Lies, Damned Lies, and Twisted Statistics

Peter B. GiblettStarred Page By Peter B. Giblett, 8th Sep 2014 | Follow this author | RSS Feed | Short URL http://nut.bz/bkbe0wjc/
Posted in Wikinut>Writing>Society & Issues

People love to quote statistical evidence as a proof that their argument is true or valid, yet statistics can be misquoted and misunderstood so easily. The statistical train is in motion, but what do we really know about the results they can provide? Take any two people with an agenda and they can prove the opposite from the same statistical information.

Tied to a Railway Track

You are tied to a railway track and told the locomotive is traveling at one hundred kilometers per hour and will reach this point in 12 hours time. Six hours later you are told there is good news - the train is going 30% slower than it was before - is this really good news? Of course it is not because the train is still coming to kill you, it will simply take a little longer to get there. The trouble is statistics can seemingly be bent to mean anything to anyone any time you choose them to and the reason is figures often beguile the reader into thinking something is true - when the opposite is actually the truth.

Consider for one moment the political party leader that states that the voting public now understand our position - "we gained 10 percent of the popular poll" yet the truth is that at the previous election they attained 9.4% and on this occasion they achieved 9.6% of the vote - the difference is hardly significant in real terms but it does mean that if the figures are rounded to whole numbers then they achieved 10%of the poll . Huh!

Truth is you can prove anything you wish through clever use of numbers and some very dubious people set about doing precisely that every time they quote statistics.

("Train Conductor" image used in this section from a Clip-art collection purchased by the author)

Falling sea levels?

Just imagine what I thought when I saw a statistic that stated the rising sea levels have reduced by 30% over the past 9 years.

Does that statement mean that sea levels are no longer rising? No! Quite the contrary they are still on the rise, but if we believe these figures they are doing so at a reduced rate. The simplest way to understand this is by knowing whet is happening in reality - the US National Oceanic and Atmospheric Administration (NOAA) states that the sea-level is rising by about 3.3 millimeters per year currently. So if we follow this mathematics the sea level rise in 2003 (prior to the 9 years) would have exceeded 4 mm per year.

Now I need more figures to compare against in order to prove or disprove this supposition.

Compare these figures to a study by Church and White (A 20th century acceleration in global sea level rise) concluded that between the years 1870 and 2004, global average sea levels rose at a rate of 1.46 mm per year, for a total of 195.64 mm in 134 years. It is always important to understand the real impact of phraseology before making a statistic mean something it does not.

Now let us ask the obvious question - which rate is greater 1.46 mm per year or 3.3 mm per year? and you do not have to be a statistical genius to understand the answer.

I would have to conclude that the inhabitants of Florida should be greatly concerned about the inevitability of flooding and it is not a case of if but when this happens.

Yet the average resident of my city, located just about 3,500 ft above sea level will probably care very little about such statistics, unless it ruins their vacation some time in the future, so campaigning about rising sea levels will probably not get you elected onto our city council.

Unhelpful Statistics

The problem of reading multiple statistics is that it can be like comparing apples to oranges - two different studies held at the same time produce two different results and it is all a matter of the questions asked and how results are measured and interpreted.

You wish to know which brand of washing machine is the best one to purchase and you will come across a baffling array of figures that all bobble the mind - one uses 90% less water but another completes the wash cycle 40% quicker and a third uses 15% less soap powder when used on a cold-wash cycle.

Trouble is no two figures are comparable and probable none helps the consumer find the best product when they are seeking a new product. In this case the old saying "buyer beware" is probably the best policy.

Proven Statistics?

Statistical proof is a rational demonstration of degree of certainty for a proposition, hypothesis, or theory usually supported by a statistical test based on a mass of supporting evidence. Trouble is inferences can be drawn from certain test scores and this regularly happens even when those drawing the inferences do not understand the mass of the data they are looking at.

Statistical methods are key to understanding the facts that are available, too many people jump to conclusions without using any such methods - often without any care or knowledge of the assumptions made before the study commenced and reasons why the results may have been skewed in favour of one proposition or another.

A common demarcation between scientific proof and non-science statistical applications needs to be understood, scientists do not use statistical proof as a means to attain certainty but do so in order to explain theories. In some respects (according to Stefano Gatti page 277) statistical proof offers a kind of proof of a theory's falsity and the means to learn heuristically through repeated statistical trials and experimental error. It should also be noted that statistical proof can at times have implications for the legal burden of proof.

Axioms

The statistical proof axioms and important because they explain such proof in relation to probability and is included here for anyone who is mathematically minded - this copy is courtesy of Wikipedia. I know that for some readers these equations are at a glance too much to understand - the point is that when using statistics it is all too easy to infer proof when none exists.

Randomness and chance should be important aspects of statistical analysis yet to many they are to be ignored.

Proven or Not?

One of the things that is true is that people jump on statistics as if they are the answer to their proposition - yet it is all too easy to get caught in the quagmire of statistical anomalies.

The following are some of the more recent articles that I have published on a variety of topics, which is now more than 470 articles on Wikinut.

Each person has a unique voice and Wikinut is great a place for you to share some of your wisdom, insight and knowledge, you could start by adding a comment, but perhaps you need something more in which case should join Wikinut, write then become connected to others who share a passion for writing, supporting one another, and learning on Wikinut.

Tags

Buyer Beware, Disprove, Lies, Political, Prove, Prove Anything, Railway Track, Real Impact, Sea-Level, Statistical Information, Statistical Proof, Statistics, Unhelpful Statistics

Meet the author

author avatar Peter B. Giblett
Author of "Is your Business Ready? For the Social Media Revolution"

Social media consultant, with C-Level background.

Share this page

moderator Steve Kinsman moderated this page.
If you have any complaints about this content, please let us know

Comments

author avatar Sherri Granato
8th Sep 2014 (#)

Insightful and interesting article.

Reply to this comment

author avatar Peter B. Giblett
9th Sep 2014 (#)

Having been involved with statistical analysis for the majority of my life it seemed appropriate that I tell people about the challenge of statistics - because they have been used and abused recently here.

Reply to this comment

author avatar Retired
8th Sep 2014 (#)

You make some excellent points here, no doubt inspired by the recent debates over global warming!

Another aspect is how statistical evidence is presented. It is very easy to convey a false impression by creating charts or tables that present information in a misleading way. I'm sorry to say that Wikinut is guilty of this to some extent - the daily charts of user views and earnings make the mistake of using curves rather then columns - curves are inappropriate for this type of information.

Reply to this comment

author avatar Peter B. Giblett
9th Sep 2014 (#)

Yes John there were a number of articles that leaned on statistics, so I thought it was time to see if they bore out, some did and some did not and as you say it is all about presentation.

Reply to this comment

author avatar Ptrikha
10th Sep 2014 (#)

Hi John, How come curves are inappropriate, if they are based on actual values?

Reply to this comment

author avatar cnwriter..carolina
9th Sep 2014 (#)

so glad you bring all this to our attention...we know how falsified so much information is..after all they have to give the great unwashed public a panacea....or some such thing ...ha!

Reply to this comment

author avatar Peter B. Giblett
9th Sep 2014 (#)

I guess the great "unwashed" public are searching for answers on a number of topics and using statistics seems to be the easiest way to prove/disprove a point.

Reply to this comment

author avatar n.c.radomes
9th Sep 2014 (#)

Thank you for bringing this truth out, but I think not all people do the quoting just to lie.

Reply to this comment

author avatar Peter B. Giblett
9th Sep 2014 (#)

It was not my intention to accuse people of deliberately lying, it is simply that they latch onto a figure that they see as proving their point but 95% of the time that figure is not based on reality.

Reply to this comment

author avatar JayeByrd
9th Sep 2014 (#)

The Obama administration keeps saying the unemployment percentage is going down. That's the apples.
The oranges are the real number of employed is going down, too. Which statistic tells us the most?

Reply to this comment

author avatar Peter B. Giblett
9th Sep 2014 (#)

Too many politicians, from too many political parties in too many different countries have caused so much confusion over these types of statistics.

Reply to this comment

author avatar Judy Ellen
10th Jan 2015 (#)

This is one statistic that has always bothered and upset me! I know for a fact that Obama has no idea statistic wise how many people are actually not working because they can only keep track of those who have recently applied for unemployment. Our government does not know or care about all the people who have been rejected and denied unemployment due to either quitting or being fired. They also don't know how many people have just plain given up and remain unemployed because after so many months of being unemployed no one wants to hire them! I could go on and on but maybe I should instead write an article about this one! Obama uses those false unemployment statistics to make himself look good.

Reply to this comment

author avatar Retired
9th Sep 2014 (#)

Well stated, Peter. On a simpler level, meaningless statistics expressed in percentages are frequently used in advertising. Shampoo that is 90% more effective at clearing dandruff, foods that contain 30% less fat, cars that are 20% more fuel efficient! Compared to what? These percentages are simple 'statistics' aimed at dumbing the audience. One of the reasons I stopped working in advertising.

Reply to this comment

author avatar Peter B. Giblett
9th Sep 2014 (#)

The shampoo one is a nasty trap because one product springs to mind whose chemical composition causes dandruff when you switch to a competing product. Yes advertising is one of the worst users and abusers of statistics.

Reply to this comment

author avatar Phyl Campbell
9th Sep 2014 (#)

I wanted to make up a statistic about this article, but my brain just won't let me tonight. I could make up another statistic about how often my inability for a quick off the cuff remark saves people from figuring out whether to laugh or groan, but I'm 100% sure no one really cares. Excellent points made.

Reply to this comment

author avatar Peter B. Giblett
9th Sep 2014 (#)

So you produced a statistic after all. Thanks.

Reply to this comment

author avatar Retired
9th Sep 2014 (#)

Winston Churchill once said that the only statistics you can trust are the ones you falsified yourself!

Reply to this comment

author avatar Peter B. Giblett
9th Sep 2014 (#)

I can see Mr Churchill saying such a thing and I'll be frank - this is one of the few occasions I would agree with him.

Reply to this comment

author avatar Chip Greene
9th Sep 2014 (#)

Excellent article Peter. My favorite abused and misleading statistic is a warning for those who don't use seatbelts when they go for a ride. Statistics show that most accidents occur within one mile of your home. True, because most people spend most of their driving time within one mile of their home.

Reply to this comment

author avatar Peter B. Giblett
9th Sep 2014 (#)

True. I would also hasten to say that most such accidents occur even closer to home than that.

Reply to this comment

author avatar Sivaramakrishnan A
9th Sep 2014 (#)

"There are three kinds of lies: lies, damned lies, and statistics." is attributed to Mark Twain. Statistics is used to hoodwink and confuse the layman, but even the so-called experts remain confused. Even facts are twisted to favor their interests like whether there was indeed WMD in Iraq. Even wars are started on wrong premises. Thanks Peter for some frank talk - siva

Reply to this comment

author avatar Peter B. Giblett
9th Sep 2014 (#)

Siva, I looked up that attribution and there is some confusion over its source and hence said nothing but instead talked about twisted statistics.

Reply to this comment

author avatar C.D. Moore
9th Sep 2014 (#)

I agree statistics can't be trusted!

Reply to this comment

author avatar writestuff
9th Sep 2014 (#)

Interesting and informative post.

Reply to this comment

author avatar Retired
9th Sep 2014 (#)

Aha, a man after my own heart! Yes, I'm a student of the Benjamin Disraeli / Mark Twain school of thought too. Thanks for going some way to exposing statistics for what they really are.

Reply to this comment

author avatar Peter B. Giblett
9th Sep 2014 (#)

Lewis, I have worked with, produced, bent, fashioned twisted, and ironed out statistics for most of my life so I know how they can and should be used.

Reply to this comment

author avatar Ptrikha
10th Sep 2014 (#)

In IT world, statistics are always compared to what we call as a "Baseline" or "Benchmark" statistical measure, especially for Performance Metrics. We cannot have a 45% gain or a 20% loss over "nothing". Well illustrated Peter.

Reply to this comment

author avatar MarilynDavisatTIERS
10th Sep 2014 (#)

Good evening, Peter; good examples throughout your article. I’ll give you a personal truth. A competing recovery house claimed a 90% recovery rate; however, looking under the surface, many of those “sober participants” were back in jail without access to drugs or alcohol - contrary to some popular misconceptions about availability of drugs in jail. I on the other hand did not include those incarcerated in the "win" column and only came in at 87% - granted, it was still more than twice the national average - yet another figure, but a more honest reflection of numbers and outcomes. When a board member asked me to include those from our current or graduate population that reoffended and were in jail, I balked, seemed misleading and untruthful. I won and our statistic felt honest. ~Marilyn

Reply to this comment

author avatar snerfu
13th Sep 2014 (#)

If you cannot prove A + B = A + B , then there is no point in using figures to depict anything. I say this from the layman point of view because statistics has hundreds if not thousands of interpretations.
It is not always easy to use pie-chart or bar charts for all the data that a big site like WIkinut has.
But statistics gives the best unbiased view of a scenario since there are no words there, only figures.

Reply to this comment

author avatar Peter B. Giblett
16th Sep 2014 (#)

Snerfu sir - statistics should give an unbiased view - but too few people understand what they mean and how they work.

Reply to this comment

author avatar vandana7
7th Nov 2014 (#)

Yes, it is true. Not many understand it the way it should be.

Reply to this comment

author avatar Judy Ellen
10th Jan 2015 (#)

I agree and don't take any statistics as legitimate or factual because I know human nature and how facts can be manipulated. If I really want to know something I will do the research myself! This is also why I mute the commercials on TV because I refuse to be brainwashed by listening to the same advertisements over and over again. I have them all memorized anyway and wish I didn't! Excellent article!!

Reply to this comment

author avatar Peter B. Giblett
10th Jan 2015 (#)

Thank you Judy - I fast forward the ads because I rarely watch live TV, but those are frequently the worst use of statistics.

Reply to this comment

Add a comment
Username
Can't login?
Password