Fair comparison

from Wikipedia, the free encyclopedia

The fair comparison is a statistical method to be adapted to the data to be compared to each other. The method of fair comparison can usually be assigned to multivariate statistics. In a fair comparison, it is assumed that the data to be compared cannot be directly compared due to different influencing factors.

example

Assuming a Bundesliga team in men's football played against a children's team and the score after 2 halves was 38: 1, it is hardly possible to make a statement about how good the children's performance was, since the comparison with the adult opponents due to the superior physical conditions would be of no value. If there were an algorithm that could compensate for the goal score of 38: 1 using certain weighting factors depending on variable differences in the two teams, a fair comparison with good informative value would be possible. It would be conceivable to add up the lung volume of the adult team and the children's team and to form a ratio from this, which analogously influences the goal score. This procedure would even enable the children's team to emerge as the winner in a fair comparison.

For another, but real, example for a fair comparison, see also the glider index list .

The fair comparison in the Austrian education system

Under the direction of the BMUKK in Vienna, the Austrian Federal Government is concerned with implementing so-called educational standards with competence models in the Austrian school system. This implementation corresponds to a changed understanding of school, away from the curricula (input orientation) towards output orientation, according to which school should help the pupils to acquire certain competencies.

These competencies are operationalized in the framework of the educational standards review (BIST) at annual intervals in grades 4 and 8 in many school types in Austria through tests. The participating schools can z. B.

  • be an urban secondary school with many migrants and a high proportion of girls
  • rural secondary school with few people with a migration background and many boys .

It should be obvious that the geographical location, foreigner quota and gender distribution - at least in some disciplines - must have a significant effect on the expectations one has of a school class with regard to its tested level of competence.

As part of the annually recurring standard checks, these variables are collected and statistically evaluated in parallel with the actual test execution. The result is e.g. B. in the context of the so-called teacher feedback, which is part of the annual test evaluation, an expectation range (expressed in the closed interval from 200 to 800) in which the average performance values ​​of an examined class group should be. So it may be possible that this fair comparison procedure must lead to the realization that a school with many obstructive factors (e.g. high migrant quota, low socio-economic status of the parents) and a characteristic value of 440 points in a certain competence area ( the expected range was, for example, 400-420 points) apparently pedagogically more solidly worked than another school whose school class achieved 550 points in the same test section, but where the expected range was estimated at 580-604 points.

Web links

  • [1] , institution for the implementation of the east. Educational standards studies and their evaluation including a fair comparison