Inferential statistics help us make generalizations or predictions about what we are measuring, based on a set of measurements. We use inferential statistics when we want to make inferences about a whole group of people or companies or events, even if we’ve only studied a smaller sample of them.
For example, when 6sense Research surveyed buyers to understand how long their buying cycles are, we produced descriptive statistics that showed they are 11 months long on average. We also produced inferential statistics that showed the relationship between deal size, the number of vendors compared, and buying cycle length. In this example, an increase in deal size of $100,000 is associated with, or predicts, one month added to the length of the buying cycle. Meanwhile, an increase in the number of vendors evaluated, from the average of four to five, can be expected to add two months to the buying cycle. Those are inferential statistics.
We often conduct hypothesis tests, which tell us whether there is a relationship between two variables. These statistical tests tell us whether a change in one variable is reliably associated with changes in one or more other variables.
The conclusions of hypothesis tests are often referred to as statistically significant or not.
