Suppose I take old data analyzed with a frequentist t-test, reanalyze it using a bayesian test, and obtain a different conclusion. Would you say this also falls under the general term replication? Or is replication strictly limited to new datasets and not new methods?
-
-
Replying to @SachaEpskamp
I think people use the terms differently, but a terminology I like: Replicability=(same analysis, new data), Reproducibility=(same analysis, same data), Robustness=(different analysis, same data), Generalizability=(diff operationalization & analysis, new data).>
12 replies 108 retweets 395 likes -
Replying to @dingding_peng @SachaEpskamp
Remi Gau Retweeted Berna D.
I would like to point out to this paper for more food for thoughts for definitions. H/T
@zerdevehttps://twitter.com/zerdeve/status/1062733416678846465?s=19 …Remi Gau added,
Berna D. @zerdeveOur new preprint is out: “A Model-Centric Analysis of Openness, Replication, and Reproducibility” We aim to provide some precision and clarity about how open science is related to replicatind experiments and reproducing their results. 1/4 https://arxiv.org/abs/1811.04525Show this thread1 reply 2 retweets 7 likes -
Federico Adolfi Retweeted Federico Adolfi
Federico Adolfi added,
Federico Adolfi @fedeadolfiReplying to @lakens @dingding_peng and 2 othersSome researchers have a (more nuanced, perhaps?) view which could sound confusing given the above definitions if the full context is not available, I think. Take a look at this example - I suppose this would sound backwards to more than a few. pic.twitter.com/QxjKE4FTGL1 reply 0 retweets 2 likes -
Replying to @fedeadolfi @RemiGau and
Ha! That excerpt sounds awfully familiar
(Taken out of context all of these terms would be confusing to some. That’s why we took care to define them precisely for our purposes. Not necessarily as a suggestion for broader use.)1 reply 0 retweets 6 likes -
Replying to @zerdeve @fedeadolfi and
I guess the general advice is "at least give some definition of what you mean by those terms in the intro of the paper / blog / twitter thread and try to be internally consistent for the time of that piece of work".
1 reply 0 retweets 6 likes -
Yeah, I do the same too every time I talk about all the different types of reproducibility (and related terms). In one article we had six different versions going just for computational rep: http://oliviaguest.com/doc/guest_rougier_2016.pdf …
1 reply 1 retweet 3 likes -
Reading this atm and they have tables with yet another list of definitions for replicate, reproduce, repeat, rerun, reuse... Trying to get everyone to agree on those would make herding cats an easy task.pic.twitter.com/sOtrdWK5FI
1 reply 1 retweet 2 likes
This idea of a "true" prescriptive definition (not within an article or person) is kind of pointless IMHO (because of the nature of the terms/conceptS), but comes up on Twitter every so often.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.