What a small survey can say
For a language project, a group invented a small survey about places to read. The fictional dataset contained responses from twenty students: eight preferred a quiet library, seven chose a comfortable place at home and five preferred an outdoor space. The numbers were created for practice, not collected from real students. The group labelled this clearly above its diagram.
Their first draft said that most teenagers preferred libraries. Another student noticed two problems. Eight out of twenty was the largest single group, but it was not more than half. Moreover, an invented set of twenty responses could not support a claim about all teenagers. The sentence sounded confident while saying more than the data allowed.
The group replaced it with a precise comparison: in this practice dataset, the library was selected more often than either of the other individual options. They also calculated the corresponding percentages and checked that the three categories added to the total. The visual used written labels as well as different colours so that meaning did not depend on colour alone.
Next, they discussed possible explanations. A library might offer quiet and fewer interruptions, while home might be more convenient. These were reasonable ideas, but the dataset contained no answers about reasons. The explanations therefore had to be presented as possibilities, not findings of the survey.
The teacher then gave the group a different fictional set with two categories selected equally often. This helped them practise describing a tie without inventing a winner. A third example allowed participants to select more than one answer. In that case, the total number of selections could exceed the number of people. The students learned to read the question and the response rules before assuming that every set of percentages should add to one hundred.
They also examined two diagrams that used the same original numbers. One vertical axis began at zero, while another showed only a small part of the range. Both contained readable labels, but the second made the differences look much larger. The group discussed how a visual choice could influence an impression even when the printed values remained correct. Their own diagram used a clear scale and a short explanation of what each bar represented.
For the spoken presentation, each student had to choose one comparison and explain why it mattered. Reading every label aloud would take time without necessarily helping a listener. A useful description identified the largest category, noted that it was not a majority and explained one limitation. If a listener asked about a reason that the dataset did not contain, the presenter could say that a follow-up question was needed.
Finally, the students exchanged reports and checked one another's claims. They looked for unsupported words such as everyone, always and proves. The aim was not to remove every interesting idea. It was to distinguish an observed pattern inside the exercise from a possible explanation and a wider claim requiring additional evidence.
The final paragraph suggested a follow-up question about what made each place helpful. It also explained the limitations of the exercise. The project showed how to compare categories, describe a diagram and qualify an interpretation. It did not discover the preferences of a real population. Being clear about that boundary made the report more trustworthy and gave the students a stronger basis for future work.