Click to reveal answer

In everyday language, the term odds is used interchangeably with probability. But in statistics, there is a specific definition. Let’s imagine a sample of 100 adults in the UK. We find out that 92 of those 100 have had at least one dose of a covid-19 vaccine. The probability is straightforwardly: $\frac{92}{100}=0.92$

But, the odds of observing the outcome are:

So here:

We can interpret the odds to mean that being vaccinated is 11.5 times more likely than not being vaccinated. Note that probability and odds are two ways of expressing the same idea

Say we also have a sample of children aged 5-11. Among these, only 11 of 100 have been vaccinated. In this case the odds are;

Click to reveal answer

Exactly 1

So, we have seen that a probability >50% produces odds with a value > 1 and a probability <50% produces odds with a value < 1. And when a probability is exactly 50-50 the odds will be exactly 1. We’ll come back to this later when we interpret the output of our logistic regression models.

Click to reveal answer

A log (or logarithm) expresses the inverse of an exponential. If you did A-level maths, this should sound familiar, but if not, here’s quick overview. Exponentials refer to ‘power’ functions, such as $2^3$ (two to the power of three, also known as two cubed), which means $2\times 2\times 2$ (and equals 8).

$2^3$ can be expressed as:

Which says that the log base 2 of 8 is equal to 3.

More generally, when $a^b=c$ then ${\log }_{a}c=b$

Click to reveal answer

${\log }_{10}1000=3$

Click to reveal answer

A: The question we are trying to solve is: two to the power of what equals 16? The answer is 4 (because $2^4=16$).

Click to reveal answer

$x=0$ (because anything to the power of zero = 1).