21 Pages


In Arisz and Kanis (1999) a statistical approach is proposed as a possible means to monitor the number of participants in a user trial concurrently. In this approach, the total number of distinctive phenomena of interest found after an unlimited number ( ‘ oo’ ) of participants, , is estimated on the basis of the findings after n participants. This insight can be used to assess the expected payoff (new information) of the extra effort (time, money) to involve more participants. In the present paper, the underlying mechanism of the statistical approach is discussed, especially as to the emergence of biased estimates. In addition, application of the approach to empirical data sheds light on the tenability of the rule of thumb that 4-5 participants would be enough to gather 80% of F„ . In the rest of this paper, the term usability problems is used as a short-hand indication of all kinds of possible phenomena of interest.