The generally prescribed procedure for choosing a decision strategy from a
decision tree employs a backward induction analysis that entails 3 fundamen
tal consistency principles: dynamic, consequential, and strategic. The firs
t requires the decision maker to follow through on plans to the end, the se
cond requires the decision maker to focus solely on future events and final
consequences given the current state of events, and the third is the conju
nction of the first 2. Five experiments were reported to test these princip
les using different subject populations, different procedures for estimatin
g consistency, and different factors for manipulating the attractiveness of
the gamble at the final stage of the tree. The main findings were that str
ategic and dynamic consistency principles were violated at rates that excee
ded choice inconsistency.