Friday, November 04, 2005
www.moral-politics.com test results
moral politics test: it's correct as far as it goes. but it's important to note the specific definition of terms used on the site as well, as they probably don't correspond well to popular notions. plus, of course, there are a lot of specifics that don't match. that's what you get when you design a test to look for only a single result. perhaps using a agree-disagree scale for each proposed answer (resulting in 64 instead of 16 questions) would lead to more acccuracy, plus also relay distribution information as well - as in, don't plot just the single balance point, plot a shaded distribution corresponding to confidence interval or standard deviation or something, and also plot the individual answer points themselves. i know that for me i have a couple of interesting outliers in each of the other quadrants.
take the test yourself, post the two axis scores in a comment.
the other thing that seems evident with some thought is that using a square grid misrepresents the realistic possibilities somewhat. for example, the upper left and lower right corners are unusual combinations of non-conformity / socialism in the first case, and conformity / individualism in the latter. conversely, political expressions along the other diagonal (lower left to upper right) could be considered more “natural”. so what we have is more of a leaning oval of possibilities. this also points out that systems along the edges of the oval, instead of the square, are the extremes, and this is where we find the form of conservatism common in the u.s.. in this analysis, u.s. democrats actually have a more natural mainstream political expression.
interesting observations may be made about other off- and on- axis situations as well. for example, it seems that the social democracy common in the world is not exactly “natural”, but the novel combination of traits seems valuable.
but - when you look at the website's maps of the distribution of test results, you find most people aligned along the “unnatural” diagonal. however, i think this is actually an artifact of the mis-shaped grid in the first place: if you compress everything into the oblong, it becomes more clear that the variation is fairly evenly spread along the “unnatural”, while opinions are generally centrist in terms of group/conform vs. individual/non-conformity - a seemingly healthy result.
i guess it's to be expected that the two variables used in this test have a high degree of interdependence. there should be a statiscal method of determining the degree of this correlation, and then perhaps normalizing for it. and it occurs to me that the scale used along each axis, and away from the axes as well, will vary in a non-uniform matter.
yeah, ok, overthinking things yet again ;-)