Graphing how long it takes for us to make a selection
Last year when we had finished making selections from proposals submitted in response to ROSES 2007, we released the bar graph and waterfall plots showing that we had made selections faster for ROSES 2007 than we had for prior years' ROSES. You can still find those old plots on the bottom of this page.
Now that we have completed the selections for all proposals submitted via ROSES 2008, I can compare how long selections took this past year to prior years. In the first figure that appears below we represent how many days elapsed from due date to selections for ROSES 2003 through 2008 as a "waterfall". Each year the general shape of the curves are the same, it starts at the top with no selections made, and by the time all have been made the curve has reached the bottom. The the faster the decisions are made the further to the left the curve reaches the X axis. The drop for each selection is proportional to the size of the program. Whereas in 2003-2006 the mid point was at 200 days or more, and in 2007 the midpoint of the curve was approximately 180 days, for ROSES 2008 it was 160 days That means that on average proposers heard 20 days faster than last year, and much sooner than in prior years. Similarly, whereas the slowest program in the past varied from over 700 to roughly 350 days, for ROSES 2008 the longest delay was 276 days.
For a simpler representation we have a bar graph where each program is treated as equivalent, and put into a bin, depending on how much time it took for the decision to be made. For the new bar graph including ROSES 2008 I decided to just use data from the last three years so its not too busy, see directly below. For last year's bar graph that goes back to 2003, see the multi-colored bar graph at the bottom of this page. Again, this improvement in the time is the direct result of hard work by our program officers.