Here are two questions. How tall am I? and How tall is everyone in our class?

One of these is worth surveying the whole class about. The other you could answer with a single measurement. Which is which, and what makes a question worth asking lots of people?

Let's sort some questions together. A statistical question is one that expects lots of different answers when you ask a group of people. A question with just one fixed answer is not worth surveying.

• What is your shoe size?
• What colour is the sky?
• How many minutes do you spend on homework each night?
• What is 7 times 8?

Let's try this in your groups. Draft one statistical question you could survey the class with, then agree what kind of answers you'd expect to get back, and whether they'd really vary.

Now let's run our agreed class question together. Our class question is the one we agreed on at the end of the discussion, written at the top of the board now.

Each of you writes your answer on a response card. As the cards come up, the recorder reads them out and writes each one into a list on the board, in the order they are read. This complete, unsorted list is our raw data: every answer written exactly as given, before any sorting or counting.

Today we learned

• A statistical question expects lots of different answers across a group, not one fixed fact.
• A good question can be reworded to give a wider spread of answers.
• Raw data is the full list of responses, recorded exactly as people give them.

Coming up

Next we take a raw list like the one we built today and start to organise it — sorting wide-spread values into equal groups so the mess becomes a tidy table.