Sign In

How Logical Fallacies Can Ruin Data Analyses

2 minutes

Analytics results can be easily undermined by logical fallacies. It is important to understand, avoid, and identify them when they occur. This discussion outlines key fallacies and shows how the inductive approach helps in preventing their occurrence.

Type 1: Misplaced Intuition

These fallacies typically come from non-quantitative perspectives, and one of the main goals of analytics is to avoid them through sound, objective analysis.

Hasty generalization is deriving conclusions from limited examples and is a result of not understanding basic statistics and outliers.
False analogy is comparing things that aren't completely comparable and is due to a superficial understanding of the things to be compared.

Type 2: Methodological Manipulation

Quantitative practitioners who lack sufficient domain expertise often commit these fallacies. Under pressure to produce results, there is a belief that if something is thoroughly analyzed with mathematical tools, the results must be valid. After all, data is seen as powerful and expected to always provide answers.

Analytical manipulation is the use of analytics tools to justify a conclusion (storytelling) rather than to test a hypothesis (scientific method).
Lack of understanding is due to mistaking data observations for understanding the underlying mechanism. Data does not measure the mechanism, and correlated data doesn't necessarily indicate a casusal relationship.

Type 3: Illogical Reasoning

These are some of the most insidious fallacies and can affect anyone involved in or outside the analytics process. It is important to remember that logic is the foundation of any analysis, and no matter how compelling the results appear, they must follow logical principles.

Incorrect inference is the claim that a factor is causal based on observation and not the underlying mechanism. Positive observations only provide evidence, not proof of validity whereas a single negative observation proves invalidity of the factor.
A non-causal factor will inevitably fail and there are many logical fallacies used to explain its failure.
Recommending non-factors is harmful and should not be done. Common sense can't override logic.

Top-Down Problem Construction

Are we at a total loss? Is there nothing that can be done? Of course not. What is needed is a carefully organized analytics process that assigns each resource to the role where they can deliver the best results and prevents situations where logical fallacies could arise and reach the outcomes.

Customers and consultants are most likely to correctly identify factors and should be assigned to this role. Analysts are best at calculating factors and machines are best at optimizing them. Utilizing each resource for what it does best is a cornerstone of the inductive approach.

Beware of Ultracrepidarians

Logical fallacies, especially when combined with inherent content bias, are a serious issue across many fields. See how many you can spot the next time you watch a news report. This is not about being pedantic or making analytics harder to execute, but about ensuring quality and delivering the best results for customers, which builds trust in analysts and helps both succeed.

One quick poll question (3 seconds):
I’d really value your input.

Sign up for early access to articles, delivered every two weeks. Please check your spam folder for the confirmation email.

Ready to see Advanata in action? Explore more resources, read related articles, and access the full application at advanata.com.

Want a personalized walkthrough? Contact info@advanata.com to get application access and schedule a demo.

Leave a comment