Tony Roberson
@tonyjroberson.bsky.social
Asst Prof UH-Clear Lake / #RStats, Methods, & Measurement / Youth Wellbeing / #SchoolPsych / 🎸Radical🤙Behaviorism / #OpenScience / 🎞 Da Moviesh 🎥 / #DoubleBass
Reposted by Tony Roberson
Intervening on a central node in a network likely does little given that its connected neighbors will "flip it back" immediately. Happy to see this position supported now.
"Change is most likely [..] if it spreads first among relatively poorly connected nodes."
www.nature.com/articles/s41...
"Change is most likely [..] if it spreads first among relatively poorly connected nodes."
www.nature.com/articles/s41...
Transformation starts at the periphery of networks where pushback is less - Scientific Reports
Scientific Reports - Transformation starts at the periphery of networks where pushback is less
www.nature.com
September 29, 2025 at 9:16 AM
Intervening on a central node in a network likely does little given that its connected neighbors will "flip it back" immediately. Happy to see this position supported now.
"Change is most likely [..] if it spreads first among relatively poorly connected nodes."
www.nature.com/articles/s41...
"Change is most likely [..] if it spreads first among relatively poorly connected nodes."
www.nature.com/articles/s41...
I didn't change the default correct answer to the actually correct answer for 2 items on my multiple choice Psych Measurement final when I built it so there were a couple students who ended up doing better than I did on the test lol #AcademicSky #SchoolPsych #ClinicalPsych #Psychometrics
July 3, 2025 at 9:47 PM
I didn't change the default correct answer to the actually correct answer for 2 items on my multiple choice Psych Measurement final when I built it so there were a couple students who ended up doing better than I did on the test lol #AcademicSky #SchoolPsych #ClinicalPsych #Psychometrics
Reposted by Tony Roberson
Google AI is useless. This made me laugh.
June 25, 2025 at 12:13 PM
Google AI is useless. This made me laugh.
Reposted by Tony Roberson
It’s a good thing Congress isn’t alive to see this
June 22, 2025 at 12:47 AM
It’s a good thing Congress isn’t alive to see this
Do you use behavior rating scales that offer a general total score and subscale scores?🚨SPOILER!!!🚨You probably shouldn't be interpreting most of those. Come join and learn how to tell which scores are uniquely interpretable and which are wasting your time! #schoolpsych #clinicalpsych #psychometrics
❗️Our next workshop will be on July 10th, 6 pm CEST, on Assessing rating scale dimensionality through bifactor model-based indices in R @tonyjroberson.bsky.social!
Register or sponsor a student by donating to support Ukraine!
Details: bit.ly/3wBeY4S
Please share!
#AcademicSky #EconSky #RStats
Register or sponsor a student by donating to support Ukraine!
Details: bit.ly/3wBeY4S
Please share!
#AcademicSky #EconSky #RStats
June 21, 2025 at 9:28 PM
Do you use behavior rating scales that offer a general total score and subscale scores?🚨SPOILER!!!🚨You probably shouldn't be interpreting most of those. Come join and learn how to tell which scores are uniquely interpretable and which are wasting your time! #schoolpsych #clinicalpsych #psychometrics
Reposted by Tony Roberson
CLARIFICATION :
The Trevor Project's hotline will still be live. But basically this means LGBTQ people can no longer be connected to it through the US national 988 suicide prevention hotline.
They will need to be contacted DIRECTLY
So spread this around
They can be texted. They can be called.
The Trevor Project's hotline will still be live. But basically this means LGBTQ people can no longer be connected to it through the US national 988 suicide prevention hotline.
They will need to be contacted DIRECTLY
So spread this around
They can be texted. They can be called.
June 18, 2025 at 5:48 PM
CLARIFICATION :
The Trevor Project's hotline will still be live. But basically this means LGBTQ people can no longer be connected to it through the US national 988 suicide prevention hotline.
They will need to be contacted DIRECTLY
So spread this around
They can be texted. They can be called.
The Trevor Project's hotline will still be live. But basically this means LGBTQ people can no longer be connected to it through the US national 988 suicide prevention hotline.
They will need to be contacted DIRECTLY
So spread this around
They can be texted. They can be called.
Reposted by Tony Roberson
There are only 4 useful decisions for peer reviewed journals.
Reject
R & R Major revision (goes back out for full review)
R & R Minor revision (reviewed only by AE and editor)
Accept
Everything else is just confusing and unhelpful.
Reject
R & R Major revision (goes back out for full review)
R & R Minor revision (reviewed only by AE and editor)
Accept
Everything else is just confusing and unhelpful.
May 29, 2025 at 12:32 AM
There are only 4 useful decisions for peer reviewed journals.
Reject
R & R Major revision (goes back out for full review)
R & R Minor revision (reviewed only by AE and editor)
Accept
Everything else is just confusing and unhelpful.
Reject
R & R Major revision (goes back out for full review)
R & R Minor revision (reviewed only by AE and editor)
Accept
Everything else is just confusing and unhelpful.
Reposted by Tony Roberson
It came up, for highly scientific reasons, during yesterday's Data Science Lunch in my department that one can estimate the speed of light using a microwave oven and marshmallows. I've successfully done this experiment in my own home. Here's a nice video explaining. www.youtube.com/watch?v=HwRE...
Finding The Speed Of Light With Peeps | SKUNK BEAR
YouTube video by NPR's Skunk Bear
www.youtube.com
May 28, 2025 at 7:13 AM
It came up, for highly scientific reasons, during yesterday's Data Science Lunch in my department that one can estimate the speed of light using a microwave oven and marshmallows. I've successfully done this experiment in my own home. Here's a nice video explaining. www.youtube.com/watch?v=HwRE...
- Twin Peaks
- 30 Rock
- Mad Men
- Curb Your Enthusiasm
- Better Call Saul*
*MAYBE swap for Sopranos given the day but Kim Wexler is the best written female character in TV history and I would miss her
- 30 Rock
- Mad Men
- Curb Your Enthusiasm
- Better Call Saul*
*MAYBE swap for Sopranos given the day but Kim Wexler is the best written female character in TV history and I would miss her
Sopranos
Friends
Letterkenny
Big mouth
The Office
Friends
Letterkenny
Big mouth
The Office
May 25, 2025 at 6:08 PM
- Twin Peaks
- 30 Rock
- Mad Men
- Curb Your Enthusiasm
- Better Call Saul*
*MAYBE swap for Sopranos given the day but Kim Wexler is the best written female character in TV history and I would miss her
- 30 Rock
- Mad Men
- Curb Your Enthusiasm
- Better Call Saul*
*MAYBE swap for Sopranos given the day but Kim Wexler is the best written female character in TV history and I would miss her
Reposted by Tony Roberson
This Thursday I will give a talk on (causal) mediation analysis -- happy to have finally worked out what I want to tell people.
And happy to pilot a new mode of slide sharing: just putting them on my website (juliarohrer.com/resources/)
And happy to pilot a new mode of slide sharing: just putting them on my website (juliarohrer.com/resources/)
May 5, 2025 at 10:42 AM
This Thursday I will give a talk on (causal) mediation analysis -- happy to have finally worked out what I want to tell people.
And happy to pilot a new mode of slide sharing: just putting them on my website (juliarohrer.com/resources/)
And happy to pilot a new mode of slide sharing: just putting them on my website (juliarohrer.com/resources/)
Gave the final today in my doc stats class. One student brought these cookies to share and I've never been so proud. #RStats #AcademicSky
April 29, 2025 at 12:42 AM
Gave the final today in my doc stats class. One student brought these cookies to share and I've never been so proud. #RStats #AcademicSky
Reposted by Tony Roberson
Do you have a favorite empirical example of a situation in which there is a (documented) lack of measurement invariance? Send them all my way 😊!
April 17, 2025 at 12:43 PM
Do you have a favorite empirical example of a situation in which there is a (documented) lack of measurement invariance? Send them all my way 😊!
Reposted by Tony Roberson
This is a *key* new paper in the world of school mental health interventions
A very large trial (N=6388) testing a universal CBT-based app for adolescent depression (13-14y)
No effects found (on depression, anxiety, distress or insomnia)
(🧵)
mentalhealth.bmj.com/content/28/1...
A very large trial (N=6388) testing a universal CBT-based app for adolescent depression (13-14y)
No effects found (on depression, anxiety, distress or insomnia)
(🧵)
mentalhealth.bmj.com/content/28/1...
April 25, 2025 at 8:30 AM
This is a *key* new paper in the world of school mental health interventions
A very large trial (N=6388) testing a universal CBT-based app for adolescent depression (13-14y)
No effects found (on depression, anxiety, distress or insomnia)
(🧵)
mentalhealth.bmj.com/content/28/1...
A very large trial (N=6388) testing a universal CBT-based app for adolescent depression (13-14y)
No effects found (on depression, anxiety, distress or insomnia)
(🧵)
mentalhealth.bmj.com/content/28/1...
Reposted by Tony Roberson
My first #TidyTuesday:
I really enjoyed exploring the Pixar dataset and getting more comfortable working in R with new data. Enjoyed @andrew.heiss.phd's program evaluation course this semester. Excited to keep building my R skills over the summer in #dataviz!
#rstats
I really enjoyed exploring the Pixar dataset and getting more comfortable working in R with new data. Enjoyed @andrew.heiss.phd's program evaluation course this semester. Excited to keep building my R skills over the summer in #dataviz!
#rstats
April 23, 2025 at 8:52 PM
My first #TidyTuesday:
I really enjoyed exploring the Pixar dataset and getting more comfortable working in R with new data. Enjoyed @andrew.heiss.phd's program evaluation course this semester. Excited to keep building my R skills over the summer in #dataviz!
#rstats
I really enjoyed exploring the Pixar dataset and getting more comfortable working in R with new data. Enjoyed @andrew.heiss.phd's program evaluation course this semester. Excited to keep building my R skills over the summer in #dataviz!
#rstats
Reposted by Tony Roberson
‼️ New paper @ Child Development discussing why it is inappropriate to use traditional common factor models to model adverse childhood experienced (ACEs) and other stressor inventories ‼️
Read full (brief) paper at srcd.onlinelibrary.wiley.com/doi/full/10....
@srcdorg.bsky.social #PsychSciSky
1/9
Read full (brief) paper at srcd.onlinelibrary.wiley.com/doi/full/10....
@srcdorg.bsky.social #PsychSciSky
1/9
<em>Child Development</em> | SRCD Journal | Wiley Online Library
Adverse childhood experiences (ACEs) are highly impactful stressors that increase individuals' risk for a plethora of negative developmental and health outcomes. Furthermore, minoritized groups and u...
srcd.onlinelibrary.wiley.com
April 22, 2025 at 2:24 PM
‼️ New paper @ Child Development discussing why it is inappropriate to use traditional common factor models to model adverse childhood experienced (ACEs) and other stressor inventories ‼️
Read full (brief) paper at srcd.onlinelibrary.wiley.com/doi/full/10....
@srcdorg.bsky.social #PsychSciSky
1/9
Read full (brief) paper at srcd.onlinelibrary.wiley.com/doi/full/10....
@srcdorg.bsky.social #PsychSciSky
1/9
Reposted by Tony Roberson
Hey, just a quiet reminder: it’s not true that people get more conservative as they age. *Populations* get more conservative as they age because poor and marginalized people die sooner.
nymag.com/intelligence...
nymag.com/intelligence...
Seniors Are More Conservative Because the Poor Die Off
Sometimes older voters aren’t just naturally more conservative. It’s just that wealthier and whiter people tend to live longer and healthier lives.
nymag.com
April 13, 2025 at 12:02 AM
Hey, just a quiet reminder: it’s not true that people get more conservative as they age. *Populations* get more conservative as they age because poor and marginalized people die sooner.
nymag.com/intelligence...
nymag.com/intelligence...
Reposted by Tony Roberson
📚😅🎉
Yay!! I just submitted the complete manuscript of my upcoming book to the publisher!
Learn to easily and clearly interpret (almost) any stats model w/ R or Python. Simple ideas, consistent workflow, powerful tools, detailed case studies.
Read it for free @ marginaleffects.com
#RStats #PyData
Yay!! I just submitted the complete manuscript of my upcoming book to the publisher!
Learn to easily and clearly interpret (almost) any stats model w/ R or Python. Simple ideas, consistent workflow, powerful tools, detailed case studies.
Read it for free @ marginaleffects.com
#RStats #PyData
April 10, 2025 at 7:06 PM
📚😅🎉
Yay!! I just submitted the complete manuscript of my upcoming book to the publisher!
Learn to easily and clearly interpret (almost) any stats model w/ R or Python. Simple ideas, consistent workflow, powerful tools, detailed case studies.
Read it for free @ marginaleffects.com
#RStats #PyData
Yay!! I just submitted the complete manuscript of my upcoming book to the publisher!
Learn to easily and clearly interpret (almost) any stats model w/ R or Python. Simple ideas, consistent workflow, powerful tools, detailed case studies.
Read it for free @ marginaleffects.com
#RStats #PyData
Reposted by Tony Roberson
After seeing FiveThirtyEights implementation of this idea (thank you @andrew.heiss.phd for sharing), I've been working on a medicine themed p-hacking simulator called Placebo playground: medicines.charlwood.xyz#phacking
#pharmsky #healtheconomics #medsky #pharmacy
1/3 🧵
#pharmsky #healtheconomics #medsky #pharmacy
1/3 🧵
April 3, 2025 at 7:36 PM
After seeing FiveThirtyEights implementation of this idea (thank you @andrew.heiss.phd for sharing), I've been working on a medicine themed p-hacking simulator called Placebo playground: medicines.charlwood.xyz#phacking
#pharmsky #healtheconomics #medsky #pharmacy
1/3 🧵
#pharmsky #healtheconomics #medsky #pharmacy
1/3 🧵
Just one more scale bro, please bro 🙏 we just need to make one more scale bro then it will fix everything!
I asked DeepSeek to develop a scale to measure addiction to addiction scale development and of course it delivered.
March 26, 2025 at 11:36 AM
Just one more scale bro, please bro 🙏 we just need to make one more scale bro then it will fix everything!
Reposted by Tony Roberson
I’ve long used FiveThirtyEight’s interactive “Hack Your Way To Scientific Glory” to illustrate the idea of p-hacking when I teach statistics. But ABC/Disney killed the site earlier this month :(
So I made my own with #rstats and Observable and #QuartoPub ! stats.andrewheiss.com/hack-your-way/
So I made my own with #rstats and Observable and #QuartoPub ! stats.andrewheiss.com/hack-your-way/
March 20, 2025 at 6:30 PM
I’ve long used FiveThirtyEight’s interactive “Hack Your Way To Scientific Glory” to illustrate the idea of p-hacking when I teach statistics. But ABC/Disney killed the site earlier this month :(
So I made my own with #rstats and Observable and #QuartoPub ! stats.andrewheiss.com/hack-your-way/
So I made my own with #rstats and Observable and #QuartoPub ! stats.andrewheiss.com/hack-your-way/
Reposted by Tony Roberson
Who has an academic website they’re proud of?
I’m especially interested in websites by grad students, and in websites built with Quarto. I’ll be offering a local workshop on building a website with Quarto, and I’m keen to give my audience examples. #rstats
I’m especially interested in websites by grad students, and in websites built with Quarto. I’ll be offering a local workshop on building a website with Quarto, and I’m keen to give my audience examples. #rstats
March 19, 2025 at 2:27 PM
Who has an academic website they’re proud of?
I’m especially interested in websites by grad students, and in websites built with Quarto. I’ll be offering a local workshop on building a website with Quarto, and I’m keen to give my audience examples. #rstats
I’m especially interested in websites by grad students, and in websites built with Quarto. I’ll be offering a local workshop on building a website with Quarto, and I’m keen to give my audience examples. #rstats
Reposted by Tony Roberson
Hey beautiful causal inference people! I was invited to give a workshop on mediation analysis, ca. 90-120 min.
What would you cover? Thinking both in terms of topics to be covered (e.g., mediator-outcome confounding, sensitivity analyses?) & practical implementations (R packages worth introducing)
What would you cover? Thinking both in terms of topics to be covered (e.g., mediator-outcome confounding, sensitivity analyses?) & practical implementations (R packages worth introducing)
March 14, 2025 at 9:15 AM
Hey beautiful causal inference people! I was invited to give a workshop on mediation analysis, ca. 90-120 min.
What would you cover? Thinking both in terms of topics to be covered (e.g., mediator-outcome confounding, sensitivity analyses?) & practical implementations (R packages worth introducing)
What would you cover? Thinking both in terms of topics to be covered (e.g., mediator-outcome confounding, sensitivity analyses?) & practical implementations (R packages worth introducing)
Predictor -> Oracle
Prediction -> Prophecy
Outcome -> Destiny
Prediction -> Prophecy
Outcome -> Destiny
This is a dumb #stats writing problem to have, but how can I talk about outcome and predictor variables in linear regression without using the word "predict" (which has a specific meaning to many). So:
lm(y ~ x)
That is, how do I convey to the reader what is y and what is x grammatically?
lm(y ~ x)
That is, how do I convey to the reader what is y and what is x grammatically?
March 13, 2025 at 3:43 PM
Predictor -> Oracle
Prediction -> Prophecy
Outcome -> Destiny
Prediction -> Prophecy
Outcome -> Destiny
Think this is a good way of talking about the disconnect
i think in psych people are only taught "b hat" problems (estimating effects), while the prediction language for variables is from the "y hat" problems (like in econ, where you're just building the best predicting/fitting model).
March 13, 2025 at 3:39 PM
Think this is a good way of talking about the disconnect
Reposted by Tony Roberson
Here are some cool animations I made a couple years ago for teaching DAG-based causal inference
Confounding!
Confounding!
February 20, 2025 at 9:40 PM
Here are some cool animations I made a couple years ago for teaching DAG-based causal inference
Confounding!
Confounding!