Monday, October 31, 2011

x% of Infinite

Those of us who got some "stimulus" funding (ARRA) in the form of a research grant a few years ago have to do quarterly reporting on the progress of this research, as opposed to annual reporting for a regular grant; in my case, the grant was from NSF. This quarterly reporting is not arduous; it just takes a few minutes of checking the info generated automatically (e.g., number of persons employed by the grant) and updating some text about research activities.

I have no idea who, if anyone, reads these quarterly reports. Unlike annual reports, which have to be approved by a program officer (at least at NSF, I don't know about other funding agencies), once these quarterly reports are done, they seem to be done.

The one part of the reporting that always makes me have to think a bit is the part where I have to report the % of the research that is completed. In some senses, of course, a research project is never really done. You can always do more, and then some more. Particular research questions lead to more research questions which lead to .. etc. That's one thing that is so great about research.

Realistically, though, a grant has a finite time span, so things do come to an end (financially). Therefore, I could answer the question in terms of "How much money is left in the grant?". That is a number that, in theory, can be determined well enough for reporting requirements, although never exactly at any given moment, given the complexity and vagaries of my university's accounting system.

Another way to phrase the question is:"What % of what you said you would do in the grant proposal is done at this time?" That's a tricky question for research projects that veer -- for scientifically valid reasons -- from what was proposed in the proposal. I have written about this before: my research group's grant proposals are our best guess for what we will do to solve the questions and problems posed, but, once the research starts and we get some results, we may find a different/better way to approach these problems, at least in some ways.

In that sense, the % completed of what we proposed to do may be a very different number from the % of the project that will be "completed" by the time the funding runs out.

I have been thinking about this over the last few years as I do my quarterly reports and have to select an answer from a pull-down menu with possible responses to the % completed question. For a long time, I was answering with low numbers because we were in the data-gathering stage (for well over half the duration of the grant), but once we were >>50% in terms of time and money left on the grant, I realized that the % completed should probably kick up a bit, even if we still have a lot more to do.

I suppose at some point I will have to say that the project is 100% completed; that is, when the money runs out, even if the research is not really completed. For accounting purposes, this is probably fine, but it will not be entirely true.


Stephan said...

You are fortunate not to have experience with grants from DARPA. My adviser has two DARPA grants, one with quarterly and the other with monthly reports.

Mary said...

In the corporate world you're supposed to break the project up into tasks, and assign a budget to each task. Your "percent complete" is the weighted average, percent complete of each task times fraction of the budget assigned to that task. If you need to change your list of tasks or your budgetting during the project, that's called a replan. You just have to submit your new plan along with your monthly or quarterly progress report and make it clear what has changed. If the total dollar amount changes that is "re-baselining" and requires approval. You are also supposed to make a schedule showing when these tasks should be completed and which ones depend on which prior ones. Somewhere on the schedule you put dates for key "milestones" you think must be met along the way. You can change the schedule as long as you report it, but if you miss a milestone or move the date of a milestone, that requires approval. This is how we organize our DARPA projects too, but I don't know if they expect it of academics. It actually gets a lot more complicated that this, with this raw data being used to calculate all kinds of schedule and budget performance metrics. You are barely dipping your toe in the ocean of "project management" with this post. :-) No matter how rigorous you try to be, "percent complete" still feels arbitrary wit a research project.

Anonymous said...

Today's PhD comic seems relevant:

Anonymous said...

The disconnect between the various definitions of completion is like the various definitions of "effort" by bean-counters, funding agencies, PIs etc.

Anonymous said...

I think % of money left is what they actually want to know for ARRA. After all ARRA is about stimulating the ecnomy. Therefore, they want to know the following, really: 1) How many "jobs" have been created (e.g. if you have money for 4 person-years of funding, they want to know how much of that you've spent). 2) How much economic activity have you generated (e.g. by buying supplies/ordering services). So the % of money left is what I would report in your situation.

John said...

Seems like the question "What % of your life is completed so far?"

Anonymous said...

Interesting - all I have to do for our ARRA grants is write up the summary of research in 2000 characters or less. No percent complete, nothing on expenditures, just the narrative.

I wonder if different institutions are differently interpreting the reporting mandate, or if different funding agencies had different requirements.

Anonymous said...

For an ARRA grant, the "right" answer is that the % of the project completed is approximately equal to the % of the funding that has been spent. This question emanates from Earned Value Project Management, which was developed for construction activities. Remember that ARRA funding was intended mainly to be for "infrastructure building" (in a literal, not a metaphorical sense).