Advanced Quantitative Toolkit

Trying to wrap your head around how to apply Bayesian reasoning to multivariate modelling? This article illustrates how these concepts intersect, with practical steps for applying it to your own research.

This article offers a step-by-step guide on how to conduct a multilevel analysis, with an accompanying R script and dataset that make it possible to practice the concepts outlined.

As a social scientist, you may deal with context-dependent clustered data. This chapter will showcase one statistical model to account for the relationships between observations in this type of data.

Working with clustered data and unsure how to account for within-group variability? This lecture introduces Bayesian hierarchical models as a way to model the relationship between an outcome and predictors for clustered data.

Review the general framework for GLMs, then learn how Bayesian modelling can improve estimates about the relationships between variables in your data.

This rigorous discussion of Bayesian modelling is grounded in an approachable example.

Learn how to apply Bayesian reasoning to estimate linear model parameters.

Sometimes frequentists statistics’ assumptions don’t apply to qualitative data- this video explains the advantages of Bayesian statistics in these cases.

Unsure about how to solve classification problems? Learn more about the statistical models used in LDA and QDA.

Learn how to conceptualise LDA and QDA within the context of Gaussian Maximum Likelihood Classification methods.

Step-by-step guides for discriminant analysis and logistic regression procedures

Model selection can sometimes be more of an art than a precise science, and this video tutorial illustrates this by walking through an example in R.

This model allows you to get an accurate estimate of how a certain variable changes over time, by factoring in past observations of the same variable.

This concise explanation of time series models is a quick guide or refresher on when to use which.

This article offers a primer on GAMLSS, as a robust model for data with a distribution which deviates from the normal because of its shape (and hence violates the assumption of other commonly used models like GLMs). The authors provide examples in R using 4 different examples.

Trying to go beyond the mean with your regression model? When neither GLM nor GAM provide an adequate picture, GAMLSS might be the place to start.

You may know about GAMs, but do you know how to make one in R? Watch this code-snippet-infused lecture and learn how to make your own model.

Are you unsure which model to use for your truncated data? This lecture may give you a better sense of what each model does and when to use which.

Sometimes a linear model just doesn't cut it- in these cases, we can extend the model to deal with cases where data violates important assumptions such as linearity.

If you’re unsure of when to use a simple linear regression model or how to interpret one, this tutorial explains the basics with a simple example.

urious about what makes up a linear regression model? This tutorial provides a simple introduction to generalized linear models and the structure of the most common types.

Article providing an accessible and detailed introduction to multilevel models, with examples.

Short article discussing methodological challenges with secondary data analysis.

Choosing an appropriate and well-powered sample is crucial for effective research. This paper explores best practices for qualitative research.

How do you develop sampling designs in mixed methods research? This paper presents a framework for researchers.

How do you make sure your study is valid and reliable? This video provides guidelines for researchers.

R is a powerful and efficient tool for survey question analysis- this tutorial explains how to harness it for your own research.

How we frame research questions dictates the types of conclusions we can make from our data. This resource explains key considerations for survey question design.

This article offers guidance on the survey process, with a section on designing questions that are best suited for rigorous analysis.

This brief overview of DiD explains the method’s assumption, strengths and limitations, and best practices.

This comprehensive guide to DiD is organised into the sequence of steps needed to get an estimate of an effect

This resource starts from the basics, providing a simple introduction to DiD methods.

This resource is an overview on Differences in Differences methods, including treatment effects estimation and region-based methods.

Starting with a review of the canonical model, this resource advances to the deviations from the model that researchers use currently.

Learn about time-series cross-section methods in this approachable book chapter.

An introduction to TSCS models that explores its advantages. This chapter also describes methods for minimizing errors and generate more robust conclusions.

Learn about the complexities of TSCS data analysis.

TSCS modelling is an iterative process- while there is no single way to analyse this type of data, this article provides strategies to model its dynamic and cross-sectional components.

Learn how to choose the best model for your panel data with this thorough introduction to panel data modelling.

This resource explains how to estimate causal effects in longitudinal data using OLS and fixed effects models.

Unsure how to apply the theory of tree partitioning to social science examples? This tutorial is an accessible tutorial.

Machine learning methods have a lot to offer for social science research- this article introduces basic principles of tree based partitioning and explains key terminology.

Unsure how decision trees, root nodes and leaves can help you learn from your data? This resource explains decision trees and partitioning with plenty of visual aids.

How to we visualise the process of argument logic? This journal introduces AVIZE, a diagramming tool for argument schemes.

How do we use RDF triples to carry out efficient SPARQL queries? This video breaks down four methods.

How do we make sense of massive amounts of population data? Data dashboards use large amounts of data to convey information at a glance.

How do you manage large amounts of information as a researcher? This textbook chapter introduces database management systems and how to use them.

How do we use mathematics to study people, and what assumptions are baked into our methods? This talk challenges researchers to thoughtfully consider the statistical methods they follow.

An introduction to content analysis and its application to the largely implicit concept of identity.

Article providing an in-depth overview of multilevel analysis, with examples

Short article presenting a systematic process for using secondary data to conduct research

What type of experiment can justify causal claims? This video presents guidelines on experimental design

This book chapter provides a conceptual introduction to MDS.

How to you conduct a multidimensional scaling analysis in R? The idea of multidimensional scaling is explained using intuitive examples.

How do you quantify the similarity among items in your data? This paper offers an overview of multidimensional scaling, a technique that allows you to visually represent similarity in your data.

What scaling technique is best suited to your data? These book chapters provide a concise definition of data theory, followed by a discussion of scaling techniques.

What can geometric data analysis offer social scientists? The book chapter provides a nontechnical introduction to this method as a way to link geometric representations to statistical interpretations.

What are the conceptual underpinnings of correspondence analysis? This guide provides a thorough answer to this question by using an example on crime statistics.

Unsure how to translate your conceptual knowledge about PCA to practical data analysis skills? This resource breaks down PCA in R in a beginner-friendly way.

This resource provides a comprehensive introduction to principal components analysis with relevant examples such as crime statistics and determinants of economic expenditures.

This video explains the power of factor analysis as a way to reveal the underlying basic factors from a more complicated set of variables.

This article makes it clear the differences between descriptive methods like PCA and modelling methods like factor analysis; both should be used for the most robust research outcomes.

Unsure how to evaluate your MIRT model? This book chapter explains how to diagnose these models and how to choose the best one.

A thorough guide on summated rating scales based on classical test theory.

What distinguishes the social science approach to machine learning (ML) from the rest of ML? This review encourages readers to apply conventional ML techniques to data collection and causal inferences.

What can computer science offer social science researchers? This keynote from a computational social scientist describes how machine learning modelling is used to uncover the latent structure in count data.

How can machine learning move you from unstructured text to well-defined vectors? The bag of words technique allows you to convert text into the numerical data which machine learning models rely on.

Feature extraction is a technique that encompasses several techniques. This blog post provides a summary and evaluation of the most popular ones.

Unsure how sentiment analysis works, or how to apply it? This book chapter provides a conceptual and practical introduction to sentiment analysis.

Learn how to tailor apply word-count techniques to a pre-defined group of words in R.

This video introduces quantitaive text analysis, including character encoding, GREP and cleaning text data in R.

This video explains topic modelling with an analysis of scientific journal abstracts.

What is topic modelling and how can it enhance your language data analysis toolkit? This article introduces topic modelling, exploring its applications for social science research.

What do natural language processing (NLP) methods have to offer? This talk explores the intersection between NLP and two questions in social science.

How do you minimise selection bias when estimating treatment effects? Regression discontinuity designs allow you to leverage quasi-randomisation by looking at data points around a threshold.

How do you estimate treatment effects of a continuous variable, without selection bias? Regression discontinuity has emerged as a powerful experimental design for this purpose.

This book chapter is an introduction to conjoint analyses, with theory grounded in political science examples.

How do people choose between options that vary in multiple ways? Conjoint experiments and analysis are well suited to ask these questions, where people rate multidimensional options.

Understanding the nature of your data, and whether it is parametric or nonparametric, will dictate the types of inferential statistics you can do with your data. This guide provides a simple explanation.

An article providing an in-depth overview of multilevel analysis, with applied examples using the the TIMSS dataset.

A short article discussing biases that arise with secondary analysis of large national survey datasets and related considerations.

A detailed article presenting a simple process for using secondary data to answer research questions.