Political polling has faced a number of challenges in recent years. The failure of polls to correctly forecast certain high-profile election outcomes has invoked many a post-mortem on the polling industry. For other elections however, polls have proved remarkably accurate despite operating in difficult conditions (Gelman, Reference Gelman2021). One can thus appreciate why understanding contemporary polling may seem like a daunting task to many.

In Polls, Pollsters and Public Opinion, Clifford Young and Kathryn Ziemer draw on their extensive experience of working at the polling firm Ipsos to provide an informative and user-friendly book on how to navigate the world of opinion polling in the 21st century. Their aim ‘is to aid the user of public opinion data in developing a systematic analytical approach for understanding, predicting, and engaging public opinion’ (p. 5). They are upfront from the beginning that the book is ‘presented through the lens of the principal purveyor of such data, the pollster’ (p. 1), which is important for readers to be cognisant of.

At the core of the book is their concept of the ‘three-hatted pollster’ in which pollsters possess (up to) three different identities. The first of these is that of a data scientist who assesses the quality of surveys and what deductions can be made from the data at hand. This is put forth as the essential skill for ‘any pollster worth their salt’ as otherwise ‘they risk losing their credibility’ (p. 210). The second is that of the fortune teller who seeks to use public opinion to predict future events. And finally there is the spin doctor who offers advice on how public opinion can be moved. While these are presented as distinct, it is important to remember that – like other kinds of identities – more than one of these may be activated at any given time.

The first section (Chapters 1–4) introduces what public opinion is and its societal relevance informed by especially classic but also more modern literature. While most of this material will already be familiar to (political) scholars of public opinion, it serves a key purpose in providing an accessible overview for individuals who may not have had the opportunity to study the topic before. The illustrated examples – drawing heavily from US polling – bring the concepts to life. Moreover, the 'bringing it all together’ sections that conclude the chapters are effective in focusing the readers’ mind on the key takeaways.

The second section (Chapters 5–7) centres on the data science role of the pollster. It begins with Chapter 5 on understanding bias and error, and the various considerations that polling companies have to take into account to minimise these. As the authors set out at the start of the chapter, the two key questions that pollsters need to ask themselves are ‘Is our public opinion data unbiased?’ and ‘If not, can we still use the findings to draw credible conclusions?’ (p. 55). Their answers on how to approach these issues are equally as important for researchers to be informed about if they are planning to carry out or commission original polling, or indeed for anyone who wishes to interpret survey data published by others.

Chapter 6 uses Reuters/Ipsos polling for the 2016 US presidential election as a case study to illustrate how potential bias and error in polls can be evaluated. After running through one by one various scenarios for why their polling for Hillary Clinton's vote share in their vote intention question was an outlier among the various polling firms, the authors conclude that it was as a result of providing a combined ‘neither/other’ response to respondents. This prompted them to alter the question in the midst of the time series. This is not only a thoughtful application for readers less familiar with issues of error and bias that could easily be applied to other data and contexts, but also a fascinating insight for more experienced scholars or consumers of polls into the internal mindset of a pollster when they are taking high stakes decisions. Chapter 7 takes a welcome in-depth look outside the US in a postmortem of the Greek ‘Grexit’ referendum of 2015.

The third section (Chapters 8–10) turns its attention towards the role of the pollster in predicting future outcomes drawing on examples from the US, Argentina and Brazil. The key takeaway is the importance of triangulation by drawing on information from outside of polls to further inform forecasts as ‘relying on a single source of data is a recipe for prediction disaster’ (p. 124). These may include various indicators on the state of the economy or other political indicators such as approval ratings. Such a lesson is a very valuable one to highlight that often gets lost in the midst of election campaigns when polls can become the centre of attention. It also echoes the recommendations of Jackson et al. (Reference Jackson, Lewis-Beck and Tien2020, p. 26) who had previously remarked that ‘forecasters should be extremely wary of relying on polling data alone’.

In the fourth and final section (Chapters 11–13), the authors build on the previous parts of the book by presenting the role of pollsters as spin doctors who can ‘provide inputs to the marketer or PR professional, so that they might devise impactful communications pieces’ (p. 167). The Brazilian presidential election of 2022 is invoked to show how context, the key issues of the day, and public perceptions of candidates all combine together to reveal important information for the possibilities for campaigns to attract the support of key population segments.

In sum, Polls, Pollsters, and Public Opinion is a welcome addition to the public opinion literature. It delivers on its objectives to ‘produce a toolbox for the practitioner’ and to act as a resource ‘for anyone who wants to analyse public opinion and use it in decision-making’ (p. 210). Its clear structure and engaging examples make it an accessible introductory text for students, researchers and aspiring pollsters who wish to gain a foundation in key practical and conceptual issues in survey research. Given that ‘political and economic actors so often underestimate or misunderstand public opinion’ (p. 1), as well as broader issues of public misunderstanding of polls (Graefe, Reference Graefe2021) and sometimes problematic media reporting that either misrepresents polling data or uncritically reports upon findings from low-quality polling (Graham et al., Reference Graham, Hillygus and Trexler2024), the focus of the book in equipping readers to critically engage with polling is to be welcomed.