Many people are unduly intimidated by the thought of teaching probability. At its most basic level, probability revolves around being able to predict whether a future event is certain, possible, or impossible. These basic concepts grow and develop into a more mathematical understanding of likelihood as children learn to quantify that…

Probabilistic machine learning increasingly informs critical decisions in medicine, economics, politics, and beyond. To aid the development of trust in these decisions, we develop a taxonomy delineating where trust in an analysis can break down: (1) in the translation of real-world goals to goals on a particular set of training data, (2) in the…

Aside from being culturally relevant, artificial intelligence is also supporting companies in making business decisions. Consequently, "workforce needs have shifted rapidly," resulting in a demand for applicants who are skilled in "data, analytics, machine learning, and artificial intelligence" (Miller and Hughes 2017). This…

A strategy to facilitate understanding of spatial randomness is described, using student activities developed in sequence: looking at spatial patterns, simulating approximate spatial randomness using a grid of equally-likely squares, using binomial probabilities for approximations and predictions and then comparing with given Poisson…

Artificial intelligence-enabled adaptive learning systems (AI-ALS) are increasingly being deployed in education to enhance the learning needs of students. However, educational stakeholders are required by policy-makers to conduct an independent evaluation of the AI-ALS using a small sample size in a pilot study, before that AI-ALS can be approved…

During modern general election cycles, information to forecast the electoral outcome is plentiful. So-called fundamentals like economic growth provide information early in the cycle. Trial-heat polls become informative closer to Election Day. Our model builds on (Linzer, 2013) and is implemented in Stan (Team, 2020). We improve on the estimation…

This post is one in a series highlighting MDRC's methodological work. Contributors discuss the refinement and practical use of research methods being employed across the organization. Across policy domains, practitioners and researchers are benefiting from a trend of greater access to both more detailed and frequent data and the increased…

The "radioactive dice" experiment is a commonly used classroom analogue to model the decay of radioactive nuclei. However, the value of the half-life obtained from this experiment differs significantly from that calculated for real nuclei decaying exponentially with the same decay constant. This article attempts to explain the discrepancy and…

Logistic regression is slowly gaining acceptance in the social sciences, and fills an important niche in the researcher's toolkit: being able to predict important outcomes that are not continuous in nature. While OLS regression is a valuable tool, it cannot routinely be used to predict outcomes that are binary or categorical in nature. These…

Tossing a fair coin 1000 times can have an unexpected result. In the activities presented here, players keep track of the accumulated total for heads and tails after each toss, noting which player is in the lead or whether the players are tied. The winner is the player who was in the lead for the higher number of turns over the course of the game.…

Education systems are increasingly creating rich, longitudinal data sets with frequent, and even real-time, data updates of many student measures, including daily attendance, homework submissions, and exam scores. These data sets provide an opportunity for district and school staff members to move beyond an indicators-based approach and instead…

"Probability is a difficult concept to teach, because children and adults find it counterintuitive." This is impetus to consider the detailed planning of a set of lessons with a "mixed", in many senses, group of fourth graders. Can the use of prior experience, and the knowledge associated with that experience, make probability a concept that is…

The linear regression model is one of the most popular models in statistics. It is also one of the simplest models in statistics. It has received applications in almost every area of science, engineering and medicine. In this article, the authors show that adding a predictor to a linear model increases the variance of the estimated regression…

The methods and models for categorical data analysis cover considerable ground, ranging from regression-type models for binary and binomial data, count data, to ordered and unordered polytomous variables, as well as regression models that mix qualitative and continuous data. This article focuses on methods for binary or binomial data, which are…

This article describes the development of early warning indicators for high school and beyond in the Milwaukee Public Schools (MPS) by the Value-Added Research Center (VARC) at the University of Wisconsin-Madison, working in conjunction with staff from the Division of Research and Evaluation at MPS. Our work in MPS builds on prior early warning…