Re: False quantification to the nth degree

Christopher J. Morgan (Christopher.Morgan@JCU.EDU.AU)
Wed, 15 Jun 1994 23:37:39 +1000

>i forgot to add, that as every mathematician and scientist and i would
>hope every social scientist should know, if your input has no significant
>digits, then your output cannot have any either. therefore if student
>scores are reported as 1,2,3,etc., then their grade averages must be
>reported as 1,2,3,etc. by that system, 1.999 rounds up to 2 as does

Tru Ya, Bonnie - I was just waiting for someone to make that obvious point.
However, one arbitrary aspect of mathematical custom further obfuscates the
inherent bullshit in such reasoning - 2.500000000 rounds to 2 also. At this
institution we're currently undergoing all sorts of angst re "quality
assurance" in our teaching - which I suspect (how goddamn negative I am at
this hour of the antipodean night) will result in similar reductionism.
Fortunately, this particular debate has been fought & won in (shudder...)
statistical circles some years ago, under the aegis of the perils of
converting ordinal to interval scales. Mind you, this has not deterred
colleagues from 'cognate' disciplines (read psychology) from routinely
performing e.g. ANOVA on Likert scales. Shit happens... & happens....

Arf Arf

* | *
* Christopher J. Morgan | Tel: 0561 77 81 4304 *
* Dept of Anthropology & Archaeology | Fax: 0561 77 79 5435 *
* James Cook University Q 4811 | *
* Australia | *
* _____________________________________|______________________ *
* "In the government of primitive is recognised *
* nowadays that the old methods of slaughter and slavery were *
* not only inhuman but also unscientific...." *
* - Sir Hubert Murray, 1933 *
* *