What problem does sample size solve in research? It answers a plain but difficult question: how many cases are enough to make a study useful, without making it wasteful.
Sample size is the number of people, records, objects, or observations chosen from a larger group. If the sample is too small, the study may miss the pattern it is trying to see. If the sample is too large, the study can become costly, slow, and hard to manage.
I think of sample size as a practical balance. The goal is not the biggest sample possible. The goal is an optimum sample, which means a size that gives useful precision at a reasonable cost.
That idea matters in both humanities and social research. A small archive study, a survey of readers, or a content analysis of texts all depend on the same basic question. The sample has to be large enough to support the claim being made.
The first step is to define the population. That is the full group the study is about. In a library setting, it might be all visitors in a month, all articles in a journal run, or all records in a collection.
The second step is to decide how much error can be tolerated. This is often called acceptable error or precision. A smaller error means the estimate needs to be closer to the true value, so the sample usually needs to be larger.
The third step is to choose a confidence level. Confidence tells how certain the researcher wants to be that the sample reflects the population. Higher confidence usually calls for a larger sample.
For a large or effectively endless population, a common approach uses a formula with three parts. It relies on the desired confidence level, the expected spread in the data, and the acceptable error. In simple terms, more variation and less tolerance for error both push the sample size up.
For a population that is known and limited, the same logic applies, but the size of the whole population is added into the calculation. That keeps the sample from growing too large when the total group is not vast. The point is to match the formula to the kind of population being studied.
There is also a different path when the study is about a proportion. A proportion is a share of a group, such as the percent of responses that fall into one category. Here, the estimate of that share, the margin of error, and the chosen confidence level work together to shape the sample size.
A small concrete example helps. Suppose a librarian wants to estimate the share of patrons who use a new digital catalog feature. If the librarian wants tighter accuracy and more confidence in the result, the sample has to be larger than it would be for a rough internal check. If the librarian can accept a wider margin of error, the sample can be smaller.
That is the whole logic in plain form. The sample size changes because research goals change. A study that needs exact estimates has different needs from a study that only needs a broad picture.
Researchers also make choices before the calculation begins. They have to decide whether the study is about counts, proportions, or measured values like test scores or reading times. Each kind of data puts different pressure on the sample size.
The method is only as sound as the assumptions behind it. If the expected spread in the data is guessed poorly, the sample estimate can be off. If the population is poorly defined, the size can look precise while the study itself remains weak.
That is why sample size is not a mechanical number. It is a judgment made with rules. The rules are about precision, confidence, variability, and cost.
In practice, the same question keeps returning: what size sample gives enough information to answer the research question well? A tiny sample may fail to show the pattern. A huge one may add burden without adding much value.
For librarians, digital researchers, and humanities scholars, this is a useful way to think about evidence. A database may hold millions of records, but a study rarely needs all of them. It needs a sample that fits the question and the limits of the project.
I have found that this is the point many beginners miss. Sample size is not only about statistics. It is also about restraint, clarity, and purpose. A study becomes stronger when its size is tied to what it is trying to show.
With this lesson, the reader can now explain what sample size means, see why it matters, and tell how population size, confidence, and error shape the number chosen. The next step in any real project is to define the population carefully and match the sample plan to the question, which is the kind of practical discipline The Source List exists to keep visible: one digital source worth knowing, one search tip, and one honest limitation.