Pharyngula

Posted by: MolBio | January 23, 2010 12:40 PM

Polls are more a psychological thing than anything else.

Structure leading poll question you know target audience will go for.

Make target audience feel they're participating.

Report result and let hive-mind do the rest.

Posted by: Zeno | January 23, 2010 12:52 PM

My algebra students (especially business majors) often express exasperation at being required to take any math as part of their program of study. It's so irrelevant! Since algebra is a prerequisite for statistics, I like to point out that statistics is math for self defense. If they want to stop being duped by scoundrels with bogus numbers, they need to learn what is significant and what is not. Some of the complainers are mollified. All of them appear to be surprised.

Posted by: William | January 23, 2010 1:00 PM

The 'target' audience is key here. As we've witnessed, and participated in, polls can be manipulated by an 'invasive' audience holding views contrary to those initiating the query - example, the silly ass 'afterlife' turd - surprise!

Posted by: eddie | January 23, 2010 1:03 PM

Surely sixth grade is about 5 years too late.

Posted by: 'Tis Himself, OM | January 23, 2010 1:08 PM

Proper use of a graph, coupled with correctly chosen error bars and judicious use of a finagle factor, can be used to prove anything. Consider the last two panels of this cartoon.

Posted by: Glen Davidson | January 23, 2010 1:28 PM

But if evolution's so right it should be really easy to convince the public that it's right. That's the "argument" for the importance that a lot of people propagandized and socialized to react against evolution in fact do not accept it.

And so stupidity builds upon stupidity to keep from learning something as simple as the basic case for evolution.

Glen D
http://tinyurl.com/mxaa3p

Posted by: irarosofsky | January 23, 2010 1:35 PM

"Lies, damn lies, and statistics," said that great skeptic, Mark Twain.

Innumeracy also means that anecdotal evidence is taken as evidence, and the belief that an instance of something counter to a trend is the trend.

"It's snowing" or "This year is colder than last" means there's no global warming trend. I doubt that 1 out of 10 people could explain the difference between climate and the weather.

NASA recently concluded that the past decade was the warmest in recent history, but we all know they'r a bunch of Commies.

Another common fallacy is concluding that the non-saintly behavior of scientists mean their conclusions are invalid. Science is not a beauty contest or a democracy.

Statistics may be lies, but facts are facts.

"Climategate: Scientists are mean girls--so what's new?"

http://www.psychologytoday.com/blog/adventures-in-old-age/200912/climategate-scientists-are-mean-girls-so-what-s-new

Posted by: mothwentbad | January 23, 2010 1:40 PM

Seconding comment #11 - it's Cham, with an m.

Here's hoping PZ makes the correction at his first convenience.

Posted by: 'Tis Himself, OM | January 23, 2010 2:03 PM

"Lies, damn lies, and statistics," said that great skeptic, Mark Twain.

The quote is also attributed to Benjamin Disreali, Walter Bagehot, Arthur Balfour, Leonard Courtney (President of the Royal Statistical Society) and that prolific author, Anon.

Posted by: irarosofsky | January 23, 2010 2:16 PM

@"Tis Himself, OM"

I was careful not to say he originated the quote, only that he said it, which is documentable:
North American Review, No. DCXVIII., July 5, 1907

Or as Picasso, or was it T.S. Eliot? may have said, "Good artists borrow. Great artists steal."

Posted by: a_ray_in_dilbert_space | January 23, 2010 2:33 PM

Disraeli said the line about lies and statistics. I say that any fool can lie with statistics. What takes skill is using them to tease out the truth!

Posted by: Brownian, OM | January 23, 2010 3:35 PM

The second panel in the comic, Medieval News Network, reminds us of a program we used to watch on PBS that presented historical events, like the story of Joan of Arc, in the format of a modern newscast.

The newscasters were dressed in period clothing. It was an innovative way to present "boring" historical events in an entertaining way.

There is a children's show called History Bites that presents history lessons in comedy sketches.

Unfortunately it's Canadian, so it occasionally highlights events that are of no interest to most American viewers, such as every single that's ever happened that isn't the Revolutionary War.

Posted by: Knockgoats | January 23, 2010 3:56 PM

Unfortunately it's Canadian, so it occasionally highlights events that are of no interest to most American viewers, such as every single that's ever happened that isn't the Revolutionary War. - Brownian, OM

Now, that's just unfair! Americans are also interested in the 19141917-1918 and 19391941-1945 wars!

Posted by: Knockgoats | January 23, 2010 3:58 PM

How wonderfully ironic that "Lies, damned lies, and statistics" has itself become one of those quotes that's always misattributed!

Posted by: Legion | January 23, 2010 4:06 PM

Brownian:

There is a children's show called History Bites that presents history lessons in comedy sketches.

Sounds cool and interesting. Will check it out. Thanks.

Posted by: Brownian, OM | January 23, 2010 4:24 PM

It really is a clever show, Legion, but it is pretty Canadian. I wonder if many non-Canucks would recognise a parody of Don Cherry.

Posted by: Alverant | January 23, 2010 6:06 PM

I loved how in #3 the scores are 99% and 67% for a grand total of 166%.

Posted by: agenoria | January 23, 2010 6:50 PM

All the talk of questionable polls reminds me of this:

Opinion Polls: Getting the results you want

(Yes, Prime Minister, BBC)

Posted by: Davidpj | January 23, 2010 7:01 PM

I hesitate when using/recalling statistics to back up a point for precisely these reasons (it runs much deeper than polls, but they are a good case study). That said, considering how many students make it to university without realising what a percentage actually is, I'm not sure if it makes much difference how these numbers are reported.

My guess is that enough sportspeople use the phrase "giving 110%" that a report claiming 110% of people agree with a position wouldn't raise too many eyebrows (clearly, they just interviewed extra people for it!).

Posted by: Jadehawk, OM | January 23, 2010 7:15 PM

What do you mean? A spiral? Or just a pie chart with random numbers on it that don't add up?

Posted by: 'Tis Himself, OM | January 23, 2010 7:23 PM

Unfortunately it's Canadian, so it occasionally highlights events that are of no interest to most American viewers, such as every single that's ever happened that isn't the Revolutionary War.

Oh pshaw. Americans are interested in Canadian history. Many Americans know that Wayne Gretzky's number 99 was retired by all NHL teams. We're fully aware of the Toronto Blue Jays' back-to-back World Series wins in 1992-93. And who can forget during the 1988 Calgary Olympic games Canada became the only host nation in Olympic history not to win a single gold medal? See, Americans do know about important events in Canadian history.

Posted by: Feynmaniac, Chimerical Toad Superhero. | January 23, 2010 7:28 PM

Or the hilariously ironic poll reported by Fox News that asked "Did scientist falsify research to support their own theories on Global Warming". It added up to 120%.

According to the poll, 35 percent thought it very likely, 24 percent somewhat likely, 21 percent not very likely, and 5 percent not likely at all (15 percent weren't sure).

Fox News' graphics department added together the "very likely" and "somewhat likely" numbers to reach 59 percent, and called that new group "somewhat likely." Then, for some reason, they threw in the 35 percent "very likely" as their own group, even though they already added that number to the "somewhat likely" percentage. Then they mashed together the "not very likely" and "not likely at all" groups, and threw the 15 percent who were unsure into the waste bin. Voila -- 120 percent.

Posted by: SC OM | January 23, 2010 7:31 PM

This, of course, is the best pie chart ever,

:D I have to agree.

Posted by: craig.mcgillivary | January 23, 2010 8:21 PM

Consider this case:
A poll is conducted where the margin of error is 5% but 51% of respondents said yes while 49% of respondents said no. In addition 999 seperate polls are taken each with the same question and the same population and the same margin of error. Now you are given the first poll and are asked to predict whether yes is more common in the other 999 polls or no is more common.

If you don't predict yes you are making the same mistake as this cartoon. Given the available evidence yes is more likely to be the majority answer in the other 999 polls. In other words poll results within the margin of error do provide information.

Posted by: sandiseattle | January 23, 2010 9:25 PM

Amusing little comic. Need more (intentional) comedy here.

Posted by: Sauceress | January 23, 2010 9:40 PM

OT
News I'm sure all Haitians will be absolutely thrilled to hear (as will you all I'm sure)is that neither the AiG missionary translator nor any of the recently translated AiG propagander materials were damaged by the earthquake.

AiG Booklets Available in Haitian Creole

I wrote about (and posted some photographs of) AiG materials that had been translated into Haitian Creole—the main language in Haiti. We were concerned for the missionary who translated these materials and who lives in Haiti and what was happening concerning his ministry in the aftermath of the recent massive earthquake there. Matthew wrote and told us he is fine and also sent the following:

Thanks for lifting us up to the throne of grace in our time of need. God is faithful and has enabled us to keep going, and we trust that He will continue to do so in spite of the continued earthquakes and tremors. In answer to your question about the materials that we have translated and their availability in Haiti, we do have them available here.

Posted by: Creature of the Universe | January 23, 2010 10:28 PM

Joan Jett gets it right...in a different way. This stuff isn't hard...unlike maths and statistics can be.

...over and over...just for fun... HA!

http://www.youtube.com/watch?v=hdhonK8NMm8

Posted by: claire-chan | January 23, 2010 11:01 PM

Does it even take "a basic statistics course" to come to this conclusion with regard to poll data?

Posted by: Ichthyic | January 23, 2010 11:26 PM

PZ is leading that poll by 4% over "Stephen Hawking finds his ALS is eliminated, enjoys health of the body."

*shrug*

I voted for Hawking.

Posted by: A Relevant Haiku | January 23, 2010 11:30 PM

@ 51

On the internet
PZ Myers leads a poll
Water is wet, too

Posted by: Scrawny Kayaker | January 24, 2010 2:27 AM

Thanks for the Joan Jett link. I needed that!

Leading to this was the icing on the cake:
http://www.youtube.com/watch?v=HONFeMNXW0U&feature=related

Posted by: Cactus Wren | January 24, 2010 5:30 AM

Legion @21:

The second panel in the comic, Medieval News Network, reminds us of a program we used to watch on PBS that presented historical events, like the story of Joan of Arc, in the format of a modern newscast.

"Newscasts From the Past". An entertaining series, even if it did repeat the long-discredited notion that medieval people used spices to disguise the taste of rotten meat and the "ring around the rosy" urban legend. One scene that has stood out in my memory had the announcer describing a wonderful technological innovation developed by builders at work on _____ Cathedral: "This new device, we are told, enables one man to do the work of two. They call it a 'wheelbarrow'."

Posted by: Sili, The Unknown Virgin | January 24, 2010 7:52 AM

"Which miracle do you think Americans would most support?"

I had to vote for the staring into the Sun one. This is the general, Faux News watching public we're talking about, after all.

Posted by: blf | January 24, 2010 8:26 AM

This episode is full of win: Newscast from the Past-1642

That was good. The obvious quibble is not all those events happened on that day (16-Oct-1642), with perhaps the most glaring examples being Galileo Galilei was already dead, having died in January of that year, and so could not have been interviewed the previous day; and William Shakespeare died years earlier, in 1616, not on that day. I assume some of the other events have also slightly moved in time?

But that's all a quibble, it was certainly entertaining, full of win, and a very cleaver idea—and I assume a good teacher could use the time-shifting and other stuff to teach a number of lessons.

Posted by: shreddakj | January 24, 2010 9:20 AM

News media are very good at making a story out of nothing. That's what we get for having 24 hour news channels though right?

Posted by: miketv | January 24, 2010 10:35 AM

Major News outlets respond: "Whats a margin of error?"

Posted by: Edmund Berven | January 24, 2010 11:36 AM

To: A Relevant Haiku...and totally off the subject.

This does not follow the normal structure of 17 moras.

After all it's not a haiku, but a Hai-poo!

"Dad kept the bible in the toilet
I started to read it one day
It was begat,begat and begat as I shat."

I believe that's 29 moras...poetic license.

Posted by: deriamis | January 24, 2010 11:56 AM

craig.mcgillivary #48:

If you don't predict yes you are making the same mistake as this cartoon. Given the available evidence yes is more likely to be the majority answer in the other 999 polls. In other words poll results within the margin of error do provide information.

The fact that the results of a poll have predictability isn't relevant to the conclusions that are drawn from the results. It's just an interesting fact. Predictability is a function of probability, which is not the desired measurement of such polls. And even if you were interested in such a thing (because you are conducting scientific surveys and not non-scientific polls, you would still need a large enough difference between the data and the noise to be able to make a claim that the data are predictable.

Besides which, how many polls have you ever seen presented on the boob tube that ever told you what questions were asked, how many people were asked which questions, the order in which the questions were asked, how the data were collected, etc., etc, etc.? That's even more important than the numbers themselves!

Posted by: Rey Fox | January 24, 2010 1:07 PM

"PZ is leading that poll by 4% over "Stephen Hawking finds his ALS is eliminated, enjoys health of the body.""

Nice to know that they care more about someone's religious tribe affiliation than someone's physical health.

Posted by: Carlie, ghoul of deluded buffoons | January 24, 2010 2:33 PM

"PZ is leading that poll by 4% over "Stephen Hawking finds his ALS is eliminated, enjoys health of the body.""

Nice to know that they care more about someone's religious tribe affiliation than someone's physical health.

Of course, they also know it's hella more likely to change someone's mind than to have a miracle that leaves actual physical evidence.

Posted by: John Morales | January 25, 2010 4:10 AM

Ruttee, you're not an atheist, then?

Posted by: Rorschach | January 25, 2010 5:00 AM

Ruttee, you're not an atheist, then?

Irrelevant to judgment of her arguments.

And I hate Atheists too. Jesus and Mo is Mallard Fillmore repackaged for a new audience.

An ad hominem argument.Weak.
More a rant then an actual argument.

Posted by: Rorschach | January 25, 2010 5:18 AM

LOL, sorry John, I'm evidently in "slow joke pickup mode"...:-)

Posted by: sqlrob | January 25, 2010 9:36 AM

Rutee, terrible art, unlike the high quality art styles of xkcd, Dilbert, and Pearls Before Swine?

Posted by: KOPD, SAOTI | January 25, 2010 11:06 AM

computers don't have whims

True, but sometimes their behavior does approximate whims.

Posted by: John Morales | January 25, 2010 4:59 PM

Rutee,

What is the redeeming value of J and M supposed to be, exactly?

For me, it's the satirical yet pithy and wry commentary on religiosity.