Understanding how to use Statistics
The World of Statistics is not so Cut and Dry
People are bombarded by statistics on a daily basis and many fall for them, hook, line and sinker. Why? Because most people do not understand how statistics work.
Funny thing about statistics...they are usually very biased, slanted or give a poor representation of the real world and most people's situations.
See, I took a few statistics classes in college and they taught us some interesting things...
People LIE about statistics
Statistics are taken from a SAMPLE demographic of usually 1200 people (and there are how many people in the US??? A little disproportionate, don't ya think?)
There are books about how to lie with statistics, including the book "How to Lie with Statistics" (Darrell Huff)
Statistics take for granted that there is a "typical" price or behavior, but, remember, it is out of the SAMPLE of 1,200 - does it apply to you? You actually have less than a 3% chance that it does (based on US population of 301,139,947 as supplied by CIA World Fact Book - July 2007)
Statistics often leave out pertinent information
For instance, a scholarly journal in 1995 stated that "every year since 1950 the number of American children gunned down has doubled." (The author claimed that the statistic came from the Children's Defense Fund.)
Hmmm, let us take that into consideration for a moment. Suppose just 1 child was gunned down in 1950. In 1951, the number of children gunned down would have been 2. In 1952, the number would have been 4 (remember, we are doubling!) and so on. Well, by 1965, it would have been 32,768 children gunned down, but in 1965, the FBI only identified 9,960 criminal homicides in all of the US, including adult and child victims combined. To jump to the chase, by 1995, when this article was published, the annual number of children gunned down would have been more than 35 trillion. Where are all of these extra children? And why haven't we heard of this mass gunning down of children that escalates so exponentially?
Because it is a part of flawed statistics and an assertion to my point that people use statistics to twist "facts" to suit their soap box rally or rant at the time. It is usually very weak when you look at the specifics - and this includes governmental and non profit statistics.
In truth, the Children's Defense Fund did indeed publish a statistic regarding children being gunned down. In The State of America's Children Yearbook - 1994, it was stated, "The number of American children killed each year by guns has doubled since 1950."
Difference in wording equals different meaning. It is just a matter of twisting the statistics to suit your needs.
So, how do you make sure that YOU do not fall for faulty statistics?
1. Be wary of statistics spouters who fail to direct you to the exact location where you can view the statistics for yourself. (book, article, journal, link)
2. READ the fine print that explains the sample (for polls and certain statistics) and situation, including environment, geographical region, ages, etc.
3. Don't believe everything that you read. If your source does not have information to divulge the conditions of the survey or poll, does not provide a link to how the poll or survey was conducted or other pertinent information that allows you to discern the validity of the poll or survey, disregard it as bunk.
4. Ask these questions:
Where did the data originate from?
Who conducted the survey?
Does the administrator of the survey have an ulterior motive for slanting the results in a particular direction?
How was the data collected?
What questions were asked?
How were the questions asked?
Who asked the questions?
5. You should be careful of comparisons. When two things happen at the same time, it does not mean that the two things are necessarily related. This is basic logic, but many people who are not skilled in logic and statistics rush to put, what they erroneously think, are two and two together. It ain't equaling four, that is for sure! Many people use this slanting of information to "prove" their point. Politicians are famous for this, but people who are desperately grasping at straws to substantiate an argument are very guilty of it as well.
6. Watch for numbers that are taken out of context. Affectionately referred to as "cherry picking," this slanting refers to adjusting the analysis so that it concentrates solely on the data that supports a specific claim and ignores or shuts out everything else. In other words, certain "facts" that suit the person's claim are "cherry picked" or selected while other pertinent information is swept under the rug.
One of the primary things that statistics will teach us is that there are no averages. If 50% of pet owners are responsible, then 50% of pet owners must be irresponsible. It does not help to change the definition, there must always be a population that is 50% below and that is substantiated by bell curve graphs.
This, in turn leads us to the next issue where people have problems with interpreting statistics. They want to make the statistics fit the normal distribution. However, this is significantly flawed because there are non-normal distributions. So what happens is that the statistics that are used for normal distributions are usually not appropriate for distributions that are blatantly non-normal.
Finally, most people do not understand the terminology behind statistics. They mistakenly assume that the term "mean" means the same as "average." This is inherently wrong. Mean is a mathematical term while the word average is used to describe a person or data item. In mathematics, however, average means "a number that typifies a set of numbers of which it is a function." When used in the mathematical context (and the context in which it is used when referring to statistics), average can mean "mean," "median" or "mode."
So, what do these terms mean?
Mean - a number the typifies a set of numbers
Median - the middle value of a distribution
Mode - the value or item occurring most frequently in a series of observations or statistical data
These two examples will aid in understanding this:
2, 5, 5 6, 9, 12, 13
Mean - 7.71 (typifies the set - add the numbers, divide by 7)
Median - 6 (is the number directly in the middle of the distribution)
Mode -5 (occurs most frequently in the set)
4, 5, 5, 5, 8, 12, 86
Mean - 17.857
Median - 5
Mode - 5
Mark Twain said, "There are three kinds of lies: lies, damned lies and statistics." Numbers are provocative and have a certain power, but in the wrong, uneducated hands are nothing more than a mess. Even accurate statistics can be used to try to strengthen inaccurate arguments. And honesty and accuracy is therefore compromised.