Statistics and Manipulating Public Perception

[HippoTwilight]

This seems extremely fitting to the news these days. Pundits from every side use numbers and statistics to prove themselves right, or their opponents wrong.
These videos provide a glimmer of insight into what happens behind the scenes. To put it bluntly, the general public doesn't have a chance to learn the "truth" when you have teams of people with advanced math and psychology degrees working to manipulate data to fit the desired narrative. You could use the exact same data, present it in two different ways, and make two "informed" people argue and fact-check to the death that the other is wrong.

How to make statistics lie:


Manipulating perception:

 

[Bacpacker]

Same with polling numbers. Ask questions a certain way, or in a certain area, skews the info badly.

 

[dademoss]

There are three kinds of lies: lies, damned lies, and statistics. Mark Twain.

 

[backlash]

Our China virus number made a huge jump yesterday. Seems like the imported data from several weeks ago. The numbers are all suspect.

 

[bkt]

Thanks, Hippo, those are great videos and they bring me back to my college years of studying sadistics (yes, that's what we called it).

Confirmation bias is, I think, a huge player these days. People live in an echo chamber. We listen to news we trust and deem other sources of being biased. We look for confirmation of our beliefs without analyzing the actual data.

That we have had so much contradicting dialogue and that there has been so much censorship, we are left either holding onto what we think is true, or losing all faith in all sources of information and acknowledging we CAN'T know what's true unless and until we dig into the data ourselves. The latter is a huge undertaking virtually no one is capable of doing.

So when we see news reports like "96% Of All People Tested Are Positive, Most Asymptomatic", how many of us actually get on the phone to call the facilities that provided this information to the news source? I posted an article a couple days ago where Florida was found to be in error with regard to positive C-19 test cases. Rather than 96%, the number was 9.6%. That's a big fargin' difference!! So was the error deliberate or accidental?

Evil or stupid?

That seems to be the recurring question of our times.

Thanks again, Hippo. These videos help show how data can be (and are) skewed to mean whatever the person commissioning a study wants them to mean.

 

[dademoss]

Sadly, the data analysis doesn't scare me off, the integrity of the data does

 

[HippoTwilight]

Our China virus number made a huge jump yesterday. Seems like the imported data from several weeks ago. The numbers are all suspect.

It's bad enough when solid data is manipulated. In the case of COVID, we have officials openly admitting to fudging the data.
Any death with a positive COVID test are counted as a COVID test. Regardless of the actual cause. That's inflating the death rate.
Then there's "clinical" COVID deaths. Where people who showed COVID symptoms (I think the list covers about everything now) are reported as probable COVID deaths. This probably causes an even bigger inflation in the death rate.

Aside from COVID, this stuff gets used for absolutely everything. Statistical algorithms run the world. Facebook and Google use it to send you targeted ads. Insurance companies use it determine rates. Banks use it to determine loan approvals. Wall Street has been running historical data through algorithms trying to quantify risk.

 

[Meerkat]

There are three kinds of lies: lies, damned lies, and statistics. Mark Twain.

LOL, liked Twains books and philsophy." how you know a lawyer or plotician is lying,there mouth is opem".I think Twain said something like that?
I'm more on the G.Carlon, " its a club and you ain't in it" [ minus 200 F words]

 

[dademoss]

(I think the list covers about everything now)

I thought I had the Kung Flu, but a couple Benedryl worked wonders.