posed workers and of the control subjects on the Farnsworth
panel D-15 test and on the Lanthony desaturated panel D-15
test. We calculated the differences of the CCIs before and after
cleaning and the control situation, respectively. The medians
of the differences were not statistically different (
2
0.08
and
2
0.02, respectively, Wilcoxon–Mann–Whitney test, onesided
tested). The results of the panel tests were confirmed by
those of the SPP2 test. No subject made any real error. The
median difference of the number of the possible blue–yellow
errors after and before cleaning and the work of the controls,
respectively, was 0.
Chronic exposure.
All eight control subjects performed
well on the Farnsworth Panel D-15 test, but the color confusion
indices of three printshop workers were slightly elevated
(Table 3). Regarding the Lanthony desaturated panel D-15
test, which is more difficult, the performance of the exposed
workers was worse (medians 1.18 and 1.05, respectively,
p
5
0.06, Wilcoxon–Mann–Whitney test, one-sided tested), but
the difference did not yield statistical significance. The individual
results are given in Table 3. The errors made by the exposed
workers were of blue–yellow type. All subjects recognized
all figures on the SPP2 test correctly. The number of the
possible blue–yellow errors was not elevated in the exposed
group, when compared with the control group (medians 0.25
and 0.5, respectively).