Confidence intervals

Question 1

What are the 2 things central limit theorem leads to?

Answer 1

For an infinite number of samples of size N:

1. mean of the sample means =
   population mean
2. s.d. of sample means =
   s.d. population / √ N

Question 2

What is another name for standard deviation of sample means?

Answer 2

Standard error the mean (sem)

Question 3

What is SEM?

Answer 3

Indicates the likely variability in
experimental estimates

Question 4

If a population variable is normally distributed…

Answer 4

…then the sample means is normally distributed

Question 5

For a normal distribution…

Answer 5

…the proportion of
values of a variable within any range can
be calculated using the mean and the
standard deviation

Question 6

Do we usually overestimate or underestimate the population standard deviation?

Answer 6

Underestimate

Question 7

Why is a t-distribution used?

Answer 7

To compensate for underestimating the population s.d.

Question 8

What degrees of freedom do we use when calculating confidence intervals CI?

Answer 8

N-1

Question 9

How do we calculate the 95% confidence interval?

Answer 9

CI = T-score x SEM either side of sample mean

Question 10

What does the 95% confidence interval mean?

Answer 10

If we repeated our experiment and calculated the 95% confidence
interval for each, 95% of CIs would contain the population mean

Question 11

What is the different between a T-table and a Z-table?

Answer 11

T-table for samples, Z-tables for populations
Z, as in overall

Question 12

As the sample size increases, how does the t-score compare to the z-score?

Answer 12

It approaches the value of Z