Logistic regression analysis was carried out in order to examine
potential relationships between external illuminances and lighting
status (i.e. lighting is on or off). Logistic regression analysis has been
widely used to investigate the effects of environmental stimuli on
the use of environmental control systems [13,15–16].
Logistic regression results imply that there was no relationship
between external illuminance and the use of lighting. All regression
coefficients for external illuminances in the logistic regressionmodels for the investigated offices were zero with standard error
of zero. This indicates that a variable, external illuminance is not a
statistically significant factor in predicting lighting use patterns.
Table 5 shows time that has been passed until the lighting was
first on since the occupant’s first arrival at an investigated office in
the morning. There was a close link between the start of daily occupancy
and switching-on of lighting. First light switch-on events in
the investigated offices occurred within 11 min after the start of
daily occupancy. For example, in the office A, occupants turned on
the lighting 3 min (SD = 1 min) after they first opened a door in the
morning. One-paired T tests indicate that the differences between
the start of the daily occupancy and the first light switch-on event
in the investigated offices were not meaningful.