sdaza/survey-methods

Homework 1

sdaza opened this issue · 7 comments

sdaza commented

Here you can post all your questions on assignment 1. Questions I got via email:

  1. I am not sure if Frame Error is the same as Coverage Error.

The coverage error may be caused by sampling frame error or other factors, such as excluding certain population groups for convenience.

  1. When calculating the MOE for the entire sample, which value is "n"? Is it the total number of participants in the sample, which is 3642, or the 300 that have been selected? I have attached the calculations I have made.

Sample is just a fraction of the population. MOE is related to a sample, not the population.

EDIT: I think I actually found the answer I was looking for with the proportions definition and by using binomial distributions. So never mind!

About question 2:

I'm confused as to how we get the standard error, the standard error is only related to the sample, right? And can we say that 174 is the mean of the sample in this case? Or if not how do we find the mean and standard deviation?

Thanks,

sdaza commented

EDIT: I think I actually found the answer I was looking for with the proportions definition and by using binomial distributions. So never mind!

Good!

I'm confused as to how we get the standard error, the standard error is only related to the sample, right? And can we say that 174 is the mean of the sample in this case? Or if not how do we find the mean and standard deviation?

Standard error is the standard deviation of the sampling distribution (so, it relates only to samples). Use the corresponding formula to the problem at hand. There are examples in the notebooks and slides.

Hi! For question 3-D: MOE only for group C, should we compose the MOE completely by itself like this: serr(61, N = 343, p = 0.51) (note that the the n for this strata is computed with w=0.3 to make the group larger in relation to the other groups) or in relation to the other strata serr(300, N=N, p=prop_eng) (note that N is the total population and p = proportion of engagement). This function gives the MOE for all for all 4 strata and hence we can read the MOE for just C as well.

sdaza commented

Hi! For question 3-D: MOE only for group C, should we compose the MOE completely by itself like this: serr(61, N = 343, p = 0.51) (note that the the n for this strata is computed with w=0.3 to make the group larger in relation to the other groups) or in relation to the other strata serr(300, N=N, p=prop_eng) (note that N is the total population and p = proportion of engagement). This function gives the MOE for all for all 4 strata and hence we can read the MOE for just C as well.

To calculate the margin of error (MOE) for C, you can use group C as the sample. To estimate the MOE for all strata, you will need to apply the stratification formula (which we will discuss in the next class).

Thank you!

Hello! I was trying to use the package of sampler, but I am still getting the following error: Skipping install of 'sampler' from a github remote, the SHA1 (99b279a0) has not changed since last install. Use force = TRUE to force installation. How can I solve this?

sdaza commented

Hello! I was trying to use the package of sampler, but I am still getting the following error: Skipping install of 'sampler' from a github remote, the SHA1 (99b279a0) has not changed since last install. Use force = TRUE to force installation. How can I solve this?

It seems that the package is already installed. Could you please try running library(sampler) and then execute ssize(.05)?