A very small magnifying glass, for helping people with small hands to see things better
17.06.2025 03:19 β π 1 π 0 π¬ 0 π 0@peter-hayes.bsky.social
PhD researcher working at an intersection between quant psych methods/longitudinal modelling/metascience. - researcher practices - personality/assessment/individual differences - child and adolescent development - Obsidian evangelist
A very small magnifying glass, for helping people with small hands to see things better
17.06.2025 03:19 β π 1 π 0 π¬ 0 π 0have you been baking Ben?
03.06.2025 05:11 β π 0 π 0 π¬ 1 π 0it sounds like you want a plan for a '1-page essay', of indeterminate size/length
02.04.2025 09:15 β π 2 π 0 π¬ 1 π 0Thanks Peter, I'll have to read the chapter and see where you land on how Australians think about the normative aspects of income and wealth distribution. looking forward to it.
06.03.2025 00:19 β π 0 π 0 π¬ 0 π 0Thanks Peter, do you think Australia's 'fair go' narrative is fundamentally changing, or are (might) these (be) temporary disruptions?
06.03.2025 00:10 β π 0 π 0 π¬ 2 π 0omg, this is so funny. UCD is my alma mater and this has brought me joy. thank you
05.03.2025 23:28 β π 0 π 0 π¬ 1 π 0i like to randomly change between camelCase and snake_case and i'd prefer you not to question me, thank you very much
04.03.2025 09:52 β π 2 π 0 π¬ 0 π 0great: resources provided ahead of time enabling people to be properly set up. amazingly detailed resource pack provided so that I could refer to it after the fact.
also great: thoughtful use of discord as the conversation channel alongside google codelab
terrible: reading code off pdfs
Model cough Context cough Protocol
04.03.2025 01:57 β π 1 π 0 π¬ 1 π 0it's almost like the people haven't read "Causal and Associational Language in Observational Health Research: A systematic Evaluation"
02.03.2025 22:07 β π 2 π 0 π¬ 0 π 0i look forward to people bemoaning the state of this new meta-meta-science and wishing for a return to the simpler days of meta-science.
28.02.2025 06:23 β π 1 π 0 π¬ 0 π 0i'm something of a many-many-analyst-analyst myself
28.02.2025 05:26 β π 4 π 0 π¬ 1 π 0how many many-analyst studies do we need before we can do meta-science of many-analyst studies?
quick back of the envelope calculations only please
this is an interesting argument and one that has implications for the 'just do better science' argument
i think we should think about the many different ways that disciplinary norms shape and constrain science and scientific progress
- could be theory, methods, reporting, data wrangling
anyway, i better get back to my actual work -
28.02.2025 02:19 β π 0 π 0 π¬ 1 π 0it also seems that you can conceive of a scientifically useful contribution from this kind of study
this study provides an empirical test of something we can otherwise only speculate about.
.
I'm not sure how you would even begin to deal with the issue of selection bias by handing this type of study to a select set of subject domain experts.
it seems like there is a cake and eating it problem
i would counter that it may be useful to develop an understanding of whether there are systematic researcher biases that may cross-cut substantive topics/domains of research and to have information about which researcher decisions have greatest impact on model estimates, etc
28.02.2025 01:59 β π 0 π 0 π¬ 1 π 0and I guess that's your fundamental argument - that understanding how researcher decisions can contribute to variability in findings doesn't appear to be useful to the scientific endeavour. 1/
28.02.2025 01:57 β π 1 π 0 π¬ 1 π 0fine, but that doesn't preclude the usefulness of understanding factors that may influence variability in modelling
28.02.2025 01:44 β π 0 π 0 π¬ 1 π 0Even aligning theory to a statistical model is not trivial and may involve variability due to subjective/researcher decisions.
Being able to model those sources of variability is a useful scientific contribution
It's striking that you speak about the blue-tit claim, but not the grassland claim
What can you say about contexts in which there is no univocal theoretical framework? What can you say when there may be multiple causal mechanisms? /1
thanks, I guess I was getting excited about that potential paper. "psychology without its eugenicist foundations."
20.02.2025 00:57 β π 3 π 0 π¬ 0 π 0Yeah, so given this looks like an introductory paragraph, I'd be interested in whether the following argument is about how a statistics not based on eugenics might have contributed to a different empirical psychology.
Do you have a view?
Is it really secondary to what made them foundational to the field?
Briggs' article on Galton in Historical and Conceptual Foundations supports the view that his commitment to eugenics were pretty essential/core to the statistical methods that he developed
that is wonderful, but I imagine a fair proportion of students don't receive such thoughtful feedback on their writing.
19.02.2025 21:37 β π 1 π 0 π¬ 1 π 0I wonder if this Toulmin Argument Model Ontology could be a useful approach for developing a json_schema for your project?
sparontologies.github.io/amo/current/....
OMG outlook search is absolute trash - trying to find an email that I know I have sent - never ending cycle of torment
06.02.2025 03:02 β π 0 π 0 π¬ 0 π 0your important question is not answered in this thread: www.reddit.com/r/asklinguis...
but other, related, important questions are!
this is pretty amazing stuff!
it seems like there are gains to be made with algorithmic efficiency and that it's not all brute compute!