contracts and grants to universities fall within the broad category of net space research, the principal focus of the task group's study. This estimating approach was applied consistently for each of the 4 fiscal years for which detailed data were collected. These data suggest that over the decade from FY 1986 to 1995, the total number of NASA awards to universities and colleges for the performance of space research increased by about 1,000—from 1,631 awards in FY 1986 to 2,645 awards in FY 1995. Awards for purposes other than research increased more rapidly during the decade, with the result that the proportion of awards for research declined from about 58 percent of the total in FY 1986 to 52 percent in FY 1995.

Many task group members were concerned that the use of simple average statistics (especially the use of mean values) would not give an accurate sense of the variability or the typical value of award sizes, especially since these distributions are known to include relatively small numbers of very large awards and relatively large numbers of small-dollar-value awards. To address similar concerns, statistics on new versus continuing awards were tabulated to allow for the fact that some awards continue for very long periods whereas other do not. During a period in which the total number of awards is increasing, the average duration of awards tends to decline because of the varying proportions of new versus continuing awards.

In the process of developing data for analysis, the task group created a lengthy listing of large NASA awards that were then classified on an individual basis for purposes of this study. This listing, sorted by the major classification codes used to generate many of the statistical series reported in the text of this report, provides a basis for assessing the validity of the coding scheme used by the task group both for correlating activity types with the various awards and for assigning them to the major NASA science disciplines.