About Chandra Archive Proposer Instruments & Calibration Newsletters Data Analysis HelpDesk Calibration Database NASA Archives & Centers Chandra Science Links

Bibliography Statistics - Notes and Caveats


Papers Included

Only papers already published in refereed journals that present specific data from Chandra observations are included in these plots.

Statistical Significance and Apparent Trends

Many apparent spikes and trends in many plots are due simply to small-number statistics. The plot values and corresponding sample size are listed in the data table linked below each plot. For plots using average values, e.g. all 'per ksec' plots, the mean is used in the plots while the median and standard deviation of the distribution are included in the corresponding data file.

Trends With Time

Several temporal factors affect the apparent statistics of publications. Only programs whose observations are complete are included, so programs in the latest Cycle are underrepresented. The introduction (Cycle 3) of Large and (Cycle 5) Very Large Programs decreased the total number of programs observed. The unavoidable delay between completion of an observation and publication of resulting papers causes a downward trend for papers (or citations) per program as Cycle increases. This of course does not imply that the number of Chandra publications is decreasing, only that fewer papers have so far been published based on data from the most recent Cycles. DDT observations have shorter proprietary times and so may publish more quickly. Long observations or large datasets generally take more time to complete and analyze.

Proposal Types

CAL, DDT GO, GTO, TOO are all mutually exclusive definitions in the Chandra databases. So for instance, a TOO from a General Observer is not also counted as a GO. Program Types LP and VLP when shown, are those proposals defined and peer-reviewed specifically as LP and VLP.


Proposal Types

Large Project (LP) programs were only begun in Cycle 3. Very Large Project (VLP) programs were only begun in Cycle 5. Therefore, comparisons of V/LP programs to normal GO programs by program type can only be made for very recent Cycles, for which statistics are still poor. In addition, it has always been optional whether to designate a 1Msec/300ksec or larger prgram as a V/LP. Plots that cover arrayed by ksec may thus be more informative, but since V/LPs do undergo a different review process, we also provide plots by program type, separated by Cycle.

Judging Science Return

Many plots, e.g., "Papers per ksec by time" seem to indicate that short exposures provide a larger science return per ksec. However, the relative productivity (papers) and impact (citations) of larger projects appears to have a longer latency. See for instance "Ratio of large-project to small-project papers per ksec" and "Ratio of large-project to small-project citations per ksec".

What Defines A Program

Data or plots referring to programs simply tally ObsIDs together that are from the same approved proposal in a single proposal Cycle. ObsIDs grouped by science objective (probably but not necessarily with the same PI) might be a better definition. To that end, we expect to eventually identify 'datasets' as distinct countable categories, for instance the Lockman Hole, Chandra Deep Field North, M101, ... Datasets defined for archive data retrieval will be different (a superset) of those defined for bibliographic statistics.

The Chandra X-Ray Center (CXC) is operated for NASA by the Smithsonian Astrophysical Observatory.
60 Garden Street, Cambridge, MA 02138 USA.    Email: cxcweb@head.cfa.harvard.edu
Smithsonian Institution, Copyright © 1998-2004. All rights reserved.

Last modified: 04/27/06