Honesty in Politics

Posted on Sunday, 28 August 2011


Questioning a politician’s honesty seems like a pointless endeavor yet one website dedicates its resources to this task. In an era when news channels are pandering to our emotions and working on obvious agendas, PolitiFact acts as a mediator and fact-checker for the raconteurs of our society.

On the PolitiFact site, The Truth-O-Meter® rates a statement and classifies it into one of six categories (as described on their site):

  • True  – The statement is accurate and there’s nothing significant missing.
  • Mostly True – The statement is accurate but needs clarification or additional information.
  • Half True – The statement is partially accurate but leaves out important details or takes things out of context.
  • Barely True – The statement contains some element of truth but ignores critical facts that would give a different impression.
  • False – The statement is not accurate.
  • Pants on Fire – The statement is not accurate and makes a ridiculous claim.

This system is built with the understanding that the truth is not always easy to define. Sometimes statements have modicums of truth to them that are wrapped up in agenda building lies. Sometimes lies are bolstered by small truths to increase their credibility.

This scale helps to quickly identify the truth in an individual statement, but what happens when you take the ratings and compile them? When this data is compiled and evaluated as a whole, does it tell us anything about our society? I decided to find out.

Experiment Procedure

When I first came up with the idea to do this, I went through the entire list of people and copied the data by hand into a spreadsheet. It was tedious and didn’t amount to anything at the time; however, the idea never left me. Earlier this year, I decided to write a program to scan the site for me and put the information into spreadsheets. I assigned numerical values to each of the ratings; True would be equal to 1 and False would be equal to 0. The other values were set to even intervals between those. Pants on Fire is a special case that is arguably worse than a simple lie so I gave it the value of -0.5.

With the rating scale set, the next step was to organize the data into meaningful categories. I decided to split the data by: political party, state, and house or senate. There was some data that did not fit into these categories but I felt that these categories would be the most revealing. Using these categories, I then calculated averages. To do this, I took a basic weighted average of each person or organization by multiplying the number of each type of statement they gave (true, half-true etc.) with the value I assigned to each type of statement.

For example if John Doe had 5 true statements, 3 half-true statements, and 2 lies the formula would be:

Weighted Average Equation

This gave me an average for the person. I then used this value and similar values of the data in each category to find averages for the set.

In finding the averages for the set, I could have instead found the total number of each type of statement within the category and found a total weighted average but I found that doing this would lead to tremendously talkative politicians throwing off the average. While looking at the results, keep in mind they are averages of the subject’s honesty, not averages of total statement honesty.


The data is given as a percentage. The higher the percentage, the more honest the statements in a set were. I captured the data on July 15, 2011.

Total Data

  • 35.90% Honest
  • 1255 Data Points
  • 3763 Statements

This is a compilation of all the data on the site that can be traced to a human or an organization. It does not include such things as chain-mail.

Total Republicans

  • 29.00% Honest
  • 393 Data Points
  • 1409 Statements
Everyone who is identified as a Republican on the site

Total Democrats

  • 40.53% Honest
  • 360 Data Points
  • 1268 Statements

Everyone who is identified as a Democrat on the site



  • 49.33% Honest
  • 53 Data Points
  • 394 Statements

Senate Republicans

  • 42.40% Honest
  • 26 Data Points
  • 306 Statements

Senate Democrats

  • 56.00% Honest
  • 27 Data Points
  • 88 Statements

House of Representatives


  • 33.01% Honest
  • 101 Data Points
  • 310 Statements

Senate Republicans

  • 25.19% Honest
  • 60 Data Points
  • 193 Statements

Senate Democrats

  • 43.18% Honest
  • 41 Data Points
  • 117 Statements


For this, I only compiled data for states that had at least three data points. These are not just state politicians; they are any politician that hails from a state. It is more of a measure of what kind of politicians a state produces rather than an evaluation of a state’s current politicians. 

  • Alabama: 4.17% Honest
  • Arizona: 22.89%
  • Arkansas: 36.63%
  • California: 50.95%
  • Colorado: 50.83%
  • Connecticut: 76.46%
  • Delaware: 49.81%
  • Florida: 41.47%
  • Georgia: 43.05%
  • Illinois: 42.46%
  • Indiana: 15.38%
  • Iowa: 4.17%
  • Kansas: 51.57%
  • Kentucky: 52.94%
  • Louisiana: 54.17%
  • Maryland: 66.07%
  • Massachusetts: 36.81%
  • Michigan: 64.58%
  • Minnesota: 20.36%
  • Missouri: 28.13%
  • Nevada: 52.73%
  • New Hampshire: 58.33%
  • New Jersey: 30.42%
  • New Mexico: 51.04%
  • New York: 34.29%
  • North Carolina: 18.89%
  • Ohio: 49.38%
  • Oklahoma: 42.50%
  • Oregon: 31.09%
  • Pennsylvania: 45.83%
  • Rhode Island: 32.48%
  • South Carolina: 28.67%
  • Tennessee: 60.00%
  • Texas: 40.09%
  • Utah: 66.67%
  • Vermont: 42.50%
  • Virginia: 43.44%
  • West Virginia: 33.33%
  • Wisconsin: 22.72%
  • Wyoming: 37.50%
  • Number of States evaluated: 43
  • Max: 76.46% – Connecticut
  • Min: 4.17% – Alabama
  • Average: 42.15%
  • Standard Deviation of States: 16.55%

People of Interest

From the data, I choose a few people who have a large political influence and a large number of statements. I know there are other Republican Candidates however, these are the only ones which had a large enough number of statements to evaluate.

Issues with the Data

Before I get to the conclusion of this, I figure I should cover some issues that many of you will see with this whole experiment.

The data is [biased, not complete] – There is nothing I can do about this. I found the biggest issue with the data was that some politicians don’t say much. The majority of the statements come from a minority of the people. Give your representative a call and make them more active in defending your freedoms.

The data is old – I compiled the data on PolitiFact on July 15th. Since that date, the debt-ceiling debates have occurred. This may have changed the data some and I would be willing to update this in the future. However, I have been tracking this data since last November and little has changed in the averages.

Politifact.com is biased – This may well be; the results do tend to favor one side over the other. Whether this is evidence of bias or just the truth is up for you to decide. They seem mostly centered politically however; no person or organization is without some bias.

This process is flawed. – In some ways yes. I am assigning numerical data (or values) to somewhat subjective ideas and them compiling it. Of course it is not perfect. I do not claim that this experiment is the last word on the honesty of our politicians; it is just an interesting concept to explore.

You are biased. – I am biased, which is why I tried to explain the process as much as possible and give you the data. If you wish to check my work, all the data is freely available on Politifact.com.


Nothing in the data was surprising. We all know politicians lie. The point of this experiment was not to prove this or to point fingers at the liars. This experiment hopefully gave you an opportunity to reevaluate your views on politics. The important question is not how much politicians lie; the important question is: why do they lie? … The obvious answer is because they want to please their constituency, but how does lying to us please us? They say what they believe the public and perhaps more importantly, the media, would like to hear. We are the people voting for them though; do we have unrealistic expectations of our politicians? It seems so.

In recent polls, a “generic Republican” is ahead of Obama but any specific Republican is behind (polls of course fluctuate but that seems to be the trend). This is not the result of a poor field of Republican candidates. This is the result of each individual having a different idea of what an ideal candidate is. No one candidate can please everyone yet they still have to try to appeal to as many people as possible. This results in a simplified platform where little is left but stories that bear little resemblance to reality. These stories can be considered as lies that fit comfortably on PolitiFact’s scale.

As citizens of the United States of America, it is our duty to be well-informed and make educated decisions when we vote. Nevertheless, we all vote for liars. We vote based on dreams and we are disappointed when our dreams don’t come true. We idolize our representatives and then trash them when we realize they are only human. Politicians lie because we want them to. They lie because we want to hear stories about how great things will be if they get elected. We are all optimists when we vote; or at least optimistic that our guy will do better than the other guy.

PolitiFact is owned and copyrighted by the St. Petersburg Times. The author claims no affiliation with this source. 

Posted in: Politics