Who Answered the Survey?

01

An 80% beside the timetable

Seojin is arranging the next introductory session for a neighborhood board-game club. The club has 120 registered members this season. Seojin wants to know whether these members prefer a weekday morning or a Saturday evening. This is a fictional situation designed for reading a survey, not a report about a real club.

Seojin posts a survey link in an online announcement channel. Sixty current members belong to that channel; the other 60 do not. Assume that the link was not shared elsewhere and that each member could submit only one response. The survey opens on Tuesday morning and closes the following morning.

Of the 20 people who answer, 16 choose a weekday morning and four choose Saturday evening. Seojin calculates 16 divided by 20: 80%. The draft announcement says, “Most members want a weekday morning.” The response count is correct. Does “members” in that sentence describe the same group as the 20 respondents?

02

Following the route the link could take

The membership list contains 120 people, but only 60 of them appear on the channel list. The 60 outside the channel had no route to this link. Within the channel, 20 people respond. The record does not tell us whether the other 40 read the post, missed it, or decided not to answer.

What this survey establishes is that 16 of its 20 respondents selected a weekday morning. The preferences of 100 members remain unknown: the 60 outside the channel plus the 40 inside it who did not respond. We cannot fill in their answers as matching the respondents, or as opposing them.

Perhaps the channel includes many people who already attend weekday-morning sessions. Perhaps people who check announcements frequently simply answered sooner, regardless of their time preferences. The available information does not establish either explanation. Whether the response process is related to time preference remains a question to investigate.

03

How many answered, and how did they get there?

The whole group we want to learn about is the population. Here, it is this season’s 120 members. A sampling frame is a list or structure used to locate or select people for a study. In practice, the channel list of 60 defined the range Seojin could contact. The group that actually provided answers contains 20 people.

Before carrying some people’s answers over to the whole group, we can ask who was more likely to be included or left out. This is a question about representativeness. People who were outside the contact route from the start present a different issue from people within its reach who did not answer.

If respondents and nonrespondents differ systematically in ways related to the preference being studied, an estimate for the whole group may be biased. A low response count alone does not reveal the direction or size of that bias. This was not a randomly selected survey, so we also do not attach an invented margin of error to 80% and present it as a population estimate.

04

What remains before collecting more answers?

More answers in the same channel would reveal more about people within it. They would not create a contact route for the 60 members outside it. A growing response count and broader coverage of the group of interest are separate things to examine.

Using the current membership list to arrange other ways of making contact could give previously missed members an opportunity to participate. Adding channels would not, by itself, complete the work of establishing representativeness. There would still be a reason to record whom the invitation reached and who actually answered.

The question and the response period matter too. Choosing a preferred time from weekday morning and Saturday evening is different from reporting when someone can actually attend. If the question changes, the new answers should not be combined with the old ones as if everyone answered the same question. Expressing a preference is not a promise to attend either.

If Seojin wants to hear from people who have not yet joined the club, the population of interest changes. A list of the current 120 members cannot cover that new group by itself. Naming whose views are wanted helps clarify which sentence can properly contain the 80%.

05

Keeping a label on the people who answered

The same questions can be carried to craft-class reviews or library-program surveys. Who received the question? What methods and time were available for answering? Who actually responded? A large-looking collection of comments is different from evidence that the whole relevant group was reached.

A limited response group does not require throwing away the views already collected. The preferences those 20 people reported still matter. Before expanding them into a conclusion about all members, describing the respondents and the method lets other readers see the number’s limits.

Seojin underlines “most members” in the announcement and replaces it with “16 of the 20 people who answered this survey.” Weekday morning remains on the list of possibilities. Beside it, Seojin makes a small space for ways to contact members outside the channel and a period in which they could answer.

The two time slots remain beside each other on the timetable. Seojin opens the current membership list next to the empty space. Between the answers already heard and the ones still unknown, there is a little more room to write the next invitation.