Are you constantly bombarded with data, facts and statistics?
We used to believe that if we had enough data about a topic or a situation that we could understand anything. That was until we got mountains of data and no way to parse it, no way to distil it down into something we can use. Of course there are data scientists whose job it is to make big data usable for us but what about the everyday stuff? The sales dashboards with granular detail about every lead, the spreadsheet with the production rate of a team on the factory floor. How do we know what information to trust and follow? Knowing what data to keep and what to ignore is becoming an ever more important skill. Use this podcast to understand how to interpret statistics to enable you to make quicker, more laser focused, better-informed decisions.
What you will learn:
3 types of probability –
- Classic – conditions are perfect, roll a die/flip a coin
- Frequentist – parameters must be estimated i.e. will this drug work?
- Subjective – this is the degree of belief a person has about something occurring
3 ways to acquire information –
- Discover it ourselves
- Absorb it implicitly
- Be told it explicitly
To identify expertise –
Ask where the information came from: Institutional bias – is there a conflict of interest? Who links to the webpage? – use link:[URL] to find out Is the research peer-reviewed? if so, by who? Is the information current? or discredited? Is there supporting information?
Correlation is not causation –
i.e. 2 lines on a graph matching up does not mean that one line caused the other e.g. ice cream sales increase as the number of people wearing shorts increases. This is NOT causation.
Sampling can be flawed
- Participation bias
- Reporting bias
- Measurement error
- How something is defined
- Lack of standardization
Overlooked, Undervalued, Alternative explanations –
- Is there a control group in the experiment?
- Cherry-picking – think ‘windowing’ what is outside what is being presented to you?
- How big is the sample size?
Logical fallacies –
4 pitfalls:
- Illusory correlation
- Belief perseverance
- Persuasion by association
- False causality
Identifying Expertise –
Expertise is a social judgement – we compare one person’s skill to the rest of the world
Expertise is relevant
Axis Shenanigans –
Our brains did not evolve to mentally crunch numbers
So we look for patterns in visually displayed data