HP RPN SCIENTIFIC WP 34S Owner's Manual page 50

1. Let your WP 34S compute
their normalized distance
stack levels still, this calculation could look like the following:




Also provide the degrees of freedom for the next two steps:




2. Let your WP 34S calculate the critical limit
freedom and a probability of 97.5% now:

.975


d 
t
If then the test indicates the difference between both samples being due to
cr
random deviations only. Congratulations – you have got a robust process regarding
the parameters you changed.
3. Let your WP 34S compute a new critical limit

.995


d 
t
If then the test indicates a significant difference between both samples.
cs
Congratulations – your parameter change caused an effect.
 
t d
For
cr
your samples may contain too little data or your measurements were not precise
enough or the process is scattering too far – so you better stay silent or mumble
something like "investigation in progress".
WP 34S Owner's Manual
↕ SERR

↕ nΣ

0. 9 75
↕ t⁸(p)


0. 9 95
↕ t⁸(p)


, however, you cannot decide based on the information provided –
t
cs
and the standard error

x x

1 2
d
. Assume you are working with four

2 2
s s
E 1 E 2
returns both standard errors in
this is the complete denominator now.
and this is
recall the number of specimens measured.
calculate the degrees of freedom
and store them where they belong for the
next step.
t
cr

as mentioned above, the requested qf lives
in catalog PROB.
executes this function, taking the degrees of
freedom stored in
t
for
cs

note this function shows up immediately
when opening the catalog again.
t  
get .
cs
Edition 3.1
s for both samples, then
E

d
.
of Student's t for
J
to get
f and 99.5%:

X and Y .


f degrees of
t
.
cr
Page 50 of 211