Re: Data Sets

On Apr 6, 10:23 am, Teresa <stephenson.ter...@xxxxxxxxx> wrote:
On Apr 5, 12:28 pm, mcap <mca...@xxxxxxxxx> wrote:





On Apr 3, 11:38 am, Teresa <stephenson.ter...@xxxxxxxxx> wrote:

I am teaching my Cost Accounting II class how to use regression to
determine the cost drivers associated with overhead.  I was wondering
if anyone out there has a data set that might be appropriate to show
them how to use interactions and dummy variables (not necessarily at
the same time).  Most of the cost books I have show data sets where
only one variable ends up being significant.... I would like to be
able to show several results:

1. More than one independent variable are significant
2. Where a dummy variable has to be used
3. Where there is an interaction and the students have to multiply the
columns to create an additional variable.

If anyone knows where to find such data sets -- especially if they are
in a business setting -- it would be most appreciated.  I have some in
my stats books but they are perhaps too complex for my level of
student.

Hi Theresa;

   I ran into this problem when I was teaching epi last summer.  I
looked at a lot of different data sets.  SPSS, in particular, has a
lot of business related data sets that come with the program.

   In the end, because I wanted to show specific things I ended up
programming my own. It was really pretty easy - you just need to use a
loop function to create subjects and then "if" and random variable
functions to create groups.   For example if you want a significant
difference between groups, you can use a combination of if and random
variable functions with specified means and sds to make your
statistical test come out like you want to.  This way, you have
complete control over what you end up with.  The downside is that your
data is a lot more clean and a lot less random than actual data.  For
teaching purposes however, it may be a good thing.   The example you
specify shouldn't be that difficult to create.

   Email me if you want some examples.  Of if you have a specific
database you want, I could possibly put something together for you
when my term ends.

Marc

Thanks.  I'll give that a shot and see what I come up with.  I've
never created my own data set before :)

Teresa- Hide quoted text -

- Show quoted text -

It's really not that difficult. Get in touch if you run into
problems.

Marc
.

Relevant Pages

  • Re: Data Sets
    ... determine the cost drivers associated with overhead. ... I ran into this problem when I was teaching epi last summer. ... lot of business related data sets that come with the program. ... specify shouldn't be that difficult to create. ...
    (sci.stat.consult)
  • Re: Data Sets
    ... determine the cost drivers associated with overhead. ... I ran into this problem when I was teaching epi last summer. ... lot of business related data sets that come with the program. ... specify shouldn't be that difficult to create. ...
    (sci.stat.consult)