You are considering investing in a company stock, and you want to know how risky that investment is. In finance, a relevant measure of risk relates returns on a company stock to market returns: a company stock is considered risky if it tends to move in the direction of the market, and the more it moves in that direction, the riskier it is. You have downloaded data on daily stock prices for many years. How should you define returns? How should you assess whether and to what extent returns on the company stock move together with market returns?

You want to know whether online and offline prices differ in your country for products that are sold in both ways. You have access to data on a sample of products with their online and offline prices. How would you use this data to establish whether prices tend to be different or the same for all products?

Identifying a good research question is a vital first step in any behavioural study because the question will focus the rest of the research cycle. Four logically distinct types of question can be asked about any behaviour. These concern its mechanisms, its development (or ontogeny), its function and its evolution (or phylogeny). The mechanisms underlying behaviour can be studied at many different levels, ranging from the social or physical environmental conditions that influence the behaviour down to the neural networks responsible for behavioural output. The nature of the research question will influence decisions about what species to study. Research questions are developed through a combination of approaches, including reading the literature, preliminary observations and exploratory data analysis. A research question leads to a set of hypotheses that need not be mutually exclusive but should all be testable. Each hypothesis should generate one or more specific predictions.

Statistical analysis is usually necessary to answer questions with behavioural data. Analysis should be planned and registered before collecting data. Once collected, a dataset should be formatted and permanently archived prior to analysis. Data is checked and visualised with descriptive statistics and graphs. Models representing hypotheses about the true effects present in the population from which the dataset is a sample are built and tested with inferential statistics. Many different hypotheses can be captured using a linear modelling framework in which an outcome variable is predicted with a combination of predictor variables and interactions. Sources of non-independence in datasets can be addressed with mixed models. The robustness of findings can be examined by comparing the results obtained when analysis is done in different ways using model selection and multiverse approaches. Confirmatory analysis designed to test preregistered hypotheses should be clearly differentiated from exploratory analysis that generates new hypotheses.

You work for a company that wants to quantify the benefits of its online advertising: how many people buy its product because they see an ad posted online. How can you translate this question into something you can uncover using actual data? What kind of data do you need to get a good answer to this question? What would be the most important issues to consider with that data?

Measuring behaviour means assigning numbers to observations of behaviour according to specified rules. Converting a stream of behaviour into behavioural metrics involves choosing and defining specific categories of behaviour that can be measured. Behavioural categories can be described in terms of their physical structure or their consequences. An ethogram is a catalogue of the species-typical behavioural categories displayed by a species in a specified environment. Descriptions of behavioural categories should be unambiguous and written down before data collection starts. Behavioural categories can be designated as either events (short duration) or states (longer duration). Behavioural categories are used to generate metrics such as latencies, frequencies, durations and intensities. Two or more metrics can be combined to form a composite metric. Metrics can be at different levels of measurement, ranging from nominal (weakest) to ratio (strongest).

What are the benefits of immunization of infants against measles? In particular, does immunization save lives? To answer that question you can use data on immunization rates and mortality in various countries in various years. International organizations collect such data on many countries for many years. The data is free to download, but it’s complex. How should you import, store, organize, and use the data to have all relevant information in an accessible format that lends itself to meaningful analysis? And what problems should you look for in the data, how can you identify those problems, and how should you address them?

Predicting whether people will repay their loans or default on them is important to a bank that sells such loans. Should the bank predict the default probability for applicants? Or, rather, should it classify applicants into prospective defaulters and prospective repayers? And how are the two kinds of predictions related? In particular, can the bank use probability predictions to classify applicants into defaulters and repayers, in a way that takes into account the bank’s costs when a default happens and its costs when it forgoes a good applicant?

The goal of the book as a whole is to ‘translate’ coin evidence for a new generation of historians. The work of Michael Crawford represented a major leap forward in the study of Roman republican coins during the twentieth century while on the work of earlier generations.The major thematic structure of the book is summarized, and eight basic principles related to the use of coin evidence are laid out.

You want to predict the price of used cars as a function of their age and other features. You want to specify a model that includes the most important interactions and nonlinearities of those features, but you don’t know how to start. In particular, you are worried that you can’t start with a very complex regression model and use LASSO or some other method to simplify it, because there are way too many potential interactions. Is there an alternative approach to regression that includes the most important interactions without you having to specify them?

You want to understand whether and by how much online and offline prices differ. To that end you need data on the online and offline prices of the same products. How would you collect such data? In particular, how would you select for which products to collect the data, and how could you make sure that the online and offline prices are for the same products?