
In the
MATH
menu,
Probability
section (see page 208), there is a
function called
UTPC
(Upper-Tailed Probability Chi-squared) which will
give the critical
X
2
probability for a supplied number of degrees of
freedom and a value. In this case we would like the value for a given
probability so we will enter the formula into the Solve aplet.
Change to the
NUM
view, enter the known parameters of
D=4
and
P=0.05
, and
for the critical
X
2
value.
Since our value of 16.387 is larger than the critical value of 9.488 we conclude that we must reject the null
hypothesis and judge that the observed values do not follow this binomial distribution and hence that the coin
is probably biased. This does not say that it is not binomially distributed; just not with those parameters.
Importing from a frequency table
The import (
) facility of the Inference aplet has a small weakness in that it can’t import from paired
columns defining a frequency table.
For example, suppose we use columns
C1
and
C2
to define a frequency
table, ensuring that it is registered in the
SYMB
view as shown right.
If you now change to the
NUM SETUP
view of the Inference aplet and
try to use the import facility you will find that you can import the mean
and standard deviation from either column
C1
or
C2
but not from both
at once.
What’s needed, of course, is a way to expand the paired frequency
table columns into a single column listing all values. You can, of course,
create the expanded columns by hand but the program on the next page
will do this for you.
148