Bayesian Linear Mixed Models: Random Intercepts, Slopes, and Missing Data

This past summer, I watched a brilliant lecture series by Richard McElreath on Bayesian statistics. It honestly changed my whole outlook on statistics, so I couldn’t recommend it more (plus, McElreath is an engaging instructor). One of the most compelling cases for using Bayesian statistics is with a collection of statistical tools called linear mixed models or multilevel/hierarchical models. It’s common that data are grouped or clustered in some way.

Modeling Motivation and Emotion using Feedback Loops

If you’re anything like me, you probably set a lot of goals. Whether it’s to finish a paper by the end of the summer or to spend more time with friends and family, goals are what help motivate us to do something. Goals are also intimately tied to our feelings. You may have had the experience of falling behind in your goals, which made you upset, but ultimately motivated you to step up your efforts.

A Model and Simulation of Emotion Dynamics

Emotion dynamics is the study of how emotions change over time. Sometimes our feelings are quite stable, but other times capricious. Measuring and predicting these patterns for different people is somewhat of a Holy Grail for emotion researchers. In particular, some researchers are aspiring to discover mathematical laws that capture the complexity of our inner emotional experiences - much like physicists divining the laws that govern objects in the natural environment.

Simulating Emotions during a Basketball Game - Just a Feeling in the Crowd

Sporting events host witness to a wide range of human emotion. The emotional ups and downs are especially clear among invested fans. Fans experience the joy and excitement of a triumphant comeback, or the anxiety and disappointment of a loss. It is particularly interesting to see how emotions differ from two opposing fan groups watching the same match. I decided to perform some simulations on how a crowd of fans would react during a basketball game.

Quick Example of Latent Profile Analysis in R

Latent Profile Analysis (LPA) tries to identify clusters of individuals (i.e., latent profiles) based on responses to a series of continuous variables (i.e., indicators). LPA assumes that there are unobserved latent profiles that generate patterns of responses on indicator items. Here, I will go through a quick example of LPA to identify groups of people based on their interests/hobbies. The data comes from the Young People Survey, available freely on Kaggle.

Network Analysis of Emotions

In this month’s post, I set out to create a visual network of emotions. Emotion Dynamics tells us that different emotions are highly interconnected, such that one emotion morphs into another and so on. I’ll be using a large dataset from an original study published in PLOS ONE by Trampe, Quoidbach, and Taquet (2015). Thanks to Google Dataset Search, I was able to locate this data. The data is collected from 11,000 participants who completed daily questionnaires on the emotions they felt at a given moment.

The Face of (Dis)Agreement - Intraclass Correlations

I was recently introduced to Google Dataset Search, an extension that searches for open access datasets. There I stumbled upon this dataset on childrens’ and adult’s ratings of facial expressions. The data comes from a published article by Vesker et al. (2018). Briefly, this study involved having adults and 9-year-old children rate a series of 48 faces on two dimensions of emotion, valence (positive vs. negative) and arousal (activated vs.

Plotting the Affect Circumplex in R

I’m a strong adherent to the circumplex model of emotions introduced by James Russell in the late 1980s. Russell argued that all emotional experience can be boiled down to two dimensions: valence and arousal, with valence being how positive or negative you feel and arousal being how sluggish or emotionally activated you feel. The emotions we commonly label as anger, sadness, joy, etc. can be mapped within this affective two-dimensional space, such that joy is a high valence, high arousal emotion, whereas boredom is a moderately low valence and low arousal emotion.

A New Way to Handle Multivariate Outliers

Psychologists often have a standoffish attitude toward outliers. Developmental psychologists, in particular, seem uncomfortable with removing cases because of the challenges inherent in obtaining data in the first place. However, the process of identifying and (sometimes) removing outliers is not a witch hunt to cleanse datasets of “weird” cases; rather, dealing with outliers is an important step toward solid, reproducible science. As I’ll demonstrate in this simulated example, a few outliers can completely reverse the conclusions derived from statistical analyses.