CSR SCORING PROCEDURES

Each scored grant application is assigned a single, global score that reflects the overall impact that the project could have on the field based on consideration of the five review criteria (significance, approach, innovation, investigator, and environment), with the emphasis on each criterion varying from one application to another, depending on the nature of the application and its relative strengths. The best possible priority score is 100 and the worst is 500. Individual reviewers mark scores to two significant figures, e.g., 2.2, and the individual scores are averaged and then multiplied by 100 to yield a single overall score for each scored application, e.g., 253. Abstaining members and those not present during the discussion do not assign a numerical rating and are not counted in calculating the average of the individual ratings. For research applications, reviewers are also asked to recommend that half the applications not be scored and to spread final scores to achieve a median score of 300. (Any member of the scientific review group may request that an application be scored, in which case all members must score the application.) To the extent that the study section does not score some applications, the scoring range is altered. If half of the applications are not scored, then the remaining applications should be scored from 100-300. If only 25% of the applications are not scored then the remaining applications should be scored from 100-400. Note that these procedures for scoring only half of the research applications do not apply to fellowships and career applications; all fellowships and career applications are scored.

PERCENTILE CONVERSION

Research grant applications (R01s) reviewed in CSR study sections are assigned a percentile rank. The conversion of priority scores to percentile rankings is based on scores assigned to applications reviewed during the current plus past two review rounds. Applications reviewed by a standing study section are percentiled against all applications reviewed by that same study section for the three consecutive rounds. Applications reviewed by Special Emphasis Panels (SEPs) are percentiled against the parent study section database if at least 30% of the reviewers are current or recent (during the last 2 years) regular members of that study section. Applications reviewed by SEPs where fewer than 30% of the reviewers are current or recent members of a standing study section are given a percentile based on the distribution of scores assigned by all CSR study sections. Note that at CSR, applications other than R01s (e.g., fellowships, small business applications) are not percentiled.