Upgrade your account to a Community Supporter account and remove most of the site ads.

Reply to thread

Message: <blockquote data-quote="Ovinomancer" data-source="post: 6843827" data-attributes="member: 16814">While I don't care much about the topic, you can argue however you want on what IQ is what INT, mapping the statistical distribution such as dice throws, with defined possible states, to an ordinal distribution is invalid.  SD doesn't have much meaning in an ordinal data set, and is only useful to compare values within the set.  You can't do statistics on ordinal data structures that has validity outside the data (ie, it's only partially useful to look at relationships within the data set).This is because ordinal systems are just ranked data -- this is higher than that, and that over there is higher that those other two, etc.  A Likert scale is an example of ordinal data.  Since there's no real definition of the difference between values other than the ranking, and the actual difference can change between different values (or, for opinion polls on Likert scales, between poll takers), statistical comparison of ordinal data to other statistical models is inappropriate.So, you are correct that IQ is scaled based on population and likelihood, but that doesn't mean it's actually distributed that way.  The values are only rankings and have no value in and of themselves.  They only refer to what is better that which within the data sat. The difference between points can change depending on where you are in the data (ie, the difference between a 100 and a 105 may be very small, but the difference between a 140 and a 145 may be larger or smaller, but it's probably different from the 100/105).  This means you can't take that arbitrary distribution, even though it uses terms like SD and means, to map to a valid statistical distribution like 3d6.  They can't possible map accurately.However, if you want to do that, and it provides use to you for you game, be happy and continue!</blockquote>

Verification