@marcuscrede.bsky.social
Grumpy IO psychologist with an interest in research methods, meta-science and personality. Views are my own and not those of my employer. Begrawe my hart op Klein Tambotieboom en strooi my as oor die Bosveld horison
Has he ever pet a dog; smelled a flower; walked hand-in-hand with a loved one; did something for someone else without expecting something in return; shed a tear for someone else etc.
They would probably grow in your neck of the woods. We have one in our garden and it has started producing nice fruit (needs a cross pollinator).
We have an orchard near us that has a few dozen trees but they go fast.
Crimson crisp is obviously no 1.
Brilliant little game: get a google street view image and guess how that county voted in the last presidential election.
I scored 740/1000.
www.ballotguessr.com/play
I can't stop thinking about this. Imagine:
- 2200 studies, as they say >2000
- 2 studies per article given piloting/multi-study articles
- No p-hacking (lol); 1 test per study (lol), 5% False Positive Rate
That's 55 significant studies that are false positives
Actual publications: 64
Great. How about a reward for those people who've spent thousands of hours trying to fix the horrendous bullshit produced by our fields' "thought leaders".
Record your talk, hand out ritalin to your audience, and play it back at 3x speed.
I'd also add two additional CFA related issues: 1) chi-square comparisons between nested models are only appropriate when the less parsimonious model has a non-significant chi-square, 2) when model fit is poor (as you have on page 21) parameter estimates from the model should not be interpreted.
Interesting, but the comparison between an orthogonal and single-factor model seems wrong. Surely, the most plausible alternative is a correlated factor model or a hierarchical model? Orthogonal models like this always fit badly.
Any decade now I will have a conversation with another researcher about my work. It has to happen at some point, right?
Oh, to publish in a Nature brand journal about how *checks notes* the big math changes to small math.
This is so cool because even a complete R coding idiot like myself can follow along. Thank you so much!!
Yup! I have an early draft of their original paper and it is nuts to read that. Carney sent it to me when I contacted her to inquire about the non-significant chi-square value that they reported as significant. That's when the whole power posing crisis started. Somehow I was framed as the bad guy.
Important to remember that neither the original power posing researchers nor almost all of the replicators could be bothered to include a neutral pose control condition and that any effective on feelings was almost entirely due to the -ve effects of a contractive pose.
Ha! The original Lancet article on the dangers of reading in bed is here: doi.org/10.1016/S014...
If you contact an APA journal to note that multiple papers by the same authors claim that 2+2=7 and that all of their inferences are based on this claim what will follow is a year-long "investigation" followed by a correction noting that 2+2=4 but that the inferences are unchanged.
Science!
I just received an e-mail suggesting that I follow Snoop Dog on LinkedIn. Does this mean that I am cool?
True but I also would not have wanted to witness Trump and the Ayatollah resolving conflict the way that Bonobos do.
Definitely. I am very confident that other journals would publish this. I was naive enough to think that the journal that published the meta-analysis would see that they should publish what is essentially a major correction to work that they published.
Good point! We've appealed already but asking for a retraction might not be a good follow-up.
Today the APA journal that published that meta-analysis rejected our commentary - primarily because our findings were not interesting enough. What the actual fuck!
We recently submitted a commentary on a very influential meta-analysis. We found that: 1) 40% of relevant literature had not been identified because of lazy search, 2) a few large N included studies did not meet stated inclusion criteria, and 3) that almost all sig. moderator findings were wrong.
Users of survey data, lovers of DAGs, and general methodological enthusiasts, gather round!
I'm so excited to share this new paper, joint work with my brilliant colleagues @rjsilverwood.bsky.social, @pwgtennant.bsky.social, and Liam Wright.
🧵
Is there no informed consent form for this? This feels very dodgy. Is there IRB approval for this study?
Pfft. Mathematically impossible results are not a problem at all. Simply send them to the Journal of Management. They give awards for that kind of stuff.
i paid around $2.5 a few days ago.
Plan to take the next day off from work if you are going to get the pneumonia vaccine. Ask me how I know.
Amazing!!