant to these four choices:
Any changes should be done by publishing an errata that can be linked to the original article. That way the integrity of the original publication remain intact, but appropriate corrections are added to the literature.
We should strive to have as correct a text as possible. In the event that there are typos (misspelled names) or factual errors (like dates), these things should be able to be corrected for accuracy’s sake. I am less supportive of changing the fundamental meaning of a published manuscript, but I have learned to never say never. There may be some instances in which such changes are warranted, but they should be documented as changes from the original.
We make changes to minor errors as human labor/resources allow … The only "major" changes we've made after the fact in the last 10 years (that I can recall) is to pull a student's example when the student claims, 5 years after publication (googled themselves), not to have given permission for the piece. And, when we discovered post-pub that an author self-plagiarised by co-publishing a piece in two journals at once. We put a big note on that one after the fact, but we don't announce errata in the editor’s column.
None of the editors participating in this survey was willing to agree with the position that “Knowledge-making is always a work in progress, so authors should be permitted to revise their published work in ways that reflect their current understandings,” but one did note in a comment that,
In theory, the final reference to "knowledge-making" being a work in progress is quite true, and it would be great to have authors update their work. In practice, from an editorial standpoint, this could be a logistical nightmare. I think what it may call for is a type of interface, such as a blog/wikispace, where both authors and readers can dialogue about more recent scholarship, extending original ideas represented in the webtext, etc.
Relatively few editors, only two in each case, indicated that a sense of disciplinary obligation or anxiety about possible legal consequences played a significant role in their decisions to make (or not make) edits to published texts. An equally small number felt any concern about the effect post-publication edits would have on their journal’s standing in the field. A slightly higher number (4) did feel that they had some responsibility as editors to safeguard archival integrity by documenting “any and all changes,” but it would be unreasonable to conclude from this that the other twelve editors did not believe archival integrity was important. More likely, this is a consequence of my using the phrase “any and all” in the question, which prevented the seven editors who felt minor citation edits could be made without errata (question 5) from agreeing. One editor, however, did feel compelled to reflect on the potential dangers of historical revisionism in academic research, a concern that many other scholars may share:
Academic journals are not like Wikipedia. Preservation of the historical record in scholarly publication is important to the creation of new knowledge, because arguments and proofs are built on the research conducted by earlier scholars. If those publications can be revised, research that references that work becomes unreliable.
The fundamental tensions between print culture and digital culture are teased out in the positions articulated in choices 4 and 9:
Online journals should emulate the practices of print journals in regard to edits and revisions.
Online journals should develop new policies for textual edits and revisions that reflect the possibilities of the new medium.
As might be expected, the three editors who said online journals should follow the policies and practices of print journals were also in the group of respondents who said no edits of any sort should be allowed once an article has been published, and their individual responses to all of the other questions in the survey confirmed this conviction. Though I suspected, at first, that these answers had come from editors whose online journals had companion print publications, that proved not to be the case. Only one of the three journals existed both in print and online.
Nine of the editors (60%) indicated that online journals need to develop editing policies in keeping with the digital spaces they occupy, but none offered any specific suggestions or comments as to what those policies might be. Though there was clear resistance to the idea that online publishing should be held to (or confined by) the strictures of print technology, the specific direction(s) journal editors should follow to create and implement change – so far as this survey could reveal, at least – remain uncertain.
 The drop from 6 to 5 in some questions is likely due to their having been skipped by some respondents.