The purpose of this paper is to test a procedure that would be useful for a certain, more final purpose.
Assume that you want to know if the differences in the result of a certain effort in health care have any connection with differences between the life situations of the patients. This would be easier to investigate if the life situations of the patients had connection to the easily available variable Health Care Cost (in Swedish VK). In that case the searching for connections between VK of patient groups and the result of health care efforts would be sufficient to gain knowledge of possible connection between health care efforts and life situations. Thus that is the final purpose.
The purpose of this paper is to find a connection between combinations of variable values (in Swedish VVK) of life situations and mean and variance of the variable VK.
Mean and variance for VK is let become the variables of a two-dimensional space. In such a space every observation contains data for a group of patients. If the observations for patient groups with similar life situation are more often in the same areas than they are hazardously spread out over the space, there probably is a connection between life situation and mean and variance of VK.
The concrete purpose of this paper is to try to find groups of similar life situations, groups represented by dots mostly found within certain areas (of the space mentioned).
The data material for the paper has a somewhat odd nature that makes it a bit awkward to work with. On the other hand data are easily achievable, updated every year and chosen all from the beginning for connecting differences in VK as statistically well as possible to life situation variables.
The oddness of the data makes the understanding of the methods require a lot of text. This is why the methods are excluded from this summary.
The results seems to mean that the groups of similar life situations are hazardously spread out over the space. On the good side this seems to be able to blame on the part of the method that was not thought through thoroughly, i.e. it is not proved that the method could not work. On the other hand the reliability of the results seems to fade away...
