Factors Affecting Response Rates 
Hunting Success 
Response rates were calculated for various hunter-success 
groups (Table 6), and proved to be related to the size of the kill 
previously reported on the Mail Questionnaire Survey. The higher 
response rates of the more successful groups of hunters probably 
reflected more successful hunting and perhaps also more interest 
in coopérating among hunters in the more successful groups. There 
is also a positive correlation between the wings received per 
respondent and the reported kill. If it is assumed that there is 
a relationship between kill and number of wings sent in it might 
be inferred that there is a relationship between the number of 
birds the hunter actually kills and the number he reports on the 
Mail Questionnaire Survey. Response rates of the lowest success 
group may have been slightly reduced by the system of envelope 
allocation (see page 12). 
The Follow-up Letter 
A letter reminding hunters to keep sending wings was sent at 
or near the middle of the season to half of the hunters of the 
States of Michigan, Illinois, and Arkansas. Hunters having odd 
identification numbers were sent follow-up letters and those 
having even numbers eerved as a controle Hunters' numbers were 
originally assigned by grouping all hunters by post offices of 
Stamp purchase, placing those post offices in alphabetical order, 
and then numbering the hunters consecutively. 
In all three States, response rates of hunters who received 
follow-up letters were slightly higher than those of hunters who 
did not receive the letters (Table 7). Also, after the time when 
letters were received, those hunters that received letters returned 
relatively more wings than did those who did not receive letters. 
It is apparent that in this study the follow-up letter increased 
late-season responses. It is possible, however, that if there 
were different patterns of duck availability in different years, 
follow-up letters would have different effects and the comparability 
of the data would be reduced. For example, if there were differences 
in the time -of the peaks of migration, follow-up letters might arrive 
in an area one year during a period when large numbers of ducks were 
present and another year arrive when very few were present. Response 
rates could be greatly altered by such a circumstance. If no follow- 
up letter were used it seems likely that results would be more consistent 
from year to yeare We believe follow-up letters hould not be used 
because of the possibility of complicating rather than simplifying the 
interpretation of response rates. 
-9- 
