A Sample of One
by Phil Bartle, PhD
Live with the people
The author of a text book, Henslin, wanted to make observations about homeless people.
His method was participant observation.
He wanted to live with them and to experience what they experienced, and to find out what they themselves thought about the meanings of what they were doing.
He could not live with more than one homeless group at a time, so his sample consisted of only one community among the thousands that exist throughout North America and the rest of the world.
If we knew for sure that every homeless community was radically different from every other one, then the validity of his findings would be very low.
If we believe, as we do, that there are many similarities between those communities, his validity would be very high, much higher, for example, than if he went around with a clipboard asking questions from a questionnaire.
(See that the method and the sampling technique are not as unrelated as one might first imagine both in combination contribute to the level of validity).
His text book gives an example of a population being all the students of a university, and a sample being those interviewed with a questionnaire.
How do you choose that sample?
Do you ask every fifth person walking down a particular hallway?
Will that bias your sample towards students taking those subjects taught in the rooms in that hallway?
Would that bias affect the results?
If you were asking questions where there might be significantly different answers from students taking one subject, rather than another, then that method of sampling would bias the result.
He suggests that a random sample would be most valid.
So how do you choose a sample that is random?
Here you might exercise your imagination, initiative and creativity.
How much a proportion of the population should the sample be so that it is valid?
If you interview only ten people out of a population of several thousand, perhaps the results would not be as valid as it might be if you interviewed a sample of one thousand.
My PhD research was based mainly on participant observation. See Why Obo? My selection of a sample, one community, was far from random.
If you copy text from this site, please acknowledge the author(s)