/sci/ - Science & Math

Koko has been observed to, unprompted, "narrate" her play. (e.g. "chase tickle" when playing with dolls)
Alex on at least one occasion combined two words (banana and cherry) to create a new word (banerry) describing an object he didn't have a word for (apple) that had some characteristics of both.

of course, Alex was also known to bite, say "sorry" and then bite again. so not a complete understanding of the word's meaning so much as recognizing the convention of saying it to someone you've hurt.

>>8670433
>blog post
>paper about failure of potential drugs
>essay with no primary data (again, restricted to biomedical science)
>journal comment listing red flags that identify work as difficult to reproduce
>survey (limited to oncology) that didn't address rates of reproducibility, but rather what fraction of researchers had ever had difficulty reproducing someone else's results
>paper talking about how poor documentation of materials makes it difficult to reproduce results (again, limited to biomedical research; noticing a pattern?)
>news article talking about issues of reproducibility
>news write-up on Reproducibility Project, which found low rates of reproducibility...but failed to account for the fidelity of the reproduction to the original http://www.pnas.org/content/113/23/6454.full.pdf
>paper showing that retractions are usually the result of misconduct rather than simple error.
>another write-up of Reproducibility Project, see above

a few things stand out:
first off, you don't seem to understand what reproducibility is and isn't. you've included papers warning that methodological vagueness can make it hard to accurately reproduce an experiment (which is a separate issue from reproducing its results), and talking about retractions (which are ENTIRELY different).
secondly, there's an issue with biomedical research that is particular to the field. since drug trials are done on living beings, there are a lot of hard-to-control variables, and ethics considerations severely limit your methodology. for these reasons and a few others related to identifiability of materials (antibodies differ from batch to batch, model organisms differ from strain to strain) reproducibility is lower in biomed than in other fields.

>>8646695
>Not every scientist thinks trump is antiscience.
Nature literally put out a request for letters to the editor from pro-Trump scientists to try and find out what their reasoning is, because so few scientists support Trump.

>>8646756
just gotta pray hard enough desu senpai

>>8646875
>Scientists have failed over 25 years to make a single correct prediction.
zozzle
https://www.youtube.com/watch?v=tPSIvu0gQ90