Thread replies: 8
Thread images: 2
Thread images: 2
Anonymous
How do I know if a study is trustworthy? 2017-08-14 08:26:48 Post No. 9105294
[Report] Image search: [Google]
How do I know if a study is trustworthy? 2017-08-14 08:26:48 Post No. 9105294
[Report] Image search: [Google]
I've been having a lot of debates about all sorts of things recently - politics, psychology, anthropology, economics and so on - and by the time I bring up statistics and studies, the other party argues that all studies can be subject to fraud and manipulation, and at that point all I can do is fight anecdotes with anecdotes.
First of all I'd like to know how to rebut this claim, but even if I can get people to give up this line of thinking, the fact that a study could have been tampered with remains. What safety measures exist to ensure the authenticity of a study and how can I make sure a study is legitimate - that is, if the numbers haven't been adjusted or important details not been omitted?
>>
>>9105294
If it replicate can be trusted. Just like studies about Einstein Theory of Relativity.
Psychology by other hand rarely replicate, so cannot be trusted.
>
http://www.slate.com/articles/health_and_science/science/2016/10/why_the_replication_crisis_seems_worse_in_psychology.html
>
Replication = redo the experiment with same results by different research groups.
>>
>>9105294
make sure it comes form a reputable journal and I guess make sure it's a couple years old so there's been some time for other papers to reanalyze the claims, and check for said papers either corroborating or disagreeing with the original paper
>>
>>9105313
Couldn't someone argue that the replicating group matched their numbers to artificially confirm the results? I know this is going into conspiracy theory, but I'd like to cover my bases.
>>9105320
I think I'm gonna have to start documenting my go-to studies and their authors and keep them handy on my phone. "I read in a study from a university" isn't cutting it anymore.
>>
>>9105294
Cite more than one study, read up on the methods
>>
>>9105327
Yes someone can argue but if there is weak evidence than who cares. Science isn't about proving anything, it's about gathering evidence to support a conclusion. If someone wants to be a fag and rebuttal that data is manipulated and/or methods aren't reliable than let them, nothing you say will convince them and they most likely are only concerned with winning an argument. Just have strong evidence and tell them your argument is stronger
>>
How frequent is manipulation of data actually? I often get anecdotes about professors asking their students to adjust their methods to reach a preconceived conclusion, but I imagine it's fairly rare and gets resolved sooner than later.
Then again, wasn't the study that lead to the food pyramid paid by the sugar lobby and that remained undiscovered for decades? I might be remembering it wrong, so I could be off.
>>
>>9105294
Besides the methodologies used, any claims made in the discussion or conclusion sections can be pretty hyperbolic. It's important to not jump to conclusions especially when results seem like they could be a game changer. Don't take just one research team's word for it. There will always be related studies that when taken together provide a better representation of the subject.
Thread posts: 8
Thread images: 2
Thread images: 2