the aggregate, the same distribution as 
those used in the comparison (fig. 1). 
The difference in minutes between the 
reported and observed times of kill of the 
889 birds is summarized in table D-2. 
Nearly one-half of the reports (49%) dif¬ 
fered from the observations by less than 
15 minutes, and only 18 percent by more 
than 1 hour. To check whether the omit¬ 
ting of observations had altered the com¬ 
parison, the 1963 comparison was reex¬ 
amined using only those forms for which 
hunters had reported the exact number of 
kills that were observed in the field. Here 
again, the results were similar to the 
comparison that had been made. The data 
indicate that although hunters may err in 
reporting the time of day at which they 
kill a bird, most errors are small and 
tend to be compensating when viewed 
in the aggregate. The amount of agree¬ 
ment between reported and observed 
hour of kill is all the more noteworthy 
because 95 percent of the envelopes 
were distributed to the hunters after 
the hunt; thus the hunter had no prior 
notification to be aware of the time 
kills were made. The accuracy with 
which hunters report time of kill could 
be greater in the annual duck wing 
collection survey because all the 
hunters sampled receive envelopes with 
instructions before their first day’s 
hunt. 
Figure 1. —Percentage comparison of 889 ducks killed by hours of 
day between hunter reports and spy-blind observations, 
1961-62, 1962-63, and 1963-64 data combined. 
27 
