Statistics Concepts for Laypeople

https://www.kdnuggets.com/7-statistical-concepts-every-data-scientist-should-master-and-why

A good article on statistical concepts for laypeople.

• Statistical Significance vs Practical Significance: the difference between a result being mathematically reliable and being important in practice. For example, a medicine reduces temperature by 0.01 degrees — statistically, this might be significant across a million people, but practically it’s useless.
• Sampling Bias: when data poorly reflects the real situation and leads to incorrect conclusions. For example, a survey about fear of flying at an airport — you won’t hear from those who are so afraid they stayed home.
• Confidence Intervals: instead of a single number — a range where the true value most likely lies. For example, a navigator promises arrival at 18:00 ± 5 minutes. These 10 minutes are the interval.
• P-values: help estimate how likely the observed effect could have occurred by chance. For example, if you lost weight after a new diet and the p-value is 0.05, it means there’s a 5% chance you lost weight by chance, not because of the diet.
• Types of Errors in Tests (Type I and II): false positives vs false negatives. For example, a Type I error — the alarm goes off without burglars; a Type II error — burglars entered, but the alarm is silent.
• Correlation vs Causality: two things may be related, but one doesn’t necessarily cause the other. For example, children with larger feet write dictations better not because of their shoes, but because they are older.
• Curse of Dimensionality: too many features can worsen model results. For example, you are looking for similar apartments to estimate value. By one parameter (area), it’s easy to find 10 similar ones in a database of 1000. Add the floor — you already need 10,000 apartments. Add the neighborhood, year of construction, wall material — and for a reliable comparison, you’ll need millions of records that simply don’t exist.