Jul 12, 2012

Why such vast gap exists in statistical interpretation of a single fact?


Statistics considers raw data while people tend to value associated sentiments.

If you are going for a primary survey with a standard template of closed ended questionnaires or some kind of table to be filled with numbers you might not be doing real justice to the survey. You might have a reasonable amount of data at end of the day to conclude your assignment or research, but you won’t be able to do the realistic assessment of the situation. You might fulfill the acceptable standards of data gathering; you might even publish the result and can even bag an award or two, but if the very people who got surveyed are not being able to relate to the statistical conclusion, or if they do not validate it, all the efforts of data collection and analysis goes to vain.

You will be surprised to realize that still at many instances the situation shown of paper and the actual situation experienced by the population is drastically different, though data for situation analysis was gathered from the same population or community. Then why this vast gap of interpreted and locally experienced situation exists? Why this constant apparent difference of opinion between government’s point of view and what media and NGO’s showcase about the city, region or community etc. There are few possible reasons, first is that statistics is very flexible domain, though outcome seems very impactful there is enough scope to mould or influence the direction of outcome, still staying within legally of intellectually acceptable limit. The kind of data you select to gather, variety of assumptions you take into consideration, mode and tool of statistical analysis you choose to go with, time span considered for analysis, selection and heterogeneity of sample itself, there are so many such variables that one can interpret and influence the outcome in any possible direction depending on the purpose or intention. Larger the data set, greater the flexibility of interpretation.

Government being a powerful entity with ample manpower and resources has generally upper hand when it comes to data collection and interpretation and usually less probability of misrepresentation, but again it depends on what they want to prove or convey. Media and NGO who generally lack enough data due to manpower constraint to prove a fact applicable to a larger domain generally tend to rely on emotional values attached to the subject or region and sentiments of population or communities to prove their point. They generally choose small set of data though having high perceived or emotional value and exploit the associated sentiments. On one hand there is ample scope of playing with data and on the other hand equally or more or less scope to play with people’s emotions and sentiments but ultimately both leads to distortion of fact. And this show of constant blame and justification goes on and on. Actual fact seems to lose the significance while this argument is on, later no one really bothers to reach to the actual fact.

Unless statistics as a knowledge stream would device a method and make it a mandate to consider or quantify and incorporate the emotional and sentimental values attached to the subject of research and aspirations of communities to the extent possible along with some kind of post analytic appraisal and indigenous validation method, data representation would remain deceptive and at the mercy of decision making individuals or groups. Meanwhile we as a layman can mitigate this doubt of data misrepresentation by keeping a questioning, skeptical or neutral attitude towards impactful data being presented day to day from both government end as well as media and social welfare representative’s end, without being panicked, without arriving at immediate conclusion, cross verifying the fact from different sources, applying our own head and intuition, knowing and reminding them that their job is to inform us not to instigate or incite us. You are capable enough to judge yourself; all you need is a fair amount of fair data and statistics or simply a fair representation.

By : Anoop Jha

No comments: