How to lie with statistics

David Sigerson, a solutions consultant with Tableau, introduces the presentation topic, 'How to Lie with Statistics.' He discusses the importance of questioning the information we encounter daily and the role of statistics in decision-making. David uses the example of Peckham rock by Banksy to illustrate how people often fail to question what's right in front of them. He also emphasizes the significance of recognizing misleading information and the impact of statistics on our perception.

The speaker delves into the concept of trust, particularly in professions such as politics, and how it relates to the acceptance of statistical information. He highlights a study ranking professions by trustworthiness and discusses the British public's perception of politicians, footballers, and nurses. David then transitions into the misuse of statistics, exemplified by the Brexit campaign's misleading claims about NHS funding. He also touches on how easily people can be misled by catchy headlines, such as those suggesting weight loss through chocolate consumption.

David explores how visual analytics can be used to highlight patterns quickly through pre-attentive visual attributes like color, size, and shape. However, he warns of the potential for misuse, as these attributes can lead to incorrect inferences if data is presented in a manipulative way. He uses examples from car advertisements and Tim Cook's presentation at an Apple conference to illustrate how data can be visually distorted to tell a more compelling story that may not accurately represent the underlying numbers.

The speaker explains the dual-process theory proposed by Daniel Kahneman, which differentiates between System 1 (fast, intuitive thinking) and System 2 (slow, logical thinking). He uses the example of Linda, a fictional character, to demonstrate how people often rely on System 1 thinking, leading to potentially incorrect judgments. David emphasizes the importance of context in decision-making and the influence of biases on our thought processes.

David discusses two cognitive biases: anchoring and framing. Anchoring refers to the tendency to rely too heavily on the first piece of information encountered when making decisions. He uses the example of a shopping experience to illustrate how initial high prices can influence how much one is willing to spend. Framing is shown through how information is presented—either positively or negatively—which can significantly affect perceptions and choices. The speaker also touches on the public's misunderstanding of statistical terms like 'means' and 'medians' and how real-world data often doesn't follow a normal distribution.

The speaker addresses the importance of comprehending data distribution and the potential pitfalls of relying on mean and median values. He uses the example of adding extremely wealthy individuals to a group to show how the mean can be skewed, while the median remains a more stable measure of central tendency. David also explains the concept of compound interest and how it's often misunderstood, leading to financial decisions that aren't in the best interest of the individual.

David warns about the potential for data manipulation, using the example of a politician's misleading tweet about salary increases. He emphasizes the need for critical thinking when presented with statistical information. The speaker also discusses the challenges of data collection, such as ensuring the reliability of survey responses and being aware of the limitations and biases in data sources. He concludes by stressing the importance of always seeking to verify the accuracy and source of statistical claims.

The speaker highlights the role of data analysis tools like Tableau in fostering a transparent and collaborative approach to data analysis. He discusses how such tools can help in formulating hypotheses, testing them rigorously, and sharing findings for peer review. David also emphasizes the importance of making data accessible for broader review within an organization, allowing more people to engage with the data and contribute to a more accurate understanding of it.

In his closing remarks, David encourages the audience to take an active role in critically evaluating statistical information and to be aware of potential manipulation. He advises providing context and references when sharing data and promoting a culture of fact-checking and transparency. The speaker leaves the audience with a final thought on the importance of staying alert for statistical manipulation and the role of individuals in ensuring the integrity of data interpretation.