write:A study of intuitive numerical estimation illustratesthis effect. Two groups of high school student [choult os estimated within 5seconds a numerical expression that was written o

as starting points. Payoffs foraccuracy did not reduce the anchoring effect.Anchoring occurs not only when the starting point is given to the subject
but also when the subject bases his estimate on the result of someincomplete computation. A study of intuitive numerical estimation illustratesthis effect. Two groups of high school student [choult os estimated
within 5seconds
a numerical expression that was written on the blackboard. Onegroup estimated the product8 ×7 ×6 ×5 ×4 ×3 ×2 ×1while another group estimated the product1 ×2 ×3 ×4 ×5 ×6 ×7 ×8To rapidly answer such questions
people may perform a few steps ofcomputation and estimate the product by extrapolation or adjustment.Because adjustments are typically insufficient
this procedure should leadto underestimation. Furthermore
because the result of the first few steps ofmultiplication (performed from left to right) is higher in the descendingsequence than in the ascending sequence
the former expression shouldbe judged larger than the latter. Both predictions were confirmed. Themedian estimate for the ascending sequence was 512
while the medianestimate for the descending sequence was 2
250. The correct answer is40
320.Biases in the evaluation of conjunctive and disjunctive events. In arecent study by Bar-Hillel19 subjects were given the opportunity to bet onone of two events. Three types of events were used: (i) simple events
suchas drawing a red marble from a bag containing 50% red marbles and 50%white marbles; (ii) conjunctive events
such as drawing a red marble seventimes in succession
with replacement
from a bag containing 90% redmarbles and 10% white marbles; and (iii) disjunctive events
such asdrawing a red marble at least once in seven successive tries
withreplacement
from a bag containing 10% red marbles and 9% whitemarbles. In this problem
a significant majority of subjects preferred to beton the conjunctive event (the probability of which is .48) rather than on thesimple event (the probability of which is .50). Subjects also preferred to beton the simple event rather than on the disjunctive event
which has aprobability of .52. Thus
most subjects bet on the less likely event in bothcomparisons. This pattern of choices illustrates a general finding. Studiesof choice among gambles and of judgments of probability indicate that