Question

A researcher wanted to administer an in-person patient satisfaction survey to Californians who last visited either a primary care physician (PCP) or physician specialist. The researcher wanted to obtain a statewide representative sample of patients who were seen by a sufficient number of specialists (e.g., dermatologist) for comparison with PCPs. The researcher had a directory of physicians by specialty in California, but not a directory of patients. Which of the following stages (in order) should the researcher choose if he/she wants a feasible and economical way of surveying patients? Provide a brief explanation.

1. Stratify population/cluster by physician specialty (PCP vs. any specialty)
2. Take a cluster sample of California counties
3. Take a SRS of physicians using PPS sampling to account for differing patient loads
4. Take a SRS of patients seen by the sampled physicians

Order of Stages

1. a,c,d     
2. a,b,c,d  
3. b,c,d     
4. b,a,c,d   

Answer 1

The order of stages should be 'b , a, c , d'.

Explanation :- Here, the researcher wanted to obtain a statewide representative sample of patients.

Stage I - Firstly he will do cluster sampling in order to include all the states of California. Then we take a cluster sample from it.

Stage II - He wanted to do comparison between patients seen by a sufficient number of specialists and PCPs. For this, we stratify the cluster by the physician specialty i.e. PCP vs any other specialty.

Stage III - Now, he will choose physicians by taking a sample using Simple Random Sampling in order to account for differing patients loads.

Stage IV - He have sample of physicians. Now, we will choose patients to be surveyed of those sampled physicians by using Simple Random Sampling.

After this, he will have a sample of patients in order to get required results.