Conceptual Statistics. Define for this problem, population, Samples and Estimators, and when is Normal Dist? Students in Stanford are supposed to spend on average 3 hours of time per week for
every credit hour they take. Last year, 263 randomly selected seniors were contacted and asked how much total time they spent on their studies over the last week and how many credit hours they currently take.

(i) In this study, what is the population (be precise) and what is the
  sample?

      Population: (Senior) Students in Stanford. 
                  (My proff. told me this was, I needed to add seniors, in 
                  population) 

      Sample: 263 Seniors in Stanord


(ii) What is the population parameter of interest? How could one
  compute a sample estimate of this population parameter?

      Population parameter of interest: average total time per student    
      Sample estimate of the population parameter: 263 times the mean of the sample 


(iii) Under which circumstances would it be reasonable to assume that
  the average time spent per credit hour computed from the responses is
  normally distributed?

Concerns: I have no idea if my answers for i and ii are correct, and I need help doing the last question. If someone can help me to understand this it will be appreciated, thanks a lot!
EDITED:  My profesor said that point ii and iii , there is somemting missing(wrong). I can't figure out what it is, I know that the distribution of hours can't not be normal because are all greather or equal than 0. Also I asked point ii and iii, following the thouth of Michael Hardy and Probablemy, but he said is not correct" I'm still working on an explanation. If I found it I will posted" Thanks for all the aswers.
 A: Revisions: 2/9
If I had to answer this, I would say:

i. The problem states that you took a sample of seniors, then population is all seniors. The sample is the $263$ seniors polled.

ii. The population parameter of interest is the average number of hours spent studying per week for every credit hour the seniors take. To compute a sample estimate of this population parameter, find average number of hours spent studying per week for every credit hour using the sample data.

iii. Under the assumption that the number of units taken are iid and the number of hours spent study are also iid, then with a large sample size, the estimate should be approximately normal.
A: I believe you are overthinking this assignment, and perhaps you feel the need to do some calculations before giving any kind of answer.  You should not approach this task in that manner, as I hope my answers to your questions will illustrate.
 (i) is correct as long as you have "seniors" in your answer.  They are the people contacted for responses to the question, not juniors, sophs, or freshmen, or even janitors, or professors.
  (ii)  You answered "Population parameter of interest: average total time per student. "  Here is where you left something out.  Nobody cares about the average total time per student.  What everybody wants to know from this assignment is how much study time was spent by each student per each of his credit units !   Don't forget the credit unit part. 
     The second part of your answer was " Sample estimate of the population parameter: 263 times the mean of the sample. "   Well, nobody cares about the total time used by the 263 seniors, it is the average time that is sought as detailed above.  I might suggest you think about these kind of problems and what they might be used for.  For example,  if another university, lets say Harvard, wants to know how many credit hours it should allow it's students to take per semester, they can use your class task as one estimate for the limit.  Their classes are likely to be different from 263, so 263 times avg time means nothing to them.
(iii)  You should begin to realize that if samples are randomly chosen (no selecting the best GPA's, or worst GPA's or seniors with heavy loads or light loads), just seniors who could have been picked by anybody who knows nothing about them except they are seniors at Stanford.  With the random selection process, you can expect the distribution to be normal, the famous bell shaped curve. I hope these comments answer your questions and help you achieve your degree.    Williamo
A: Your answer to i is correct.
The answer to ii cannot be ascertained from the information in the first paragraph. A number of different quantities pertaining to the whole population could be of interest.  One could be interested in the average number of hours per week per credit-hour.  One could also be interested in testing a null hypothesis that the amount spent per week does not depend on the number of credit hours.  Or one could be interested in whether the dependence is simply linear or affine, i.e. could it be that time spent studying is some base amount plus an additional amount proportional to the number of credit hours?  Or one might be interested in how much variability there is in the number of credit-hours taken.  Or how much variability there is in the ratio of study time to credit-hours.  Possibly one could discover, based on the data, that there are two distinct groups of students: those who study three hours per week per credit-hour and those who study 10 hours per week regardless of the number of credit-hours.
iii $\text{“}$Under which circumstances would it be reasonable to assume that the average time spent per credit hour computed from the responses is normally distributed?$\text{”}$  Here we should look closely at the way in which this question is phrased.  Notice that it says "computed from the responses".  If you take a different random sample of 263 students, you get a different average, and yet another gives another average, and so one.  It's random only because the sampling is random.
Here my first thought was that the central limit theorem implies that this random variable is normally distributed, even if the sample clearly shows that the underlying distribution of study hours per week per credit-hour is not normally distributed.  However, let's look at applicability of the central limit theorem.  Let $x_1,\ldots,x_{263}$ be the reports of hours per week, and let $y_1,\ldots,y_{263}$ be the numbers of credit-hours.  The central limit theorem tells use that the sample average
$$
\frac{x_1+\cdots+x_{263}}{263}
$$
is normally distributed, and so is
$$
\frac{y_1+\cdots+y_{263}}{263}.
$$
But the question is asking about
$$
\frac{x_1+\cdots+x_{263}}{y_1+\cdots+y_{263}}. \tag 1
$$
Certainly if $y_1=\cdots=y_{263}$, i.e. the number of credit-hours is the same for all $263$ students, then the central limit theorem is applicable, since it's just a constant times $x_1+\cdots+x_{263}$.  But at this moment I am unsure how to deal with $(1)$, simply because there are some limit theorems on which I am quite rusty.
