The Southwestern Historical Quarterly, Volume 52, July 1948 - April, 1949 Page: 10

o10 Southwestern Historical Quarterly
making more than one move. Otherwise their effect is too small to
impair seriously the correctness of the information about the
sources and channels of family migration.
Because it assigns a date to each arrival, the child-ladder method
promises to measure rates of migration. The present investigation
began with the hope that a mere compilation of the annual totals
of ascertained arrivals of families would turn out to be a true
index to the relative volume of migration from year to year.
Accumulation of data eventually proved, as careful reasoning
might have foreseen, that the gross annual totals overstate heavily
the volume of recent migration and understate the volume of
earlier migration. The explanation of the distortion is that every
passing year weakens the chances of detecting a family migration
from the ages and birthplaces of children. For migrations on the
eve of the census, enumeration itself proves the arrival of the
family in Texas, and the census date can take the place of the
birthdate of a Texas child. Hence one out-of-state child, aged five
years or less, will prove a migration.7 But a migration in the
fourth year before the census, or earlier, requires for the proof
two children, one born outside and one inside Texas. Because two-
child detection is more exacting than one-child detection, there
is between the third and the fourth years a heavy drop in the
proportion of ascertained arrivals to all arrivals. Back of the
fourth year, the proportion declines slowly but steadily. The
earlier the migration to be detected, the older the two children
must be; and the older the children, the harder they are to find,
death having thinned the stand from infancy onward. The com-
parative under-detection of the earlier migrations and over-detec-
tion of recent migrations require corrections of the gross annual
totals. The requirement, entailing labor and opening another
door to error, is unwelcome. Fortunately, appropriate correc-
tions can be estimated with fair certainty and precision.8
7An out-of-state child aged above five years is also evidence of a migration. But
if the date of migration is to be inferred with acceptable accuracy, then the
number of years intervening between proof of residence in the place of removal
and proof of residence in Texas must be limited. The limit observed in this
study is five years.
sA description of the procedure is appended to this study. The corrected figures
are called adjusted totals, the original uncorrected figures, gross totals. In calculat-
ing sources of migration, gross and adjusted figures give almost identical results.