RE: Cause source - system v. employee

["Mack, Wayne" <wayne.mack@pec.com>]

> From:	David Kerridge [SMTP:dfkerridge@mac.com]
> 
> At 7:57 am -0500 29/3/02, Mack, Wayne wrote:
> >If you attempt a change to a system and do not see a resulting data
> change
> >on a control chart, there is no evidence to support the change as
> effective.
> 
> That is often true, but under what circumstances do you need (or can
> you get) evidence of this kind?

In practice, I find

1) I often need to provide justification for a suggested change when it is
not
implicitly accepted by others, and

2) I often request justification for a suggested change when I am not
implicitly certain of its value.

I often use control charts to provide such justification.  At least in the
area
of software development, I find too many changes supported only by
opinion and speculation and spend far too much time resolving disagreements
based on differing opinions.

Also, recall that for the scientific definitions, the difference between a
hypothesis and a theory is supporting experimental evidence.  I am concerned
where pushing my own hypotheses crosses over into meddling (ala the
funnel experiment).  Short of experimental evidence, how do I prove, to
myself and to others, that a change should be made and it is beneficial?

Wayne Mack