








Determination of changes in factors associated with the surve 1970 
to 1971 
Annual changes in the mean survey date, temperature at start of 
survey, and percent of route listening stations with high disturbance 
are presented for the first time in this report (table 2). These data 
have been collected from comparable routes run in both 1970 and 1971 
which were accepted for analysis of numbers of doves heard. The anal- 
ysis of the factors associated with the survey employs the same pro- 
cedure described for determining annual population changes; all data 
were weighted by the land area value of each physiographic region 
within States and management units. Differences in route totals are 
attributed to the deletion of Automatic Zero Routes and routes con- 
taining missing or unacceptable data for one or more factors. 
Computer analysis of dove call-count survey data 
A significant modification in the computer analysis of the call- 
count survey data has been accomplished through the programming efforts 
of Mr. Frederick R. Fiehrer of the Migratory Bird Populations Station. 
For the first time, nearly all data analyses were performed by computer 
and printed in final tabular form as presented herein. This program 
provides properly weighted averages and linear regression data by State, 
Management unit, and physiographic region. It yields the mean differ- 
ence, the standard error of the mean difference, and level of signifi- 
eance of the change. 
otatistical evaluation of data 
The Call-Count Survey was designed to detect major year-to-year 
changes in the breeding population index within each management unit 
(Foote, 1959). Analysis of the 1970 and 1971 data revealed that 
observed differences of about 8, 9, and 13 percent between these years 
within the Eastern, Central, and Western Management Units, respectively, 
would be statistically significant at the 95-percent level. Although 
the survey was not designed to detect a change in the BDI between 
years within States or physiographic regions, data from these areas 
were also subjected to statistical analysis. 
Long-term BDI's, adjusted to a BY for all physiographic regions, 
States, and management units, were examined to determine whether sig- 
nificant trends were present. Data from each source were analyzed 
using a linear regression model. 

