Factors affecting subjective estimates of magnitude: when is 9 > 221?

112 undergraduates separately judged the size of the numbers 9 then 221 or 221 then 9 on a 10-point scale or on a continuous line that were both anchored only at the extremes by "very, very small" and "very, very large." Replicating the 1999 Birnbaum study, 9 was given a higher rating than 221 when the numbers were judged first by different people on the numerical scale. However, 9 was given a similar or smaller rating than 221 in the other conditions. The results are discussed in terms of context effects.