Magnus Johansson
banner
pgmj.bsky.social
Magnus Johansson
@pgmj.bsky.social
PhD & lic. psychologist. Research specialist at Karolinska Institutet. R package for Rasch psychometrics: pgmj.github.io/easyRasch/
#openscience, #prevention, #psychometrics, #rstats, #photo
I've done some more work on the relative measurement uncertainty, comparing `brms` posterior draws to "plausible values" in Rasch models, and some other reliability metrics. Estimating RMU from draws adds some variation, as shown in the figure.
pgmj.github.io/reliability....
#rstats #psychometrics
October 24, 2025 at 4:49 PM
The RelRep.R function "just works" with a dataframe with items as columns and produces pretty neat output for alpha or omega. Point estimate and CI for alpha is almost identical to RMU for my example with `eRm::raschdat1[,1:20]` data. github.com/melissagwolf...
October 10, 2025 at 9:20 PM
For instance, yesterday I read a paper with a table describing participants' sickness absence days with a mean of 71 and SD = 88. Generating a random (gaussian) sample using these values produces ~20% participants with less than zero sick days.
September 22, 2025 at 1:24 PM
Grateful to have this view a few minutes walk from home.
August 15, 2025 at 10:35 PM
Inspired by a paper (linked) that added a response category to the PHQ-2 and GAD-2 screener questionnaires we did the same in Swedish with good results in a sample of 15-18 year olds. The uppermost categories performed worse, though (q25-26=GAD,27-28=PHQ). #psychometrics osf.io/preprints/ps...
August 7, 2025 at 2:32 PM
Reading the preface to "Thinking through statistics" and this section already made me like the book.
August 4, 2025 at 9:25 PM
Might be the perfect guitar string gauge? Seems ridiculous, but .005 inches thinner than the .011 I've used forever on my solid body seems really nice.
July 29, 2025 at 8:44 PM
July 19, 2025 at 7:48 PM
...since it needs revising (ongoing) to incorporate correct information about critical values for GOF metrics: osf.io/preprints/os... We basically use the 4 criteria as presented by Kreiner (2007), who refers to Rosenbaum (1989).
May 22, 2025 at 2:02 PM
Really looking forward to this. It's been six years since his last book.
April 29, 2025 at 8:43 PM
Great to see conditional reliability and the "targeting" of test/sample discussed. I've incorporated a similar approach in a TIF curve figure in the `easyRasch` R package. Note that IRT/Rasch also provides information about the reliability of the test/questionnaire itself, independent of the sample.
March 12, 2025 at 4:25 PM
Such style and grace. Have one of those too
January 15, 2025 at 9:16 PM
Made with a Makina 67 a few years ago. Rollei Retro 400S overexposed a step and developed in Adox Silvermax 1+19.
#filmphotography #blackandwhite #believeinfilm
January 11, 2025 at 7:52 AM
Sunset at 13.30 has some benefits.
December 28, 2024 at 12:46 PM
Stunning clouds, which my old phone struggles to reproduce.
December 25, 2024 at 1:29 PM
November 16, 2024 at 8:32 AM
October 26, 2024 at 4:24 PM
October 7, 2024 at 9:21 PM
Colors and sunlight this time of year is just so nice.
September 21, 2024 at 8:28 PM
The TAM author does not seem very enthusiastic about that setting though
September 4, 2024 at 8:46 PM
I also added multicore processing and introduced a bug. Here are proper results, table and figure.
September 4, 2024 at 10:57 AM
I added `pairwise` and reran the same 250 datasets (9 items, 4 thresholds each, using PCM with 720 well-targeted simulated responses for each iteration). PAIR seems to perform worst. Summary table of absolute deviations from input parameters.
September 4, 2024 at 8:56 AM
I messed up earlier, assuming the output of tam.threshold() to be comparable to similar functions in other packages. But useful item thresholds were found in the tam() output object and I could rerun the simulation. Less bad for TAM, but still worse than eRm and mirt. #rasch #rstats #psychometrics
September 3, 2024 at 7:18 PM
August 16, 2024 at 12:07 PM
Berg ur arkivet. Utsikt från Ryfjället.
July 21, 2024 at 3:14 PM