Tuesday, August 21, 2012

Polls and Surveys–An Insult To Your Intelligence

I am amazed at how many people allow their opinions and their passions to be directed merely because someone stated that “poll x claims that x% of people feel / believe ___whatever____”.

It seems that a majority of people cite these polls without ever understanding what went into the poll, how the questions were directed and what the ultimate intentions of the pollster / media were.

While many pollsters and media claim to be belief-agnostic (whether in politics, business, religion or anything else) if we believe that they can turn off their own feelings as they direct the course of their business then we are also acknowledging that they are not human.

Consider this hypothetical poll.

Let’s assume that I am conducting a survey in your city in which one of two choices must be made.

Scenario: I am driving down a steep hill when suddenly my brakes fail.  Ahead of me is a sharp curve and I know I can’t make the turn.  A crash is imminent and I have been given a choice as to how the crash will manifest.

Choice 1. There is a five-year-old child on one side of the road, playing happily and I will steer my vehicle onto this side of the road.

Choice 2. There are ten five-year-old children (including your own) on the other side of the road, also playing happily and I will steer my vehicle onto this side of the road.

I have seconds to choose which side of the road I am going to crash into and must choose one of them.

In my hypothetical poll, 85% of people choose the lesser of two evils, steering the vehicle onto the side of the road with the lone child.

1% choose choice 2 because they are not taking the survey seriously and want to see how I will react.

14% get so incensed over the survey that they walk away and refuse to answer.

Now, depending on my intention for conducting the survey, it would be completely honest to make this headline announcement.

85% of people in the city of ___your city____ would wilfully take the life of a child if they had the chance to do so.

It’s a pretty outrageous statement, isn’t it?  However, while you don’t really feel this way, the poll can be demonstrated to prove this point.

Now the pollsters conducting the survey don’t tell people taking the survey whether or not steering onto either side of the road automatically means the death of whoever is on that side of the road.

The survey results also don’t state that people taking the survey were forced to choose a choice that had the potential for death; that there was no option 3 which would have produced no deaths at all to innocent children.

But that’s not the point.

The point is that I will have made a statement using a “trusted” survey source with irrefutable mathematics citing that the poll is accurate within 98% (plus or minus 2 percentage points) 19 times out of 20 and so it must be accurate.

And yes the stated poll results will be accurate, in a twisted sort of way.

But while technically accurate, does it tell the truth?

Does it REALLY summarize what you or the average good person actually believes?

Probably not.

The bottom line is this.

Do you prefer to be informed based on how others want you to think or do you prefer to be informed based on your own assessment of knowledge, data and facts?

The answer to this question determines whether you are a leader in this world or you are one of the many sheep.

The answer also identifies whether you are living your life to the fullest or you are a stepping stone for someone else who is living their life to the fullest.

Which do you choose to be?

What are you doing about it?

Or are you waiting for a poll result to help you answer these questions also?

In service and servanthood,

Harry

2 comments:

  1. 19 out of 20 is 95%

    ReplyDelete
  2. The great challenge with ego today is that people, in their haste to prove someone wrong, only have attempts at correction as their contribution to a discussion.

    To say "accurate within 98% (plus or minus 2 percentage points) 19 times out of 20" is to suggest that the poll is accurate within 2% (the sampling error) with a confidence interval that is 19 times out of 20 or 95%.

    So 95% of the time, the sampling error is 2%. The other 5% of the time, the sampling error may be greater but is dismissed as anomalous. The greater the confidence interval (which is difficult to obtain) the more consistent the accuracy of the poll.

    Thanks for contributing.

    Create a great day!

    ReplyDelete