The problem of finding the least favorable configuration for selecting the best of k populations i.e. the one with the largest location parameter by use of six different two-stage selection procedures is considered. Each of the six procedures consists of a subset selection (screening) rule at the first stage followed by another rule based on (the first stage and) additional samples from the selected populations to decide finally which of the selected populations is the best. In the indifference-zone approach it is (or was) conjectured that the least favorable parameter configuration is of the slippage type. It is shown that this conjecture is true for four of these procedures. For a fifth procedure it is proved that at least a certain lower bound of the probability of a correct selection has this property which is analogous to the result of Tamhane and Bechhofer (1979) concerning the sixth procedure.