Standard operating procedures for a statistical consulting center (created 2010-07-30).
This page is moving to a new website.
I asked a question on one of the American Statistical Association message boards about how I setting up a consulting service at the University of Missouri-Kansas City (UMKC), where I work part-time. I wanted to develop some SOPs (Standard Operating Procedures) for this center that would supplement the guidance already available on the web. I asked if anyone else had SOPs (or anything similar) that I could look at so I wouldn't re-invent the wheel. I got a lot of responses.
Here's my original query.
I am an independent consultant, but I'm also a part-time employee at a university that is starting up a statistical consulting center. I'm helping out with this and wanted to develop some SOPs (Standard Operating Procedures) for things like insuring that we don't analyze data from a non-IRB approved study in humans. I'm also interested in a "customer satisfaction" survey that could be sent out afterward to give us feedback on how to improve. Does anyone out there have some SOPs from their academic statistical consulting center? Anything else that you can share easily so I don't have to reinvent the wheel would be greatly appreciated.
Here are some of the links that people provided. In some cases I included a slightly different link, as I wanted something corresponding to the home page of the consulting center. In other cases, I included a second link if I thought it was helpful.
I will take the time to carefully review all of these sites (as well as the ones listed through Michigan State University). Eventually those ideas will be synthesized and reproduced at this site, and (if the power-that-be approve) on the official page at UMKC:
There were other general comments and advice that I cannot easily summarize here, but I'll try. One person pointed out that asking for IRB approval would not make sense for many research projects, a point I agree with but did not make clear in my original email request.
Another person suggested that in addition to documenting IRB approval, the researcher should provide some evidence of informed consent.
In order to ensure full ethical integrity, the statistician requires that the sponsor provide proof that all patients enrolled provided appropriate consent (e.g., a letter confirming results of monitoring consents, a letter from each site's institutional review board [IRB], or a source document suitable for audit). Signature pages of consent forms are not to be submitted, as the statistician must never see patient names or initials. The results of statistical analyses will only be provided to the sponsor once these assurances are in place.
This is exactly the sort of thing that I want to implement, though I'm not sure what form it will take.
Someone asked how the consulting center is being staffed.
How is it being staffed? For example, will it be by people who already have the appropriate degrees and/or experience, or will it be by (under?)graduate students whom you'll oversee?
That certainly would make a difference in how the SOPs would be written. Right now the staffing is somewhat thin: myself, who is only 25% time, and about 25% of another person's time. We have others that we can draw on for help, but for most projects it has just been us two. We hope to expand, of course, but the plans for expansion are pretty much dependent on others and are difficult to predict. But both of us have graduate degrees and significant consulting experience.
Another person discussed the difficulty in properly getting feedback after a project is done.
"In terms of a satisfaction survey, we have been using SurveyMonkey to send out a short survey (but we still need to overhaul the process to find a way to improve the response rate). Recently, I've been focused on upgrading our client tracking system to better track and report on our activities. I think it's essential to have a good database to track cases, effort, and scholarly products such as the grant proposals you've helped with, the funded grants you have supported, the citations for publications and presentations you have supported and/or co-authored, and so on."
This is a big issue for me. I want to get good feedback, especially from people who don't come back a second time. So a good post-consulting survey and a decent client tracking system are very important.
I also got a request from someone else who had a small group just starting to share anything I found out. This page is my effort to share what I have learned.
I thank everyone who took the time to offer comments.