3 ms·
> Note that the senior author Bhattacharya and of course Ionnadis were advancing this theory before data collection began. Declaring a hypothesis publicly and
by chrisseaton 7y ago
> Note that the senior author Bhattacharya and of course Ionnadis were advancing this theory before data collection began.
Declaring a hypothesis publicly and then gathering data to see if it’s supported or not is good science, not bad.
- karlshea 7y agoIs it though? Why declare publicly at all then? You’d just be giving the impression that when you do collect the data you’d be cherry picking, exactly like what’s happening here.
- chrisseaton 7y ago> Is it though? Yes. It’s widely accepted as best practice by mainstream scientists. > Why declare publicly at all then? So people know what you were testing and so you can’t fudge it to a test more convenient for your career, politics, or bank balance.
- wahern 7y ago> It’s widely accepted as best practice by mainstream scientists. Simply declaring your hypothesis is not best practice. Publishing the details of your experiment--the protocols, analytical methods, etc--are what's best practice, AFAIU. Few scientists run an experiment they expect to fail, and without publishing the details ahead of time it's trivial to fudge an outcome, so simply pronouncing your hypothesis is entirely uninteresting.
- parineum 7y agoThis is one of the sources of the replication crisis. People would do experiments to test a hypothesis but find nothing conclusive, they would then p-hack their data until they found something worth publishing back-hypothesize. The solution is to publish hypothesis and test methodology ahead of time and publish your data afterward.
- chrisseaton 7y agoYes it doesn't work unless you include your methodology in what you declare upfront.
- MengerSponge 7y agohttps://en.wikipedia.org/wiki/Scientific_method https://en.wikipedia.org/wiki/Scientific_method
- chrisseaton 7y agoYes... that link supports what I’m saying. > The process of the scientific method involves making conjectures (hypotheses), deriving predictions from them as logical consequences, and then carrying out experiments or empirical observations based on those predictions.
- buboard 7y agothe reverse order would be bad science
- chrisseaton 7y agoI don’t know what to tell you except that what you’re saying isn’t what most reputable scientists think.
- buboard 7y agodeciding your hypothesis after collecting data ? edit: i think we agree, misunderstanding
- mbreese 7y ago> Declaring a hypothesis publicly and then gathering data to see if it’s supported or not is good science Not for something like this. This study isn’t about running an experiment to see if treatment A or B is better. This is an observational study that is supposed to determine the prevalence of SARS-CoV-2 infection in the population. There is no hypothesis here - That is to say, there shouldn’t be anything to prove. Whatever answer you measure, you measure (ignoring any other data collection issues). But if you declare “I think the rate of infection is X”, and you end up with a rate of X, then no one knows if you’ve really found the rate to be X or if you put your thumb on the scales. This should have been a fact finding mission to establish some of the basic numbers that will be needed to design the next wave of experiments or to guide policy.
- aravindgp 7y agoVery beautiful said. I always belive that bad scientist will say with certainty that "this is the reason". Good scientists will always say it's probable ,they usually don't give certainity. Policy makers want certainity which is why there's always a tiff between good scientists n policymakers. My principle- When we are not sure of data usually simple models beat complex models.