I'm not sure about this, but I will give you my opinion since you requested that I answer it:
Part A:
The survey is designed to study people who wear corrective lenses in the United States. The group of people at one local grocery store does not give a statistically significant representation of the demographics of the entire nation. While it's likely that the differences in people with corrective lenses are not majorly impacted by demographic differences, this is not known definitively. Additionally, for a standard 5% confidence interval, the sample size would probably not be sufficiently large.
Part B:
This represents the population from only the two largest cities and does not include other regions and subsets of the population. Therefore, it is not a random population, and no statistically significant conclusions can be drawn from this data.
Part C:
The method of selection for this survey is nearly randomized, so yes, this is basically a random sample; however, the sample size is far too small to make any conclusions.
