Should it be representative for the consumption of peppersauce or for the whole range of different kinds of peppersauce? That answer is important in designing your sampling scheme.

To answer your question @C.A. (Kees) Kan, only one kind of pepper sauce will be assessed in this study and that is red pepper sauce made grinding raw tomatoes and dry pepper together.

What i mean by a representative sample is by choosing precise locations (possibly towns or markets) that are evenly spread within the Accra Metropolitan of Ghana from which to buy the samples, in order to have proper coverage of the region under study.

Therefore, the sample should be representative for the consumption of the red pepper sauce by the inhabitants of the Accra Metropolitan.

Dear Ebenezer

Most likely the Accran Metropolitan Area can be divided in different areas according to e.g business or residential, different backgrounds of the population, preferred buying places super markets or street corner outlets.

The choice of distribution of your sampling also depends on the question(s) you want to answer about samples; is it price, composition, length between production and consumption, safety of the products (microbiological or chemical) etc etc.

The simplest way would be a map with the number of inhabitants per unit of land and then making 60 areas with a similar number of inhabitant; but that approach will almost certainly not give you the answer that you are looking for.

So spent some time defining your exact question and then a sampling plan should no be terribly hard to design.

And remember: sampling is always combined with uncertainties as you make assumptions in doing so!