Total Posts:4|Showing Posts:1-4
Jump to topic:

HFCS vs. sucrose

Cliff.Stamp
Posts: 2,169
Add as Friend
Challenge to a Debate
Send a Message
6/19/2011 12:03:05 AM
Posted: 5 years ago
C0nc0rdance is a scientist who does a decent amount of YouTube video's. He has a number of really nice series and has started a new one where he does paper reviews :

The nice thing here is that he talks about how to interpret a paper beyond the quote-mining that is popular in internet snipes.
tvellalott
Posts: 10,864
Add as Friend
Challenge to a Debate
Send a Message
6/20/2011 6:49:46 PM
Posted: 5 years ago
At 6/19/2011 12:03:05 AM, Cliff.Stamp wrote:
C0nc0rdance is a scientist who does a decent amount of YouTube video's. He has a number of really nice series and has started a new one where he does paper reviews :



The nice thing here is that he talks about how to interpret a paper beyond the quote-mining that is popular in internet snipes.

C0nc0rdance is one of my favourite YouTube channels. I'll definitely have a look at this when I get home.
"Caitlyn Jenner is an incredibly brave and stunningly beautiful woman."

Muh threads
Using mafia tactics in real-life: http://www.debate.org...
6 years of DDO: http://www.debate.org...
RoyLatham
Posts: 4,488
Add as Friend
Challenge to a Debate
Send a Message
7/1/2011 1:46:06 PM
Posted: 5 years ago
I like the video because it encourages looking at length at how a paper gets to its conclusions. Still, I think the speakers attention to funding sources is inappropriate. He seems t be saying that one should always look for grounds for an ad hom attack, just in case. Scientific papers stand or fall on their results, not who funded them.

A sample size of ten is too small to draw a strong conclusion, even if the results are statistically significant. With four groups tested, there is about one chance in five that a statistically significant result would appear by chance in at least one of the groups. I didn't get a clear picture of how many groups they had, but it was at least four.

This is the same problem as getting a battery of twenty blood tests. With a 5% false-positive rate, one of them is expected to be a false positive.

Still, the guy who made the video did a good job of raising awareness of problems in scientific papers. The weakness is often in the statiistical analysis, and the authors give themselves lots of room to interpret the results.
Cliff.Stamp
Posts: 2,169
Add as Friend
Challenge to a Debate
Send a Message
7/1/2011 9:34:45 PM
Posted: 5 years ago
At 7/1/2011 1:46:06 PM, RoyLatham wrote:

Still, I think the speakers attention to funding sources is inappropriate. He seems t be saying that one should always look for grounds for an ad hom attack, just in case. Scientific papers stand or fall on their results, not who funded them.

That is a very naive approach Roy, if you are funded and you keep publishing reports against the funding then you stop getting funded. Then there are the other matters such as directive focusing, and selective vectoring, not to mention of course the always present data manipulation. He was just noting that he would look to funding to be a bias, so would I and so should anyone. This does not mean it is false, science is not about absolutes, it could always be false, it just means in general if people are funded by X then they will tend to publish that which supports X and that you should look for objective perspectives. Scientists are just people and are not some holy paradigms of morality. It is no different than if you wanted to ask someone about my character for a public statement you may not want to ask the people who depend on me for their work because they have a bias to obviously give you a selective interpretation of the truth.