Stats midterm

Question 1

What is the sampling error?

Answer 1

difference between a statistic and a parameter due to chance

Question 2

what does the law of large numbers tell us?

Answer 2

the sample mean will approximate the population mean as n increases

Question 3

when can we compare means when they’re not normally distributed?

Answer 3

both n > 30

Question 4

what is needed to use the f-distribution?

Answer 4

populations need to be normally distributed

Question 5

how can we compute the relative efficiency?

Answer 5

var(x)/var(y)

Question 6

how can we compute the mean squared error?

Answer 6

var(x) + bias (x)^2

Question 7

which values can the t-distribution have?

Answer 7

any finite value

Question 8

what happens to the t-distribution when the degrees of freedom increase?

Answer 8

it tends towards the standard normal distribution

Question 9

is the t-distribution symmetric?

Answer 9

yes

Question 10

when are samples considered dependent?

Answer 10

matched pairs (members resemble too much each other) and repeated measurements (the same individual is counted twice)

Question 11

True or false: The chi-square distribution is the sum of any squared independent random variables.

Answer 11

false

Question 12

True or false: The chi-square distribution is a positively skewed distribution.

Answer 12

true

Question 13

True or false: The only way to find out the sampling distribution of a given statistic is to take all possible samples of a given size and for each sample compute the relevant statistic.

Answer 13

false

Question 14

how to calculate the MSE?

Answer 14

variance + bias^2

Question 15

what are the conditions for simple random sampling?

Answer 15

1) Every member has the same probability of being selected
2) Selection is independent
3) All possible samples of size n should be possible

Question 16

why do we set p=0,5 if we want to be conservative?

Answer 16

Setting p = 0.5 maximizes the value of p(1 − p) thus resulting in a value
of n that will for sure be large enough