I would like to cite an email exchange that I had with the developer of the software program which was used to produce the results presented in my thesis. In this email exchange the developer explains a certain process that occurs within the software program. I need to mention this process vaguely to facilitate a certain argument in my discussion.

I recently came across a random citation guide (https://libguides.dixie.edu/c.php?g=57887&p=371731) which states that all personal communication (email/telephone conversation/letter) are not recoverable data and therefore should not be included on the reference list. Rather they should appear exclusively in the main text e.g., John Clark says… (personal communication, August 4, 2020) or (John Clark, personal communication, August 4, 2020).

However, in my experience I’ve come across various instances of private communications being cited (in a journal article) where it was treated like any other reference, i.e., the main text includes e.g., [1] and the bibliography e.g., [1] John Clark, Private Communication, August 4, 2020.

Not including private communication in the main text as recommended in the above citation guide makes a lot of sense as it kind of goes against the very purpose of the bibliography/reference list. However, having seen it done the other way in many instances has left me somewhat unsure. Does anybody know if there is a best/accepted practice for citing private/personal communication?

Many thanks,
Bob

I have seen private conversations in a reference list but it may vary between fields. If you unsure, choose a couple of respectable journals in your field and see how they do it. If the journal has previously let someone cite a previous conversation in a particular format, that should be a good enough standard for your work.

I looked into this recently for a paper, and wherever I found the answer (probably APA manual) told me to do this:

"We did X, which was considered appropriate as it
was just short of the length of the shortest sample (J. Bloggs, personal communication,
August 27, 2015)."

And it was not listed in the references at the end, just in the text.

If you are referring to actual data then the best place to check might be a recently published meta-analysis. Often authors contact researchers for unpublished data and include it in their meta-analysis (if it is a very thorough one that is trying to account for publication bias and include all possible relevant data). If you can find a paper like this you can just copy how they cite the data and then whether and how they put it in the reference list at the end. It shouldn't be hard to find one.

Hi Tudor_Queen

Thanks for looking into this. I'm not referring to actual data, it's an explanation of a basic functionality within a particular software that I’ve used. It’s information that you would expect to find in a professional/commercial grade software User's Guide. But given that this is an academic/open-source software, it comes as no surprise that the associated literature is quite poor!

The above approach you’ve mentioned is what I'm leaning towards. However, I've also emailed my library services team (which admittedly I should have done in the first instance) to determine if there is an accepted practice at my institution.

Cheers!

Thanks rewt,

I've contacted my library services at my intstituion to see if they have an accepted practice. If there isn't one then what you've menionted will probably be my next step.

Cheers!

It would be more 'normal' to add the name of the developer into acknowledgements, with brief detail on their contribution. The idea of a reference is something someone can find and read; it's a bit odd to make an email exchange publicly available (and if it's not, what's the point of referencing it). Whilst it's possible, you'd envision an email exchange to be referenced if it's something social-sciencey reflecting on context and perspectives; not because it explains how something works.

This does put it on you to convey how things work themselves, but if you're confident to publish you should be confident to describe how you got the results, so it should - hopefully - be straightforward.

The risk you run if you cite it as justification/explanation, is it being interpreted as 'we didn't know how the black box worked, but we got the explanation of the guy who made the box, and present it here as fact'. This may not be as problematic as it sounds, but is it probably the reason it's rare.

Of course I would acknowledge the developer of the software used to produce the results in my thesis. It goes without saying.



You only convey how things work to the point where you can fully justify the results in front of you.



Of course I wouldn't be citing the email as justification for why my results look the way they do. I understand how the 'black box' works pretty well. Did you read my post above? What I intend to cite is a simple functionality that one would typically find declared in the User's Guide. The literature for this software just happens to be quite vague. The information was left out inadvertently (as advised by the developer). Nonetheless it is important that I mention because it facilitates a certain discussion in my thesis, by no means does it replace the entire argument or justify any of my results.

Cheers!

Sorry if my OP sounded too critical,

Yes you're completely correct you only need to explain how things work to the extent the results are justified.

There's not a right answer in many cases since, taken to the nth degree, can we be sure SPSS gave us the right answer? Probably(?); and nobody's going to question that it did, since otherwise 90% of papers would be taken up on the inner workings of peripheral software.

There's still an intrinsic academic risk in assuming closed source stuff works as claimed (I know this from experience - and this might not be from malice on the part of the creator, but no software is bug free); no matter who claims it, it's still a risk. When presenting findings it needs to be a managed & communicated risk (and in the SPSS example, simply saying 'We used SPSS' is often sufficient, since if it's later found to have a bug it can be quickly related to the findings).

There's a lot of common sense there and consequent vagueness; if the software is e.g. integrating via a finite element method, that's a well-known thing and provided results are in the bounds of common sense it's unlikely to be questioned. If it's Facebook and they're claiming x algorithm doesn't discriminate because the developer says so it will be dug into. It's likely to fall somewhere in between (probably towards the former), in which case it's not a big deal, but I'd stand by the convention of explaining the algorithm used within the paper, and citing the developer's input in acknowledgements. The risk of citing email exchanges is you give reviewers something to criticise that would probably be a non-issue if omitted.