A Note on the Existence of Starting
Point Bias in Iterative
Bidding Games
Karl C. Samples
This note further illuminates the strength and direction of starting point bias in iterative
bidding procedures. Conflicting recent findings concerning the starting point bias phenomenon
are first briefly overviewed. The hypothesis that starting values influence valuations obtained
in iterative bidding games is then tested in an experimental setting using widely disparate
starting values ranging from $1 to $8,000. Statistically significant differences in mean final
bidding outcomes were consistently detected in games using different starting values. This
evidence, combined with test results reported elsewhere, strongly suggests that starting point
value selection can have subtle but significant effects on observed final bids.
The purpose of this note is to further
illuminate the strength and direction of
starting point bias in iterative bidding
procedures. To date, most experiments
concerning starting point bias in iterative
bidding games have focused on compar-
ing responses using relatively similar start-
ing values ($1 versus $10, $25 versus $125,
or $1 versus $50). This study has pursued
the matter one step further by using
widely disparate starting values ($1 versus
$800, and $1 versus $8,000). This ap-
proach allows biases to be detected which
may be otherwise overwhelmed by other
influential survey design factors such as
choice of payment vehicle or sample se-
lection. Statistically significant differences
in mean final bidding outcomes were de-
Karl C. Samples is Assistant Professor of Agricultural
and Resource Economics at the University of Ha-
waii-Manoa.
This work was funded by the National Marine Fish-
eries Service under NOAA contract 82-ABC-00251,
“Experimental Valuation of Recreational Fishing in
Hawaii.” The helpful comments of John Dixon,
Maynard Hufschmidt, Sam Pooley, Robert Rowe,
Chennat Gopalakrishnan, and three anonymous re-
viewers are gratefully acknowledged. Hawaii Insti-
tute of Tropical Agriculture and Human Resources
Journal Series No. 2883.
Western Journal of Agricultural Economics, 10(1): 32-40
© 1985 by the Western Agricultural Economics Association
tected in games using different starting
values. This evidence, combined with test
results already reported by other authors,
strongly suggests that values selected to
initiate bidding games can have an appre-
ciable impact on observed final bids.
Research on iterative bidding games has
examined various sources and direction of
biases associated with the method. Partic-
ular attention has been devoted to mea-
suring whether inherent structural char-
acteristics of bidding procedures, including
selection of starting bid values, signifi-
cantly influence bidding outcomes. Rowe
et al. hypothesized that selection of start-
ing value may affect iterative bidding
game results through a combination of two
effects: (1) if the starting value is far away
from the true bid value, the iterative bid-
ding process may fatigue or exasperate the
respondent, thereby providing incentive
to terminate the game before the true fi-
nal bid amount is reached, and (2) the
starting value conveys information to re-
spondents about expected or reasonable
bids and thereby influences the final bid
outcome.1 These two sources of bids can
1 Brookshire et al. (1974) were perhaps the first to
explicitly recognize the starting point bias phenom-