Sampling: Probabilistic and Non-Probabilistic Techniques

We discovered in the last post that samples may be obtained in many ways.

The goal is to obtain a representative sample, though (as far as possible).

Based on the manner in which samples are obtained, statisticians may broadly classify sampling techniques as either “Probabilistic” or “Non-Probabilistic“.

What does “Probabilistic” mean? In the context of sampling, it simply means that each subject (or sampling unit) has a known probability (or likelihood) of selection. The probability (or likelihood) of selection must not be zero. 

Please note that knowing the likelihood (or probability) of selection does not mean we know who will be selected: If I know that the probability of selection is 0.5 (or 50%), I know that half of the population will be selected, but I do not know who will be in that 50%.

“Probabilistic” sampling techniques (often) involve the term “Random”.

What is meant by “Random”? In common understanding, random means arbitrary or ad-hoc. In statistics, however, “Random” means unpredictable or unguessable. (This automatically gives eligible members of a population similar chance of being selected.)

Examples:

Let’s assume that I need to choose 10 individuals from a crowd of 100 people. If I lined up everyone and then chose every 10th person, that is not “Random”, because anyone could guess who’d be selected next.

If, on the other hand, I assigned each person a number from 1 to 100; made 100 chits numbered from 1 to 100; put all the chits in a bin; shook the bin vigorously; and picked any 10 chits out of the bin, that is “Random”.

“Non-Probabilistic” Sampling techniques: Those in which one or both of the following may be true

i. The probability (or likelihood) of selection is not known

ii. The probability (or likelihood) of selection is zero for some members of the population.

How does one decide whether to use probabilistic or non-probabilistic sampling techniques?

If you intend to apply probability based statistical tests on the sample, you must use probabilistic sampling techniques.

Summary:

Sampling techniques may be broadly classified as Probabilistic or Non-Probabilistic techniques.

Probabilistic techniques usually involve the term “Random”.

Probabilistic means the probability (or likelihood) of selection is known, and is non-zero.

Non-Probabilistic techniques are usually employed when either the probability of selection is not known, or the probability of selection is zero for some members of the population.

About these ads
This entry was posted in Biostatistics and tagged , , , , . Bookmark the permalink.

2 Responses to Sampling: Probabilistic and Non-Probabilistic Techniques

  1. Pingback: Stratified Random Sampling | communitymedicine4asses

  2. Pingback: Best Performers In All Probabilistic Fields | Consilient Interest

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s