Me when my Bayesian SEM model samples properly and quickly the first time around:

Option 2

Option 1

Inspired by true events this morning, huge pet peeve of mine.

Four hours straight of class plus a two-hour commute. Then you see your neighbourhood pub has a beer from your favourite brewery.

Here in Catalunya we don't have Halloween as a blocker, so Christmas chocolate is already on display in the grocery store:

Was wondering how everyone was seeing these for the past week because I only ever browse on my phone.

Celebrating a complete rough draft of the thesis with one of my all-time favourites.

We ended up using Bayesian finite mixture models, using two distributions to model IntDiff. This worked much better, and we interpreted this as evidence that there are two pronunciation variants of the tap. Our predictors were now used to predict gradient and categorical variation in Spanish taps.

When we went to model the intensity difference between the tap and the surrounding vowels (IntDiff). We found that hierarchical linear models were misspecified, and we were left to find alternatives to model changes in IntDiff with our predictors of interest (e.g., lexical frequency, speech rate)

After a nice Christmas break, I now have approximately one month to get a copy of my dissertation to my committee. Do you know what that means? For me, it means lots of pomodoros and two cups of delicious coffee a day!

I like to think that including memes in my teaching slides helps me drive home the main takeaways and make it memorable for students but to be honest? I'm pretty sure it's mostly for my own amusement.

Working in Praat today, which means breaking out the gaming pad. A lot of Praat keyboard shortcuts aren't the most comfortable (e.g., zoom to selection), and this lets me program macros to make things easier. I can move to zero crossings, create boundaries, and paste from clipboard in a single click