On Fri, 21 Jan 2000, Grant wrote:
> I have a client who would like to determine the demand for a very
> high-tech product and would like the results to be rigorous as they
> will be used to convince investors of the need for such a product.
> The problem is that the sample size is only about 25.
25 what or whom? What population are they supposed to represent, and
how were they selected? Or are they yet to be selected? (in which case,
why so few as 25?) How many customers does your client need, per year,
in order to stay in business? Quite frankly, _I_ wouldn't invest in an
enterprise that based its market research on 25 persons, unless there
were other, compelling, reasons for investing; in which case one hardly
needs the survey of 25.
> Besides frequency counts and percentages how else can one analyse the
> results to show that there is a demand for the product?
Presumably by "to show that" you mean "to show whether". You'd better
look really carefully at characteristics of the individuals in the
sample. Why should one believe that these 25 are representative of a
population of potential customers?
> Any suggestions on important questions that need to be asked will
> also be welcome. I have read that a simple yes/no scale is better
> than a 5 point likelihood scale for demand type questions.
Any such judgement is (or should be) based on research somebody
undertook once upon a time. Before accepting the judgement as gospel,
find the original research, and see whether the conditions under which
that was carried out bear any resemblance to the conditions of your
client's universe of discourse. Be sceptical.
> Any thoughts?
Ask price questions. How much would it be worth to the respondent if
such-and-such a service (or a device with certain performance
capabilities) were available? Nobody can really answer questions like
that reliably, since the device or service is not in fact available; but
whatever responses you get will at least indicate whether the respondents
and your client might eventually get to play in the same ball-park.
Might be worth asking such questions under more than one condition:
e.g., (1) if the respondent's competitors did not have access to such a
service or device; (2) if all one's competitors were supplied with them.
But really, try to persuade your client to deal with a larger sample.
While it is not at all clear what your client means by wanting "results
to be rigorous", one suspects that "sample results close to population
results" lurks in the back of his (her?) mind: and _that_ kind of rigor
is, to speak plainly, simply not available with a sample of 25.
(The 95% margin of error around any proportion derived from a sample of
25 is roughly twice the square root of (0.25/25), or about 0.2. That is,
if 16 of the 25 look like they'd buy the thing, that's 64% plus or minus
20%, or between 44% and 84% in a population of which the sample may be
considered representative. Is that likely to be persuasive to potential
investors? (I don't know, but that's the sort of result your client will
have to deal with.))
------------------------------------------------------------------------
Donald F. Burrill [EMAIL PROTECTED]
348 Hyde Hall, Plymouth State College, [EMAIL PROTECTED]
MSC #29, Plymouth, NH 03264 603-535-2597
184 Nashua Road, Bedford, NH 03110 603-471-7128
