Hi, Kashif,

The logic of random selection requires a rigorous sampling frame. The sampling frame not only defines the population to which the finding from your sample can be inferred, but it also the subjects/units from which you randomly select. Without a rigorous sampling frame, you will not have a valid argument for external validity of your findings.

Peter A Kindle in some case we know the exact population for example telecom subscribers of any cellular company. But we dont have access to the data to pick the respondents ? What is your advice in this situation? we can get the exact sample, and symptomatically approach the respondents. Thank you

Kashif Farhat Thank you Sir, For asking this question. I am also looking for its answer. and also thank you for sharing the link. I hope experts like Peter A Kindle can advise us here. Thank you!

We are unable to use any probability sampling techniques (eg, systematic/stratified/cluster, etc) if we do not have a sampling frame.

Nguyen Tho What do you mean by sample frame?

1-List of people in any population? by list I mean, if someone want to measure the satisfaction of particular community satisfaction regarding a product, he should have everyone's name on a list ?

or

2-Exact count of people in any population? to extract the required sample

A list of all elements (eg, people) in the population. For example, if we want to study customers (of a product) in a city (a nation) we should have a detail list (eg, names, addresses, etc) of all customers in that city (nation). Note that we normally do not have a precise list of all elements in a population (due to many reasons, eg, the list is dated, errors occurred when data were collected...) we often use an available list (sometimes called a study population).

Nguyen Tho Thank you so much for you response. I would love to have reference of this that "We use available data". Just to make it very clear, I am studying telecom companies. I have list of some subscribers (Out 150 Million I have 38 Million subscribers list), So you are saying I can extract numbers from available list and can call it the systematic random sampling. Thank you again for your time and response.

What I mean "available" is available in the market. In your case, the population is 150 M (not 38 M, ie, you have 38 M, not what is available (I am not sure) in the market). Further, this is a sampling frame which can be used for any probability sampling techniques, not the "systematic random sampling" which is a probability sampling technique.

Hi, Muhammad,

A probability sample of the 150M would be possible if you could obtain the lowest and highest possible telephone numbers used. Simply use a random number generator to select the numbers for your sample.