It is with great reluctance that I serve B. Wilbrink (NRC-Handelsblad of 27 September 1974, page CS3) with a reply, because it is almost always a pointless undertaking to get into a discussion with a scatterbrain. The importance of the topic does not permit me to remain silent about it.
According to Wilbrink's definition, an eighteen year old is deemed either "fit" or "unfit" according to whether he, being admitted to higher education, finishes his studies with or without a good result [pass]. Under the broad assumption of a maximum allowed time to complete a course of study, the determination of "fitness" is indeed a finite experiment that on average would last about six years. The quality of "fitness" is thereby defined only for a 24 year old.
Mr Wilbrink however suggests that this binary quality is also defined for the 18 year old. If of a homogeneous group of 18 year olds, one out of 5 should later prove to be unfit and 4 to be fit, he (afterwards!) says one is unfit and 4 are fit, but he presents the selection problem as a problem of separating the one 18 year old from the other four. If however the group of 18 year olds would present itself as homogeneous - for the sake of the argument, it could be five identical boys of that age, which need not exclude that one out of five would fail in his studies - we can only come to the conclusion that for an 18 year old out of such a group, the chances of successful completion can be estimated at 0.8 (this under a continuity assumption based on earlier experiences with similar groups).
To make the most of limited teaching capacity, there is only one way open to us, which is admission not by lottery, but based on a comparative exam where student groups are admitted in order of decreasing chance of success. Whether the cutoff is higher than 0.5 is, contradictory to the suggestion of Mr Wilbrink, completely irrelevant.
As a reader of the NRC-Handelsblad, I hope in future to be spared semi-scientific, warped reasoning such as Mr Wilbrink's.
Prof. Dr Edsger W. Dijkstra