Statistical Investigations
What a statistical investigation is
A statistical investigation is a way of answering a question with data. It follows a cycle: pose a question, collect relevant data, choose a display, interpret what the display shows, and communicate the findings. Each stage feeds the next, and the answer is only as good as the question and the data behind it. Thinking of statistics as an investigation, rather than as graphs on their own, keeps the purpose in view: every display and every number is there to help answer the original question.
Posing a question and collecting data
An investigation begins with a clear question. A good statistical question is one that expects data that varies, such as how many pets each student has, rather than a question with a single fixed answer. Once the question is set, relevant data is collected, perhaps by a survey, a count or measurements. The data must actually bear on the question, and it helps to collect enough of it that the answer is reliable. A vague question, or data that does not fit it, weakens everything that follows, so this first stage matters most.
Choosing a display
Different kinds of data suit different displays. Data sorted into categories, such as favourite sports, is shown well by a column graph, where the height of each column is the count for that category. Data that changes over time, such as temperature through a day, suits a line graph, where the line shows the trend. Choosing a display that matches the data makes the patterns easy to see, while a poorly chosen display can hide them. The display is chosen to serve the question, not for decoration.
Interpreting the data
Once the data is displayed, it must be interpreted: read off, compared and summarised. From a column graph of favourite fruit, the tallest column shows the most popular choice, and the counts can be compared and added. Interpreting also means describing the overall picture, such as which category is largest, which is smallest, and how spread out the values are. The aim is to draw conclusions that the data actually supports, going beyond reading single values to seeing what the whole display says.
Communicating the findings
The final stage is to communicate the findings clearly, in the context of the original question. A good report states what was asked, what the data shows, and what conclusion follows, in plain language that someone who did not see the data could understand. The findings should answer the question that was posed, not wander off topic. Communicating well means connecting the numbers back to the real situation, so that the investigation ends where it began: with an answer to the question.
Carrying out an investigation
A statistical investigation joins these stages into one process: pose a clear question, collect data that fits it, choose a display that suits the data, interpret what the display shows, and communicate the findings in context. Each stage depends on the ones before, and keeping the original question in view holds the whole investigation together. With this cycle a child can plan and carry out an investigation from start to finish, ready for the larger data handling and statistics of later years.