Discrimination test action standard

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Discrimination test action standard

Thien Hop
Dear List members,

I have a question that is not directly about SPSS but somehow relates to it.
All advices are really much appreciated.

It is about action standard for discrimination test (new formula vs current
one). Someone says the correct discrimination should not be over 35% of
sample size (otherwise the new formula must be revised), the others allow it
up to 75% (if more than 75% of the sample size can correctly discriminate,
new formula cannot be launched).

==> Which standard is more applicable and what is the statistic reason for
such decision?

Could you please give me some inputs, I really get stuck.

Thank you very much & look forward to hearing from you,

Hop

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD