Psychology of Space Exploration
Additional datasets compiled for use in analyses were 1) lists of areas of known
geographic interest to crews based on publicly released biographical information,
2) orbital track parameters to relate images taken to the log of scientific requests
sent to the crew, and 3) records of on-orbit activities to determine the incidence
of EVAs, the docking of visiting vehicles, and days of light duty/holidays. We used
the orbit tracks and message logs to identify which photographs were in response
to CEO requests and which were self-initiated by the crew. Occasionally, battery
changes and camera resets were conducted on orbit without resetting the date and
time on the camera. Because of this, not all camera time stamps were accurate. We
screened those data for inaccuracies (such as an incorrect year for a specific expedi-
tion), and these records were eliminated from the analyses.
For each day, we determined the number of images of Earth that were self-
initiated, were of areas of known geographic interest to any member of that
crew, were in response to a scientific request, and used the 800-millimeter (high-
magnification) lens setup. The use of the 800-millimeter lens was tracked because
it represents a significant skill that requires much effort to achieve the best results,
and the resulting images provide the most detail (up to 6-meter spatial resolu-
tion). The crewmembers must practice tracking the motion of Earth beneath
the ISS using the camera equipped with the 800-millimeter lens and learn how
to focus properly through the lens.^21 Although this was not one of our original
hypotheses, we realized that use of the 800-millimeter lens could be an indica-
tor of crew interest in Earth photography as a challenging, self-motivated hobby.
In general terms, the analyses looked for relationships between self-initiated
image-taking and when the images were taken, as well as between self-initiated
image-taking and the geographic location of those images. For the benefit of sta-
tistically minded readers, hypotheses 1 and 2 were addressed by examining zero
order correlations and using general linear models in a statistical analysis package
(GLIMMIX [generalized linear mixed models] procedure in SAS). This procedure
fits generalized linear mixed models to the data and allows for normally distributed
(Gaussian) random effects.^22 Hypothesis 3 was tested using a related procedure that
could incorporate categorical data into the model (GENMOD procedure for gen-
- Robinson and Evans, “Space Station Allows Remote Sensing of Earth to Within Six Meters”: 185.
- SAS, The GLIMMIX Procedure (Cary, NC: SAS Institute Inc., June 2006), p. 5, available
at http://support.sas.com/rnd/app/papers/glimmix.pdf (accessed 5 May 2006).