Ernie and others
I often seem to react when all others have finished the topic. It feels
nice having a strong impulse to make clear things who are in a dark corner.
To do research using a survey or a questionaire is not at all easy. However
that does not mean that nothing can be done.
First I want to show some of the difficulties one runs into. In creating a
questionaire there are two possibilities.
1. Closed questions can be used. The subjects can answer yes or no or they
can be given some choice on a 5 points-scale.
2. The other possibility is to use open questions. The subject tells his
own story and can use his own words.
The first method seems the easier one. It results in a matrix with numbers
and the mathematician feels happy and starts computing.
Most people in trying to complete a questionair have felt irritated. The
questions seemed to be wrong: the real answer often seemed to be: that
depends....To prevent this, knowledge about the topic that is investigated
is needed. Besides, it is amazing difficult to formulate questions meaning
the same for all people.
So possessing this matrix of numbers, the researcher can't know what they
really and exactly mean. Much pilot study has to be done to overcome this
problem.
However, having done so, the troubles don't stop. The next step is to
compute which relations exist between the variables. One can expect only
weak relations. When very strong ones exist they need no search; they can
be seen at once. But even if the numbers in your matrix are actually 100%
random you may find some weak relations by chance . So the margin to find
anything that is something is very small.
It may be clear by now, that this kind of investigation is hard to execute
and that the road to clear conclusions is full of pittfalls.
The second form of investigation, the open question one does not give all
those troubles. It is a more modest approach, more as a journalist.
It does not result in clear cut scientific statements but some interesting
material can be expected. In the former method interesting things easily
disappear.
Ernie or others who are interested think it over. When you want to make an
inventarisation of some sort 10% of the listembers wanting to cooperate is
sufficient. Of course it is a biased sample but all the listmembers
together are so too.
Plans like this need some time to brood over.
Ida Kamphuis
Holland
