# Thread: How to choose analyzing method (using discrete and continuous variables)?

1. ## How to choose analyzing method (using discrete and continuous variables)?

This is my first work, doing it for university. I have nominal, ordinal and numeric (or what they're called in English) variables. I finally figured out I have to name them discrete and continuous variables according to what exactly did I measure with them.

But is there any formulas to know which analyzing method to use when I have for example nominal discrete and ordinal discrete? Like, any charts, tables to show if I should use ANOVA for example?
What formulas/methods to use to compare with each other all these combinations?

Discrete nominal/ordinal/numeric variables
Continuous nominal/ordinal/numeric variables

I have found two different charts and they have very different results of what methods to use (where one says ANOVA, other says regression...). And I am too amateur to figure out what to trust or use.

Can anyone advise about such formulas? Is there some rule of thumb to use? I'm running out of ideas and scholarly books are talking about something, but I'm too new to this to understand without help.

I need to get this thing done, but how to choose analysis method is giving me nightmares.

2. ## Re: How to choose analyzing method (using discrete and continuous variables)?

Hey Grovius.

It depends on the model to be as generic as possible [in my answer].

The data type and the data range all impact the model and each kind of statistics takes that into account [i.e. categorical, survival analysis, normal linear models etc].

You have to look at each of the assumptions and check that the technique you use satisfies them and the data that you use.

The short answer is that it's not easy and you need to build up intuition to really answer this question well.