In February I posted a short survey on ISWorld about a requirement set forth by the publisher of Journal X that articles published in that Journal should contain at least 5 references to previous articles in that  Journal..  This technique is known as self-referencing. I asked for yes-no answers to three questions:

Question 1:  Is this practice common?

Question 2: Is it appropriate

Question 3: Is it ethical?

The result was overwhelming: 126 responses, almost all within the first 48 hours on a holiday weekend (President's day in the United States) . Furthermore, 97 of the respondents added substantive comments. Obviously, the questions hit a nerve in the community.

This page tabulates the responses and discuss them in broad terms.  In addition, The 97 comments are posted at the following URL: __________________. All responses are anonymized.  Personally, I found the responses thoughtful and raising many issues beyond the three simple yes-no questions. They are worth your time reading. 

The following table summarizes the responses numerically:

Although the questions were asked in yes-no form, some people chose to write text only. That resulted in having either to interpret  the answser or create the additional columns.  Note that in many case I interpreted the text as a yes or no. Some additional categories were added, as explained next. 

1. The response mode was NO to all three questions.

2. Not everyone answered every question. That explains the approximately 10% in the "did not answer specifically" column.

3. For "Is it common?" the usual answer was No. Of the 31 who said Yes, 18 described experiences in which they were asked to add journal self-references.  Some said may,be 11 didn't know, and 6 said the incidence was real but low.

4. The ethical issue also showed strong consensus.  The interesting answers are the ones that said yes, maybe, and questionable or debatable.

Summary of text responses:

Responses came from senior and junior people in our field, including a few non-academics and graduate students.  They came from all three AIS regions.  

1. Consensus

The consensus is remarkable. People believe that the choice of references is theirs, not the journal's. There was agreement that it is appropriate for a reviewer to suggest additional references but that they need to tread a fine line in what they recommend, particularly that they don't ask that they themselves receive multiple reference (thereby increasing the reviewer's impact factor)  or that lots of articles in the journal be  added to the reference list.

2. Marketing Rationales

-ISI Journal Rankings: Almost universally, people talked about the impacts of ISI Journal Rankings and attributed the request for 5 journal self-references to a desire to boost their impact factors.  The assumption was that the request for journal self-reference was made in the belief that the ISI ratings (now owned by Thompson) credited journal self-references.  Two of the respondents pointed to literature that indicates that this is no longer the case, if it ever was.

- Others pointed out that adding references to the journal is designed to show readers of a given article that the journal contains many items of interest to these readers and  therefore makes it more likely that they or their institutions will subscribe and that the readers will send good articles to the journal.

3. Relevance of references. There was agreement that it is appropriate for a reviewer to suggest additional references but that they need to tread a fine line in what they recommend, particularly that they don't ask that they themselves receive multiple reference (thereby increasing the reviewer's impact factor)  or that lots of articles in the journal be  added to the reference list.

4. Innovation vs. Publishing what they have published

Some editors and reviewers defended the policy of journal self-reference as making sure that the article is appropriate for the journal. That principle ensures that journals publish material that extends or elaborates material that they published previously.  It says the journal is not interested in publishing any new topics or to report innovation. The implication of such a policy is that by simply counting the journal self-references, a quick decision can be made as to whether to consider a paper at all. More important, the policy  makes It makes it quite difficult for people with something new to say to be heard by shutting journals off to ground-breaking research and innovation. In a dynamic field, such as IS, it perpetuates what is and eventually creates a dull journal.

5. Fear and loathing. Reading the examples of personal experience reported, there is a tone of fear and loathing of the publishers and editors who have such policies.  One reader plaintively said: " It happened to me once with a fairly new electronic journal. It made me feel very uncomfortable, but I have to admit that I complied with the request...I felt very torn by this for a long time (and the feeling is back while just thinking about it). Strong enough for me to send this message anonymously.

Another reader said: "In general, commercial publishers like Springer are becoming increasingly unpleasant." 

6. The AIS Code of Research Conduct. . The self-referencing proposal does not match the AIS code of ethics. As one respondent put it:

I believe this request contravenes the AIS Code of Research Conduct. 

See http://home.aisnet.org/displaycommon.cfm?an=1&subarticlenbr=15

Specifically, item 5 states

"5. Do not abuse the authority and responsibility you have been given  as an editor, reviewer or supervisor, and take care to ensure that no  personal relationship will result in a situation that might interfere  with your objective judgment."

Forcing authors to cite additional papers places the authors in a  quandary with respect to item 2 of the code, which states

"2. Do not fabricate or falsify data, research procedures, or data  analysis."

7. One respondent argued that the questions may inadvertently  be "loaded and biased".  That may well be true. They were the thoughts that came to me when I read the publisher's request. Readers should take the possibility of bias into account

Final Remarks

What came through in the responses is that the journal self-reference phenomenon exists today but at a low level.  From the experiences quoted, it appears almost exclusively in commercially published journals  that are struggling for continuing  existence. Not all publishers are involved. One publisher's book editor followed up with senior people in her organization and found no evidence of the problem.

Clearly, feedback is needed to the publishers and editors that mandatory self-reference is not considered appropriate or ethical by our community.

My thanks to all of you who responded.