Chapter 13

Question 1

Resampling

aka bootstrapping

Answer 1

(aka bootstrapping) is
an alternative approach
that does not assume
any specific type of
distribution.
• By creation of
pseudosamples to
compare with the
original sample.

Question 2

The pseudosamples

Answer 2

are generated by resampling
from within the original set of observations.
Really?

Question 3

The standard error of the

mean (aka standard error)

Answer 3

does not directly quantify
scatter or variability
• common
misinterpretation
• because SEM decreases as
sample size increases (SD
does not).

Question 4

SEM

Answer 4

= SD/ √n

Question 5

With huge samples the SEM is always tiny even if

Answer 5

the SD is large. The SD does not necessarily

change with sample size.

Question 6

The SEM quantifies

Answer 6

how precisely you know the
true population mean. It is proportional to the
margin of error of a C.I. (W).

Question 7

W =

Answer 7

SEM * tcrit

Question 8

SD =

Answer 8

SEM * √n

Question 9

• If you want to show
variation use SD, not SEM.
• However,

Answer 9

other images like
the box-and-whiskers plot
are still more informative
that just a mean and SD.

Question 10

If you want to show how
precisely you can estimate
the mean, then you could
use

Answer 10

the SEM, not SD.
• But a CI would be more
informative.

Question 11

The use of SD and SEM in graphs and tables has become

Answer 11

the
conventional choice even those these are often not the most
informative alternatives.