[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Date Index][Thread Index][Author Index]

Final footnote? Oberheim Echoplex Digital Pro -- Market Demand & Availability Study

With minor equivocations, I think Sean is correct on ALL counts.

It is important to stress this sample size can only lead to _overall_
directions and trends.  That said I sense the directions and trends are
generally correct - the scale is what is arguable.

Re: the xx.x% vs. xx%  -- it's an old habit from working w/bigger
samples sizes - I didn't mean to imply more statistical authority.  It's
actually quite comical given the sample size.

I will forward Sean's critique to the folks at Gibson/Oberheim too.

As a footnote to this whole Polling excersize - it has been the
strangest fun.  I thank all of you who responded.  The results were
quite interesting -- some even surprising.   And to the other EDP owners
in Boston/Allston, I would like to meet you! (how about a show at the
Middle East; I know who to call).

Putting the calculator away...

David Kirkdorffer

        -----Original Message-----
        From:   buzzard@world.std.com [SMTP:buzzard@world.std.com]
        Sent:   Thursday, September 18, 1997 1:56 PM
        To:     Loopers-Delight@annihilist.com
        Subject:        Re:  Oberheim Echoplex Digital Pro -- Market
Demand &   Availability Stu dy

        Pretty comprehensive.  Nice job.

        I'm not a professional statistician, but I wanted to take some
        time to correct what I believe are some misleading statistics
        and interpretations.  No flame intended.

        >*      Almost 40% of EDP owners live within 80-90 miles of
        >And 54% of all the EDP owners also live between Los Angeles and
        >Berkeley, California. 

        This statistic is impossible to interpret without similar
        about the entire sample (e.g. if 54% of the entire sample lived
        LA and Berkeley, it would mean something quite different than is
        implied!)  The detailed results cover part of this (but no
        for LA-Berkeley).

        >*      Seventy-six percent (76.4%) of those who have tried an
        >indicate they also own an EDP.  This can be interpreted to
testify that
        >three-quarters of those who try an EDP go on to buy an EDP.

        This really can't be judged with much certainly without more
        data.  For example, anyone who buys one without trying it first
        ends up in this category.  (That such people exist seems likely,
        since, for example, the apparent paradox observed in another
        set of questions--don't want to try, but want to buy--can be
        explained by people with this intention.)

        >                                             This may mean that
        >contacting 1000 existing EDP owners (from user supplied
warranty card
        >information) via direct mail with an offer to buy another EDP
        >instantly yield orders for sales of 462 EDP additional units.

        I know you said "may" and "could", but I want to stress that
        the obvious general danger here is generalizing from the sample
        of "people who subscribe to loopers-delight" as being reflective
        of all consumers.  It's certainly _possible_ that all EDP owners
        are hardcore loopists like people on this list (what else are
        they doing with their EDP, after all), and certainly that's more
        likely the case than Digitech delay owners.  But even suggesting
        it translates directly into sales is questionable.

        Also, I wasn't going to tweak you for reporting results with the
        extra decimal place, but, really, 462?  Assuming just +-1
        respondent (which is probably an underestimate of the error),
        the original statistic (6/13, I assume) goes to 6/14..7/14,
        which is 42-50%, or 46 +- 4%, or 460+-40 units in your example.
        Reporting the extra decimal places just implies more accuracy
        than is really there, adding a false authority to the numbers
        which, if the reader knows better, makes them more suspicious
        of the results.

        >*      However, on the whole, the number one reason for wanting
to buy
        >an EDP is it's perceived as the "Best Overall Looper" cited by
34.2% --
        >which comes from combining the answers "Better than JamMan"
(17.1%) and

        The members of loopers-delight have access to a very current,
        up-to-date information source on the merits of various looping
        (namely the mailing list itself); perhaps this merely implies a
        of consumer education Gibson needs to engage in to increase
        But I certainly wouldn't expect people without access to this
        to have the same perceptions as things stand.

        Anyway, these are all just minor interpretation issues.  I
        just don't think statistics should be presented in the "best
        possible light to prove a point"--which is why I'll never
        play well with marketing departments...

        And, off topic, I strongly recommend a read of the old classic
        "How to Lie with Statistics" to anyone interested in the subject
        of the various ways statistics can be misleading.

        Like I said, overall, good work.  Thanks.

        Sean Barrett