From question to data: what, where and how to collect — 11–13
A statistical question has answers that vary; before collecting, you decide what data you need, where to get them, whom to ask and whether people answer in public or in secret.
A question whose answer changes from person to person
A statistical question is one whose answer is not always the same: “How many brothers and sisters does each student in your class have?” gets different answers, so you need data. “How many students are in your class?” has one answer, so it is not statistical. When the answers are names or categories (favourite sport) the data are qualitative; when you count and get 0, 1, 2, 3… (number of brothers and sisters) they are quantitative discrete.
Decide before you collect
Data collected without a plan are hard to use, so decide first what you want to find out, where to get the data and whom to ask. A primary source means you collect the data yourself; a secondary source means someone already did it, such as a table on a credible website. Then choose how: observe (count what happens) or ask (a spoken interview or a written questionnaire). Ask the whole group the question is about, not only your friends.
A curiosity becomes a plan
Curiosity: “Which sports do my classmates like?” Step 1: turn it into a question everyone can answer: “Which sport do you like best?” Step 2: qualitative data (names of sports). Step 3: whom to ask? All 20 students of the class, so you expect 20 answers. Step 4: where? Primary source, because nobody else asked your class. Step 5: how? Written and without names. Another question, “How many books did you read this month?”, gives 0, 1, 2, 3…: quantitative discrete data.
Answering out loud is not answering in secret
Imagine 20 students asked out loud, with hands up: “Do you brush your teeth before bed?” 18 raise their hands, that is 90%. With anonymous slips on the same day, only 13 write “yes”, that is 65%. The gap is not chance: people change answers out of shame or to look good. So the method changes the conclusion. The same happens with a poll where only those who want to answer take part: it tells you about them, not the whole group.
Polls, ratings and school votes
Opinion polls, app ratings, votes for class representative and surveys about how students get to school all start with a question, a group and a method. When you read a poll, ask: who was asked, how, and did they answer in public or in secret? Someone who thinks about these three things before believing a number is already thinking like a statistician.
Keep exploring
Other languages
Loading MyLeoNes™…