I'm afraid your use of hypothesis testing is
flawed. It is of crucial importance which hypothesis is used as the
null. You can't just flip a coin. The null hypothesis looks to find
just that ONE instance that disproves your hypothesis. The
alternative is to find EVERY instance that proves the hypothesis. A
formidable if not impossible task.
The alternative hypothesis will never actually be
tested. It is only inferred based on the results of testing the null
hypothesis. If you find the one instance, you disprove your
hypothesis. If you don't, well, it proves nothing.
The case, as presented, is really two hypotheses: VAN's
are less expensive and Internet is less expensive.
For the current discussion we are wondering if moving
to the internet is less expensive than the status quo of using VAN's to transact
business documents. The null hypothesis always states the opposite of what
you are trying to prove. In this case internet is less expensive.
Again, this is because it's easier to find one example that disproves the
hypothesis than the many required to prove the hypothesis. In my opinion
our test should say:
(Ho: is the standard notation for a null
hypothesis)
(H1: is the standard notation for the alternative
hypothesis)
Ho:The use of VAN's is the least expensive method of
transacting business documents electronically.
H1: The use of VAN's is the same or more expensive than
other methods of transacting business documents
electronically.
The independent (the one we can manipulate) variable is
methods used to transact business documents electronically.
The dependent variable (the one we would like to
measure) is cost.
Unless you want to delve into multi-variate statistics,
all other parameters must remain the same (see below).
Now all we have to do is find one method that is less
expensive than using a VAN. If we do we can reject the null hypothesis and
accept the alternative. Note: This does not prove the alternative, just
that we couldn't find statistical significance to reject it.
This may sound like a bunch of semantics to you but it
is vitally important if you don't want to be duped by statistics.
Statistics don't lie. But our ignorance, fueled by zealous sales
people, can be used against us. From Michael's hypotheses it is very easy
to prove both. Most ISP are less expensive than VAN's. But given the
right set of additional parameters and the VAN's will come out ahead. In
short the original hypotheses are useless unless you are trying to sell
something to an unsuspecting buyer.
What's key to this discussion is whether or not all the
other parameters are the same. I fear our hypotheses are too broad to be
tested. At least the cost is not a simple access fee. There are many
other costs which must be factored into the test. Michael did a good job
of showing us just how many other parameters are involved.
As Michael noted, to make an informed decision an
number of things, unique to your business, must be looked at. In general
internet transactions are less expensive. They offer a wider variety of
formats and options. But, unless you are in a position to take advantage
of these other options, the VAN will be the best deal.
|
- Internet vs VAN vs capabilities and cost = Business ... Michael Burbury
- Salary survey Victoria Narezhnaya
- Re: Internet vs VAN vs capabilities and cost = ... Leah Closson
- Re: Internet vs VAN vs capabilities and cost = ... Rachel Foerster
- Re: Internet vs VAN vs capabilities and cost = ... Stephen O'Shaughnessy
- Re: Internet vs VAN vs capabilities and cost = ... Beecher, Anthony
- Re: Internet vs VAN vs capabilities and cos... Ginny Crane
- Re: Internet vs VAN vs capabilities and... Michael Burbury
- Re: Internet vs VAN vs capabilities and cos... Benita Delfin
- Re: Internet vs VAN vs capabilities and cos... Michael Burbury
- Re: Internet vs VAN vs capabilities and cost = ... Freres, Catie
- Re: Internet vs VAN vs capabilities and cos... Michael Mattias
- Re: Internet vs VAN vs capabilities and cost = ... Beecher, Anthony
- Re: Internet vs VAN vs capabilities and cost = ... Leah Closson
- Re: Internet vs VAN vs capabilities and cost = ... Beecher, Anthony