Simple Random Samples

Picking people who 'look representative' seems reasonable, but your judgment carries hidden bias: you notice some kinds of people and overlook others. An SRS removes human judgment entirely. You let chance do the choosing, so every individual in the population has the same probability of landing in the sample, and every possible sample of that size is equally likely. That's the whole point: randomness, not your intuition, protects against bias.

To actually draw one, number every member of the population, then use random digits to pick that many unique numbers. Say you need a sample of 5 from 80 people, numbered 01 to 80. Read a random digit table two digits at a time. Any pair from 01 to 80 names a person; skip 00 and 81-99, and skip repeats, until you have 5 distinct numbers. Those 5 people are your sample. The same method scales down or up: from a list of 50 students, you'd read two-digit groups between 01 and 50, keep the first 4 valid distinct numbers you hit, and those are your sample.

This topic sits right after you've learned what a population and a sample are, and it feeds directly into why some sampling methods go wrong. An SRS is the baseline every other sampling method, like stratified or cluster sampling, gets compared to. Once you can recognize a genuine SRS, you're ready to spot the sneaky ways real studies fail to achieve one, like relying on volunteers or whoever is easiest to reach.