6 ms·
First, not having read the article: #EvidenceBasedPolicy is a worthwhile objective even if only because the alternative is to just blow money without measuring
by westurner 5y ago
First, not having read the article:
#EvidenceBasedPolicy is a worthwhile objective even if only because the alternative is to just blow money without measuring ROI at all [because government expenditures are the actual key to feeding the beast, the economic beast, the...].
What are some examples of policy failures where Systematic review and Meta-analysis could have averted loss, harms, waste, catastrophe, long-term costs? Is that cherry picking? The other times we can just throw a dart and that's better than, ahem, these idiots we afford trying to do science?
Wouldn't it be fair to require that constituent ScholarlyArticles (and other CreativeWorks) be kept on file with e.g. the Library of Congress?
Non-federal governments usually have very similar IT and science policy review needs. Should adapting one system for non-federal governments be more complex than specifying a different String or URL in the token_name field in a transaction?
When experts review ScholarlyArticles on our behalf, they should share their structured and unstructured annotations in such a way that their cryptographically signed reviews - and highlights to identify and extract structured facts like summary statistics like sample size and IRB-reviewed study controls - become part of a team-focused collaborative systematic meta-analysis that is kept on file and regularly reviewed in regards to e.g. retractions, typical cognitive biases, failures in experimental design and implementation, and general insufficiencies that should cause us to re-evaluate our beliefs given all available information which meets our established inclusion criteria.
We have a process for peer review of PDFs - and hopefully datasets with locality for reproducibility and unitarity which purportedly helps us work through something like this sequence:
Data / Information / Knowledge / Experience / Wisdom
We often have gaps in our processes to support such progress in developing wisdom from knowledge that should be predicated upon sound information and data and then experience, bias, creeps in.
Basic principles restricting the powers of the government should prevent the government - us, we - from specifically violating the protected rights of persons; but we have allowed "Science" to cloud our judgement in application of our most basic principles of justice - i.e. Life, Liberty, and the pursuit of Happiness; and Equality and Equitability - and should we chalk the unintended consequences up to ignorance or malice?
More science all around: more Data Literacy - awareness of how many bad statistical claims are made all day around the world everywhere - is good and necessary and essential to Media Literacy, which is how we would be forming our opinions if we didn't have better tools for truth and belief for science.
"What does it mean to know?" etc.
Logic, Inference, Reasoning and Statistics probably predicated upon classical statistical mechanics are supposed to bring us closer to knowing: to bring our beliefs closer to the most widely observed truths.
Which Verifiable Claims do we trust? What studies do we admit into our personal and community meta-analyses according to our shared inclusion criteria?
"Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)" is one standard for meta-analyses, for example. http://www.prisma-statement.org/ http://www.prisma-statement.org/ . Could the bad guys or the dumb good guys lie with that control in place, too? Can knowing our rights - and upholding oaths to uphold values - protect us from meta-analytical group failure?
Perhaps STEM (Science, Technology, Engineering, art, and Medicine/Math) majors and other interested parties can help develop solutions for #EvidenceBasedPolicy?
This one fell flat. Maybe it was the time of day? The question should be asked every year, at least, eh?
"Ask HN: Systems for supporting Evidence-Based Policy?"
https://news.ycombinator.com/item?id=22920613 https://news.ycombinator.com/item?id=22920613
>> What tools and services would you recommend for evidence-based policy tasks like meta-analysis, solution criteria development, and planned evaluations according to the given criteria?
>> Are they open source? Do they work with linked open data?
> I suppose I should clarify that citizens, consumers, voters, and journalists are not acceptable answers
"#LinkedMetaAnalyses", "#StructuredPremises"; Ctrl-F "linkedmeta", "linkedrep", "#LinkedResearch": https://westurner.github.io/hnlog/ https://westurner.github.io/hnlog/
Alright, my fair biases disclosed, on to the reading the actual article: /1