Matthew Gaertner

Great to see some more data with positive (albeit mixed) news on student learning in the pandemic. Pretty vital research coming out of commercial curriculum & assessment providers these days. #ResearchNetwork4ever @KristenDiCerbo @McKni8 @eddatadude wsj.com/articles/new-t…

@AndrewDeanHo Congrats! So impressive! I can still call you Andrew though?

I am floored by this... thrilled to be elected along with this distinguished class and honored to join this community of scholars I have looked up to my entire career.

WAAHOWWWWW youtu.be/udvYSd2TIkg via @YouTube

Check out this Senior Research Associate (Psychometrics) opportunity recruiting2.ultipro.com/WES1032WSTE/Jo…

This is great @KatieMcClarty! This is how it should be done. @DerekCBriggs @MattBrunetti @AndrewDeanHo @nsndc renaissance.com/how-kids-are-p…

Oh, tremendous, let’s debrief. The only thing worse than losing is learning from it.

@ScottFMarion @NCIEA1 @AndrewDeanHo @DerekCBriggs @TColadarci Also - I think their excuse for reporting results for registered voters rather than likely voters is a total cop out. Because it’s hard to predict? Really, I doubt COVID changes the turnout mechanisms. A not-great voter turnout model is better than none at all.

A good challenge for my geeky measurement colleagues @NCIEA1 @AndrewDeanHo @DerekCBriggs @TColadarci @MatthewGaertner. Personally, I'd like to see a way to quantify the non-response bias and build that into the confidence intervals. Have at it...

@ScottFMarion @NCIEA1 @AndrewDeanHo @DerekCBriggs @TColadarci I wonder if tipping point analysis could help us figure out if nonresponse adjustments are reasonable. It’s done for efficacy studies that use imputation; survey weights have the same purpose (dealing w missing data) ww2.amstat.org/meetings/biop/…

@ScottFMarion @NCIEA1 @AndrewDeanHo @DerekCBriggs @TColadarci But it’s true that if nonresponse bias is a real problem, confidence intervals will expand, b/c highly variable weights introduce a design effect equal to 1+ (coefficient of variation for the weights)^2

@ScottFMarion @NCIEA1 @AndrewDeanHo @DerekCBriggs @TColadarci Actually I think nonresponse bias implies we need an adjustment to the *point estimate* not the confidence intervals.

“They Sellin’ Us a Dream They Not Preparin’ Us for”: College Readiness, Dysconscious Racism, and Policy Failure in One Rural Black High School link.springer.com/article/10.100…

Integrating Timing Considerations to Improve Testing Practices taylorfrancis.com/books/e/978135…

Margolis: Integrating Timing Considerations to Improve... - Google Scholar #d=gs_qabs&u=%23p%3D7NQjyDPBBn0J" target="_blank" rel="nofollow noopener">scholar.google.com/scholar?cluste…

Learning Gains From the KinderTEK® iPad Math Program: Does Timing of a Preventative Intervention Matter? journals.sagepub.com/doi/abs/10.117…

We have four great doc students joining us for the 15th consecutive year of our summer internship program. We look forward to working with and learning from them. We just wish we could do it in person! @DerekCBriggs @SireciSteve

Catching up or Falling Behind: Measuring Middle School Achievement Trajectories for College Readiness edworkingpapers.com/sites/default/…

My favorite assumption is probably the one that spreads the impact of missing 4 months in 4th grade over 45 years of work. At age 60, those 4 lost months are right there, haunting you.

I also appreciate the authors who’ve cited this artifact, cycled through adjectives like “concerning” and “eye-popping” before settling on “unacceptable.”

Is anyone asking for these predictions? Why? The methods section has a typo. It should just read: ”I don’t even know just pick the middle one. Yeah for 45 years. Jesus dude can we go? It’s fine just send it.” brookings.edu/blog/education…