Stats Thoughts

Measuring the unmeasured/ measurable?

Much what we measure in surveys and studies stay limited to expressed behavior and established attitudes. Among our prospects, consumers , associatives and the lay public.But admittedly what we are often trying to get hang on is the deep perceptions, interceptions, motivations propensities and temperaments.
The question here that bears salience thus is — Are we able to do so with due accuracy? And the answer here is often waveringly ‘yes’ mostly at best. Some may bravely argue about this given the natural limits of qualitative techniques …. but i would hesitate to nod along being a qualitative methods expert myself.
Qualitative methods work wonderfully in exploratory mode. Lacking however the numbers and decimals are subject to some reworking that should make them readily usable in complete form.
This is necessarily and unexpectedly in direct negation to what is well known to us in cases such as ‘IQ’ test or say ‘Personality Disorder’ test.
This is explained when we become aware of the  standardized tools batteries & inventories that these ‘IQ’ and ‘Personality Tests’  as MMPI have undergone and imbibed over the last three generations. For these tests the  questionnaires have undergone the rigor of validation and reliability tests over large sample size over repeated usage. However when  a new casual and ad hock attempts in hurriedly done surveys/studies the dilemma of the accuracy of decimals & percentages or their unavailability continues to persist. A valid caveat for both researchers & clients.

On the other hand long and protracted rigor require time which we may be in short given the need to “know now!” needs of marketers. This is partly answered for via use of ‘Factor Analysis’. Factor analysis here certainly gets us out of this jam. It is standardized , casually applicable and somewhat easy to apply.Illustratively it lets us easily know about say the intelligence of a case by diligently & selectively studying & evaluating the symptoms of verbal and say mathematical ability of a subject. And hence it is treated as a panacea for most cases where direct evaluation of a phenomenon such as (say) ‘propensity to purchase’ are to be studied.
However a ‘Caveat’ here is in order . Spurious factor analysis results is a common phenomenon too. Also different versions of factor analysis are required depending on the type of data etc. Data parameters inter-co-linearity is an another rouge element … and so it (Factor analysis) remains an experts forte and not a students play toy in data analysis. It is mutating over time, and its applicability is getting more complex over the years. Modern version application of Factor Analysis is a far cry from its initial versions created & discovered more than a century ago.But none the less it one of our few standardized reliable answer so to say to the unfinished exploratory technique of qualitative study processes. Delivering the un-measurable in accurate decimal wise measurable terms.

Leave a Reply

Your email address will not be published. Required fields are marked *