# Help Required with Course Work

#### dogz111

##### New Member
Hi,

First post so here goes.

I have some uni course work which I would appreciate a view on:

We have been given a list of data and asked to provide analysis on it.

The data entries are:

Region (North or South)
Gender (Male or Female)
Sales (Continuous variable)
Age
Yrs with Company
Days absent
Job Satisfaction (1 - 5 - Discrete Variable)

I have produced descriptive stats and carried out t-test and chi squared to look at how sales are related to other factors.

What I want to do is to consider how each variable is related to each other and how it influences sales overall. I am not great at stats but I am a quick learner.

I have had a quick look on the internet and the following techniques appear to fit the bill.

ANOVA
MANOVA
Factor Analysis

I would appreciate any comments as to the best statistical tools to use to achieve my aims. Any help appreciated especially if it stops me wasting time reviewing techniques which are of no use.

Cheers

Dave

#### firefly

##### New Member
I often work with data like this so i'd be interested to see other peoples ideas on how to do this.

But my take is that you have continuous variables than categoricals so what you can do is to convert all your categorical variables to 0/1 variables and then apply regression analysis. This is often done although these 0/1 variables violate the assumption that the iv's are normally distributed.

Another option is to convert all your continuous variables to categoricals and then use ANOVA.

#### firefly

##### New Member
Correction:

there is no requirement in linear regression that the iv's need to be normally distributed. But I am sure I have heard that the 0/1 type variables violate some or other assumption. Although I don't think it is that serious as this type of transformation is used in practice all the time.

#### deltac

##### New Member
I'm not sure a factor analysis would be useful to you. I have only used it for test construction and you need a fairly large sample size. Maybe a cluster analysis would be more appropriate. I haven't had to do one in a few years, but basically it looks at how variables function when they are grouped together. The example I remember is pre-natal care, we all know that not smoking, good nutrition, no alcohal, etc. are important for a healthy baby, but what are the risks when these are grouped together in various ways.

#### dogz111

##### New Member
Thanks for the answers - Decided ANOVA is the way to go.

Cheers