Surveys and Statistics

Surveys and Statistics was Originally Posted on October 15, 2012 by lavarock

It is always interesting to see people hang their hats on statistics and surveys. While most are accurate, sometimes they can be very wrong.

If I wanted to promate a product, my ad might say that 4 out of 5 people would buy it again. I could find 4 people who would buy it again and 1 that wouldn’t and use those numbers. While accurate, not statistically valid across a larger population of the 5 people I am using. When you see these claims of percentages, do you investigate further? Probably not.

When you heafr that 9 out of 10 doctors recommend a drug, what deos that mean? The doctors might have been given free trial sizes of a drug. If they hand them out to patients, are they not recommending that the patient try it? Does recommending a product mean that they think this product is more effective than all others the doctor has used or is familiar with? You are not told the whole story.

If you think these are far out examples, let me tell you a very real example.

Many years ago, Hewlett Packard announced that they were letting go about 1/3 of their workforce. 30,000 or 40,000 jobs are targeted. As employees, we know that they will be offering early retirement to people and others may be targeted in other ways. After all, we dis perform yearly job reviews to help rank employees on a scale of 1 to 5. The first thought would be that people in the lower band are not very good workers but we also realize that they are probably not in the higher pay bands, thus letting go 5 people in the lowest performance group might not save as much as someone in the better performance range. Also, as someone stays with the company longer, they tended to make lots more money.

So employees are sitting there waiting to see who will be “let go” or offered an early retirement.

About this time, Hewlett Packard announces that they want to purchase Compaq, a computer company who is in trouble. We realize that HP may be shedding employees to absorb the employees of the company they are purchasing.

So as an employee, I receive an email from “Corporate” asking me:

1) How long I have been with the company

2) Do I think buying Compaq is a good thing

Normally, a survey like this could be a good thing and help corporate make a decision. On the face of it, it seems as though they might want to see what the old timers thing of the decision.

Had the questions come on an envelope to my home from a survey company, employees might have accurately answered the questions. However, they came in inter-office email where our responses could be directly and individually tracked. Many employees replied that they thought it was a great idea, even though they did not. I personally failed to reply.

So Mrs. Fiorina eported to the press that employyes were behind the purchase even though manyof those same employees lied. Had the company told the employees the criteria used to decide who to let go, or used an anonymous survey, they might have gotten better and more accurate survey.

Even in voting, there are people who have registered in the opposite political party, so they can vote for the person least likely to be able to win against the person they really support in their own party.

I get a small amount of money in a program that asks survey questions. Sometimes the surveys ask questions that get me into a category that is not accurate. I might be asked what type of business I work for (and the answers may not allow me to say that I own the company). The survey may ask what my business does. Sometimes food and beverage is appropriate, sometimes manufacturing or retail sales or e-commerce.

To be accurate, my “business” is a lot more diverse that others. I own a farm, am the only employee, am responsible for the purchase and use of all products, am responsible for all accountuing, specify and purchase all software, hardware and services. I create a beverage that is sold on the internet and am also the webmaster. besides that product, I also resell web hosting services, manage websites, resell credit card processing services, resell Voice over IP servoces and can install and support products such as that. I repair computers and support software and hardware on PC’s. I recently also resold travel tours on the island whilem working in the hospitality sector for another company.

So when they ask me to fill out a survey, which real business are they talking about? Who do they really think they are surveying? Someone whose only income comes from manufactoring a product or me, who has more businesses than Walmart has stores?

Perhaps someone will do a survey to see who is really replying to surveys.

In any event, as I say, I tend to take every survey and result with a grain of salt.