Friday, February 09, 2007


This week about 170.000 children in the 8th grade of primary school took the so-called CITO test. It's a 3-day nation-wide assessment that is used to determine what level of secondary education the child can attend after the summer vacation. Every year there is a lot of discussion on the predictive value of the CITO test. That's probably why additional tests and the advice of the primary school is also taken into account when choosing a school for secondary education. About 20% of the primary schools already have chosen to do another type of final assessment.
One of the major criticisms is that everything seems to depend on these 3 days. If a child feels sick or suffers from stress (it can be a stressful experience) his score may be significantly lower than in a normal non-stressful situation. The CITO test is just a snapshot within a period of several years. An alternative test (the NIVO test) claims to take account of this problem by taking smaller tests during the whole year. This way the scores are assumed to be much more balanced, thus being a better predictor of a child's ability to perform. Another criticism that can be identified is that a relatively low CITO-score can be compensated by motivation, but also the other way around: a high score is no guarantee for success.
Personally, I have three experiences why I doubt the predictive value of the CITO test. The first experience is that I went to secondary school (pre-academic level; VWO in Dutch) having the lowest CITO score of a group of 5 peers who went to the same school. My 4 peers all dropped out, changed to lower levels of secondary education, the first one already after 6 months! I was the only one who finished the VWO level and achieved my degree. My second experience is at the University of Nijmegen. During my 1st bachelor year in psychology one of the instructors was interested to know how many of my cohort had achieved a high CITO score. The result was rather dramatic: according to the CITO test about 50% was tested for a lower level. The third experience is with my own son, who took the test 2 years ago. He achieved a score of 541 which is 3 points below the entrance level at some secondary schools at pre-academic level. Taking a closer look at his score profile, I noticed that he had a relatively low score on math, which considerably influenced his overall score. Strange, because in the past three years he used to score straight A's on math. Why not on this one CITO test day, I wondered? An additional talk between his teacher and the secondary school (a Gymnasium) he wanted to attend was needed, because normally they use the strict policy that an entrance score of 544 is required. Well, he was accepted under the condition that he could not fail his first year, otherwise he would have to move to another school. My son is halfway 2nd grade now and he is doing fine. His test scores are quite well and he is 'on schedule' to go to the 3rd grade after this summer. He's doing as least as well as some of his classmates who achieved the maximum 550 for their CITO test.
To conclude this post, the CITO score is just one predictor in a series of many other ones. The relative weight and stress that is put on these three days in February is rather silly. Primary schools are keeping track records for years, which seems much more reliable to me than the rather coincidental score on three days in a whole year. Why not spread out the CITO test over the 8th grade (e.g., one day in october, december and february)? A completely other consquence is that the 8th grade children are virtually free, starting today until the first week of september when they start their career in secondary school. Of course, they have to attend their schools the next couple of months until the end of June, but everything is already set. In the first week of March they receive their score, one or two weeks later they subscribe to the school they will attend after the summer. For some children it will be tough to start studying again after a 7-month period of freewheeling.

1 comment:

Henk Roossink said...

CITO, Eindtoets primary school

Yes, there is a lot of discussion abouth the quality of the CITO-test, but very often not a discussion bases on logical arguments (except Stanley’s comments of course). I would like to react on some points of criticisme.
1 The predictive validity of the test.
The CITO-test is measuring subjectmatter-knowledge as language, math, and studyskills. That’s a lot, but that’s all.
Cito has done reasearch on the test, relating testscores to the school career of the pupils. With a success-score of 75 % as they say ‘De Eindtoets’ (since 1970) is realy a good test, according to the positive jugdments of COTAN (COmmittee for Testing in the Netherlands, a division of NIP, the Dutch Association of Psychologists).
If we want to reach a better predictive validity, than we should not confine ourselves to the small content of the Cito-test, but take into account other aspects (determinants) of studysuccess. Trying to improve the psychometrics of the test would not be succesfull. Other aspects are for example motivation, health, perseverance, intelligence, social and economical background and so on. There are other tests (accepted by COTAN) that can be used. For example Drempelvrees (knowledge and motivation). We could discuss how many instruments are usefull and feasable for primary school. Can we do more then the three days testing now?

2 Advice of the primary school is taken into account for the allocation of primary school pupils to the secundary schooltypes.
That’s fine. The CITO-test is besides the judgment of the school an independent measurement of the knowledge of the pupils. And the school that is keeping track records for years should use this information for their advice; many schools are doing so. CITO wants the pupil-allocation to secundary school being based on the school advice as well as the cito testscore.

3 A three days test, stressful, why not spread out the test?
Testing must be valid, but feasible too. The Eindtoets can be taken pretty easily and teachers regard that as an advantage. Stress and a three day’s testing period have never been a point of discussion after research or evaluation. Perhaps it is a good idea to spread out the test.

4 And can we do allocation or selection without mistakes?
No, of course not and that means that we can never stand surety for an individual that a decision is made right. An applicant for a job or a pupil entering secondary education can have bad luck. The pupil will in one of the following years need to change to another level of education, but that takes time. And a prediction on the long run, for example at primary school the prediction if someone can go to HBO or University is realy impossible.

5 Conclusion. I think the best solution is one in which the Eindtoets is part of it and hopefully the parents are playing part of the the game as well as the pupil itself and the primary school teacher. In 14 % of the cases there is a difference between the primary school advice and the Eindtoets. 90 % of the secondary schools are using the Eindtoets-scores, 7 % not. They (74 % of the schools) are using rules for admission to VWO, HAVO, and so on (Eindtoetsscore), but they talk with the primary school teacher , the pupil and the parents for flexible admission (94 % of these schools).