next »

[untenured]

Does a point exist where a researcher can over-rely on her research assistant?

For example, scholar X has no theoretical/empirical/whatever training and hires RA to handle that part of the paper.  RA runs the tests or whatnot and gives you the results.  X puts the results in the paper with no idea about how the empirical/theoretical/whatever was reached that was the basis for results.

Is this risky, unseemly, or a savvy use of an RA?

Untenured

[immigrant]

a and c!

[kohelet]

For example, scholar X has no theoretical/empirical/whatever training

I don't get it--how can a scholar have no theoretical/empirical/whatever training?  What else is left? 

To answer the question from my experience, though, as (formerly) one of a string of over-relied-upon RAs, this was THE key strategy that got my prof a job at a top program in our field.  Seriously.  I could suggest just about any quantitative technique, and she'd just say "go for it."  It was my job to keep all the BS she wrote somehow tied to the findings.  She's the master of "collaborating" with students.  And it probably helps that publication standards are somewhat lax in her niche of the field.  So in her case, c: savvy.

[untenured]

Those slashes were meant to signify "or" this "or" that.

So researchers do this? I expected a stronger negative response.

Untenured

[kissa_mau]

I can't say that I would author a paper that included data which essentially appeared from nowhere. Of course, I actually want to have faith that what I'm writing about is what I'm writing about. I'm also really picky about details and must know them all (within reason of course, but I have been accused of being a little OCD with data).

Now some complex papers can have pieces contributed by co-authors that not all the authors will necessarily understand to the nth degree, but that doesn't sound like what you're writing about.

[edwidge]

I expected a stronger negative response

I can supply a negative response, Untenured! I'm guessing that it's not all that uncommon, but I think it's bad news. It's one thing to have multiple authors working on different aspects of a manuscript, and to have the person with the greatest stats knowledge do the statistical analysis, with everyone involved 'on board' with the conceptualization of the problem, the decisions to be made re. the data, etc., but it seems like another thing (if the prof is the first or sole author) to be completely in the dark about how X was produced or arrived at, and be passing it off as your own work.

Setting the ethical issue aside, how would you be able to present the work at a conference, or address questions about it, or build on it with future work, or write proposals that rely on it, if you didn't have a reasonably firm understanding of every aspect of it?

I would feel like such a fraud--talk about Imposter Syndrome!

[kohelet]

In the situation I alluded to in my earlier post, the prof's willingness to "over rely" on me and a string of other RAs is very much a function of tenure track publication pressure--just like recycling papers, slicing up research into the smallest publishable units, obscuring substantively insignificant findings behind uninterpretable ordinal logits (or whatever--this is my current pet peeve).  The possibility of looking foolish at a conference because you can't interpret your own findings is an acceptable risk in the quest for tenure.

(Slashes = "or" --got it!)

[anthroboy]

As an example, some of my research requires statistical methods that I haven't used before.  If one of my GA's had used a certain statistical test in (for example) SPSS, I would happily enlist that GA to help code the data properly, run the analysis, and interpret the results.

It's not that I don't understand the logic of that particular statistic -- it's just that setting it up can be a pain.

[johnr]

The original question is a bit vague what with all the whatevers and whatnots, but in my field it absolutely essential to rely on research assistants (or other collaborators) to handle key parts of the analysis.  For example, I can spend all of my time operating an inductively coupled plasmometer (ICP) for the determination of metal concentrations in lake sediments, and never get tenure because it would literally take up all of my time. Or, I could hire qualified assistants to run and maintain the ICP.  It's my job to write the grant, design the experiment, hire the qualified assistants, interpret the overall results, write the papers, teach the classes and write the next grant.

[marichiweu]

I'd wager that it depends heavily on the field. The situation OP describes is probably fairly common in many sciences, where a faculty member has a lab full of students conducting their experiments for them. Some researchers will give their RAs pretty free rein and harvest the results. In my social science field, that level of participation by an RA would probably earn them a co-author slot, and a professor that depended that much on an RA w/o acknowledging them thusly would seem pretty shady. More info on field, OP? Are you asking hypothetically, or is there a situation you're concerned about? It seems like the kind of thing one might wish to inquire DISCREETLY about in one's department. How do other labs in your department work? Etc.

[untenured]

Thanks for the information and requests for further detail.  I want to target a social science journal with an empirical paper on Topic X.  I know topic X quite well.  However, my empirical training is quite poor.  I just don't have the ability to run empirical tests with proper controls and adduce results.  I've worked with empirically-trained co-authors and to good effect.

Graduate assistants are available that can run such tests.  Do researchers give general commands to GA's, such as "go see if X influences Y" and then receive results with no idea how that result was obtained?  Naturally, I'm concerned.

Untenured

[terpsichore]

One of the best skills you can develop as a professor is the ability to make effective use of an RA. And Graduate students should learn something while working for you; they are students, after all.  But you have to be patient; progress will be slower than doing research yourself.

Some questions to ask when you consider hiring the RA: Does the graduate student have the training you need, perhaps through classwork? If so, how well did the student perform in those classes? Is the student trustworthy in other respects?

Start the student on a small, well-defined project, where you can check his or her work.  Don't let the RA simply hand over completed work to you, but make the student explain his or her results. Ask questions, and when you are satisfied with the answers, move up to more complicated assignments. Build up slowly.

Working with an RA can be very satisfying as you watch the student develop skills and confidence while you also get research results. You are basically developing a new colleague.

[secretweapon]

My two cents:  I've seen excruciating research presentations by professors who clearly didn't understand what they were presenting, as their RAs cringed in the background.  If you're going to rely on your RA to bring in skills that you don't have, use it as an opportunity to learn those skills from the RA.