
and the average seasonal perform- 
ances a hunter in terms of days 
hunted and waterfowl bagged. 
2. Measure the relative changes in 
these estimates from year toyear. 
3. Assess the effects of changes in 
season length and size of daily bag 
limit on total bag and on hunter per- 
formance, 
The 1962-63 hunting season marks the 11th 
consecutive year of the survey since its in- 
ception in 1952. 
Since there is no complete listing of water- 
fowl hunters to facilitate the survey, by 
necessity, the survey utilizes as its sampling 
universe those post offices throughout the 
Nation selling Migratory Bird Hunting Stamps 
(hereafter ‘“‘Duck Stamps’’). Eachyear more 
than 2,000 post offices are designated to co- 
operate in the survey as ‘‘sample outlets.’’ 
These outlets have been selected randomly, 
within States, from two strata of post offices: 
main offices having outlying branches and 
stations under their jurisdiction (usually in 
larger town and cities), and offices without 
branches or stations (usually in smaller towns 
and rural localities). A further refinement 
in stratification—that of subdividing each 
State into several geographic ‘‘zones,’’ while 
retaining the post office class strata within 
each zone—was devised and made operative 
beginning with the selection of the 1963-64 
sample outlets. The attempt in zoning was to 
devise geographic areas wherein the average 
seasonal duck bag ahunter among post offices 
of stamp purchase, was expected to be the 
most similar, thus increasing survey preci- 
sion. Zoning also insures a more even dis- 
tribution of sample outlets throughout a 
State. 
Names and addresses of hunters are ob- 
tained by means of business-reply ‘‘contact 
card’’ distributed to all persons buying 
Duck Stamps at sample outlets. The card 
requests the individual’s name and address, 
the number of stamps he purchases, the 
reason for his purchase, and the number of 
persons in his household under the stamp 
requirement age of 16 years who might hunt 
waterfowl during the season, 
All contact-card respondents who purchase 
stamps for the purpose of hunting are mailed 
a hunter questionnaire at the close of the sea- 
son, The 1962-63 questionnaire asked each 
hunter the total number of days he hunted 
waterfowl, his total bags of ducks, geese, and 
coots, and the number of each he knocked 
down but failed to retrieve. These data, in 
combination with total reported sale of Duck 
Stamps by State, have been used toderive the 
various survey estimates. 
All estimates are subject to several 
sources of error (as is true for most type 
surveys}. In addition to chance error due to 
random sampling variation, the estimates 
may be affected by misreporting (respondents 
tend, for example, to exaggerate their bag}. 
The estimates are further subject to non- 
response bias in that hunters who fail to 
respond may have differed from respondents 
in their hunting performance. 
A further potential source of error, that of 
faulty reports of Duck Stamp sales, this year 
developed into a serious problem. Since itis 
necessary to present kill estimates no later 
than early July to be available for regulations 
meetings, the reported sales for the first 
three-quarters of the fiscal year (July 1 to 
March 31) must be used in deriving total kill 
and activity estimates instead of the full 
year’s sales (the sales for the last quarter 
of the fiscal year are not available until 
mid-August). Prior to 1959, the three- 
quarter year reports were generally identical 
to those for the full year; since that time, 
however, a small but increasing percentage 
of sales have been reported during the fourth 
fiscal quarter (April 1 to June 30). Whether 
such reported sales had actually occurred 
during the fourth fiscal quarter was difficult 
to establish: purchases by conservationists 
and philatelists were considered a possible 
explanation, This year, however, it became 
apparent beyond reasonable doubt because of 
the magnitude of the fourth-quarter sales 
reports of many State, that the three-quarter 
year sales reports were incomplete. (In 
certain States, as high as 25 percent of total 
sales were erroneously reported to have 
occurred during the fourth fiscal quarter, and 
certain post offices furnished no three- 
quarter year data whatsoever.) Unfortun- 
ately, such errors can not be detected until 
after regulations meetings. Measures to 
correct this reporting lag are now 
being negotiated with the Post Office 
Department. 
Because of the incompleteness of the three- 
quarter fiscal year sales reports, it was 
necessary to recompute the entire 1962 
analysis. The revised estimates, based on 
the full-year’s stamp sales, are presented in 
this report. 
The 1961 estimates presented here for 
comparison with the revised 1962 estimates 
have also been adjusted so as to be based on 
