Matt Hawrilenko

@quantitudepod “And this, class, is a survival curve!” said the Adjunct Professor.

Hey Quantitude friends, let’s get your creative juices flowing again. For this week’s procrastination-enabling fun, please post (by Thursday) a clever statistics, quantitative methods, or academia related caption for the image below.

@tchanpoker How much can the game have changed? Sure these kids have their math and their solvers but you know what I have? 11 years of saved up one-times!

@tchanpoker Flop like a butterfly, stick around like a cockroach.

@DepressionLab The behaviorist in me is hearing you say that you want to know what that energy is… a function of?

new goal: how do I start my next talk off with this energy?

@NeurosciTheater “First time fathers”. That’s why I had two!

@quantitudepod Dear SEMNET, my chi-square is significant. Is that ok?

For this last Quantitude caption contest of the summer, please post (by Wednesday) a fun, quantitative methods or academia related caption for the image below. Thanks!

@jacasiegel Blackberry picking! One of our very favorite places is kite hill in Magnuson park. Loads of blackberries at the top, alongside amazing views of the water and a little swimming area

Flights booked! I’ll be headed to Seattle this August for a few days. What would I be silly to miss? Additional info: 1) I’ll be there with my partner (send your date recs!) 2) I’m a vegetarian 3) We love coffee + beer + theatre + outdoorsy stuff

@billchenpoker @quantitudepod I think your prediction here is about as accurate as the predictions from the model I referenced above :-) don’t worry—you’re still the smartest person I know (that may not be saying a whole lot…)

@hoss_tbf @quantitudepod Sorry to hear. I've heard this from others too. Wonder whether it's a short term gain though like if you were separated for months whether it would persist. Maybe something worth studying.

Got covid and quarantining from my family (3 kids <= 6). On the downside, I have covid. On the upside, wow is productivity easier when I abdicate family. For example, today I tested a prediction model out of sample and produced r^2 = -13. Is that good @quantitudepod ?

@DepressionLab LOL. That’s one way to show “proof” of your productivity!

@hoss_tbf also, as fellow Massachusetts natives, first beer is on me if we ever meet in real life. I imagine you have some interesting stories to tell... you can tell me about WSOP and I'll tell you about how I accidentally started a whiskey distillery during grad school

This fills up my heart, seriously. Literally talked about @itschekkers and Spring in 3 separate meetings today. In one, was meeting w a student and used it as an example of impact and scale of exactly what you're talking about. So happy for you, & so happy for everyone impacted!

@DepressionLab Yeah you’ve done such awesome stuff over there. Just know that we’ll be here when you’re ready.

don't have any plans to leave academia currently, but as I've told Adam many times, Spring is the only place I could really imagine going to if I did. But I really love the personalized digital therapy we've developed at the DGC and Spring has their own so it wouldn't make sense.

Hey @DepressionLab we’ve never met, but just wanted to give you a shout because when I was waffling about applying to Spring’s research scientist position last year, you tweeted about it and it pushed me over the edge. So—thanks for helping me jump ship to a job I LOVE :-)

Academic Job alert! I left my faculty job at UW psychiatry in January. I liked my job and loved my colleagues. I came to Spring because I wanted to do precision mental health research, at scale, and roll it out. Best. Gamble. Ever. Come work with me! jobs.lever.co/springhealth/c…

@cjvanlissa @bringmann_laura @EaraYoung @SRAdolescence @SRCDtweets I think ALSPAC has everything you want and more! bristol.ac.uk/alspac/researc…

@HeatherTCooley The version I got was to write it “almost without references.” The rationale was that an aims page should be intuitive, uncontroversial, and out-of-the-weeds. The reviewer should just nod along in agreement the whole time. This may have been intended to balance my tendencies :-)

10pm, made myself an emergency chocolate cake-in-a-mug, had to substitute literally every single ingredient and now I finally understand the limitations of synthetic data

@quantitudepod Loved the episode! And… disappointed to learn that NPR cites studies supporting its hypotheses as liberally as I do 🙈

S3E16: Your COVID Rapid Test Result: Are You Positive You’re Positive? buzzsprout.com/639103 In today's episode, P & G use COVID rapid tests to discuss sensitivity, specificity, positive and negative predicted values, and the questionable utility of test accuracy information.

Cancelled meetings and cut off my work day 5 hours early to pick my 6 year old daughter up from school today because, apparently, she had to fart.

@EikoFried @JoranJongerling I’ve seen you make this point before and it feels really important to me. Any chance you’ve published it, or know of another citation for it?

@JoranJongerling 3) I've modeled tons of EMA data at both contemporaneous and temporal level, even with 10 measurement points per day. Contemporaneous network is always more connected than temporal, implying that we undersample even with 10x per day for most processes we curr measure (e.g. emots)

@dp_moriarity Oh! Yeah that’s more boring :-) CFA away, then, ideally just a one group model across all 5 time points. And note that you can get away with releasing some equality constraints (partial measurement invariance). Todd Little has a great chapter on this.

@hoss_tbf Thanks for all this info! But I also think there was a bit of a miscommunication, all of the visits are for the same people, the measures just changed because the cohort started as adolescents but continued into adulthood!

Have 5 timepoints of symptoms, but measure changes after 3 visits (child>adult). Measures were created to pair well so wording is near identical. Would a multigroup test (1st child vs. 1st adult) of measurement invariance suffice to see if they can be used in the same model?