Vc. Model Control beyond Outfit and Infit

Diagnosis and the residuals

An item’s global statistics (those with just the i subscript) are useful for flagging items, but any diagnosis or understanding means looking at finer detail, i.e., things that require two or more subscripts. Beginning with the items with the most disturbing values, examine them from every angle and keep reviewing, revising, and rejecting until you are chasing noise, the item pool is depleted, or the client runs out of money.

The choice of a favorite strategy may be based on unpleasant experiences from your youth or it may be based on deeply held philosophical principles. My deeply held philosophical principles are:

  • Some notion of what might go wrong is more important than how you do the arithmetic.
  • Contemplation of raw observations with an empty mind, even when it is possible, is often hardly more beneficial than not studying them at all. (Martin Wilk)
  • If your data have something to tell you, your statistics won’t stop them. (G. Box)
  • No single fit statistic is either necessary or sufficient. (David Andrich).

Continue . . . Control after Infit, Outfit

 

Previous: ala Panchapakesan                         Next: Diagnosis

Advertisements

2 thoughts on “Vc. Model Control beyond Outfit and Infit

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s