Yes!

The short answer is No. See Cochran Sampling Techniques available in the z-library. Best wishes, David Booth

I don't have sufficient information to give you a profound answer but I try to answer asuming some things.

It is ok to limit (to select) your sampling population on a group with specific properties like "place" - where the sampling units are located, if your question is likewise limited to the same property (place). For example you can limit your survey location to a supermarket - where you randomly talk to people entering the building about their preferences of soft drinks. It all depends on the scope of the question being answered. In this case your scope of research could answer the question about this specific shop and about the specific time and day.

To generalize the findings of your research you will need to randomize the sample for those properties.

Hi Janaki

It is possible to do the sampling this way. Sometimes our population is limited to some geographic area, or even it has temporal limitation. This type of sampling is called stratified or clustered sampling. The difference is that we should select the strata to reduce within strata variance to the lowest level and between strata to maximum. Clustered sampling is the opposite.

One simple example is when someone intends to do a study in a specific country/province, even a city, or even a single hospital. In such cases, selecting only specific hospitals should be supported by some hypothesizes/questions.

Hope it helps

Hani

Probability estimates can only be applied within places you have selected and should not be applied to any other places not randomly selected. If you want to include other places, then you might use a two stage sampling fame with random sampling of places at one level and random sampling at the next level. Studies have shown that simple random sampling generally gives better results than systematic random sampling. For example systematic random sampling utterly failed to predict the outcome of the Dewy/Truman election. See https://medium.com/@ODSC/dewey-defeats-truman-how-sampling-bias-can-ruin-your-model-f4f67989709e

No, it doesn’t sound right. It is advisable to either use probability sampling method or non-probability sampling method and not the mixture of both. This is because probability sampling guarantees every subject in the population an equal chance of getting selected while non-probability sampling does not. There is need to justify the reason for using a particular sampling method.

I will advise you stick to one of these methods depending on the circumstances surrounding your population of interest. Also, read more on the various sampling techniques to understand which of them perfectly fits your study.

Hope this helps.

Peter.

Both probability and non-probability sampling techniques cannot be employed at a time.