jfeltphd.bsky.social
Quantitative methodologist interested in Bayesian and frequentist approaches to longitudinal data analysis, causal inference, and measurement. Assistant Research Professor in the Center for Healthy Aging at Penn State. UC Merced Quant Psych Alum.
62 posts
866 followers
440 following
Regular Contributor
Active Commenter
comment in response to
post
If MCAR, I’ll just randomly impute 1s and 0s so I don’t lose power
comment in response to
post
Oh absolutely. Unfortunately it means you’ll have to double check these errors in lme4 before determining if they are erroneous though
comment in response to
post
Another thing to request, particularly when using lme4 or nlme in R, is the confidence intervals. If you just do a summary on the model object, you won’t actually know if your model has converged, particularly in the random effects. Asking for confint (in lme4) or intervals(in nlme) will check this
comment in response to
post
As a Sacramento kings fan, I am a fan of anyone that is a thorn in the Lakers side. Go Celtics!
comment in response to
post
Have you been telling participants to “git gud” in your studies?
comment in response to
post
Wow, this is an impressive three to have in one talk!
comment in response to
post
I once heard someone say that “they didn’t need to randomize because the data were all collected so close in time to each other”
comment in response to
post
Definitely this paper. It is written so clearly without a lot (or any equations). It helped me also see that the whole DAGs framework also fits with the threats to validity and the Rubin’s causal framework
comment in response to
post
That is a really clever poster design
comment in response to
post
I agree with all of this. Outliers are really individuals in your sample that come from a different population than you intended to collect, not necessarily extreme values.
comment in response to
post
Will shadish had a great paper on this where participants were randomized to choose their condition or to be randomly assigned to condition www.statpower.net/Content/MLRM...
comment in response to
post
Interesting. I wonder if these findings generalize across methods for establishing childhood adversity? Lots of work on how prospective and objective measures of child maltreatment (to biological measures) are related to different outcomes than retrospective self-report (to psych measures)
comment in response to
post
That’s a great question. I haven’t seen anything published on that. I only have a single anecdote with an adult male around 59 years old. when I was collecting my dissertation data, the father of one of my RAs wanted to try it and it worked despite him knowing the goal and undergrad confederates
comment in response to
post
You can do self-report, housemate/family report, and… voyeur report?
comment in response to
post
Most most may have been too strong of a word for me to use. I’ve never had any issues with articles I’ve published being made available immediately on pmc. The journals have all provided an approved version for articles funded through the nih in my experience.
comment in response to
post
Most journals only have rules about the final formatted version. You can usually post the last version you submitted to the journal before it was formatted
comment in response to
post
MPlus is definitely still used a lot. Some things are just easier/more feasible in MPlus than in R. I’d still prefer R in most cases
comment in response to
post
I may need to use Mark Jones for my word or the day generator
comment in response to
post
I’ve had this problem when I click on a link for a starter pack outside of BlueSky, but not from within the app.
comment in response to
post
We you looking for growth or how to variables change over time together? With ESM analyses I often look at how moments when a given individual is higher or lower than their typical score on a predictor is associated with the outcome. You can control for time and make it random to add growth in mlm
comment in response to
post
comment in response to
post
I’m doing a lot of statistical work in this area now (for mobile cognitive assessment). Could you add me to this starter pack please?
comment in response to
post
Could you add me please?
comment in response to
post
Please add me!
comment in response to
post
Great paper! Thanks for sharing it!
comment in response to
post
There are issues with him on the quant side too. There are a lot of other open science researchers who are more kind, professional, and doing better work. I had to stop following him long ago
comment in response to
post
Could you add me to this?
comment in response to
post
Add me please
comment in response to
post
Can they provide a citation supporting their claim?
comment in response to
post
#devpsych #stats
comment in response to
post
Nice! You can also do this in r using toupper. names(df) <- toupper(names(df))
There’s even a tolower if you want all lowercase
comment in response to
post
Does APS still do their ICPS?