Hi,

usually in ANOVA the IV are categorical, if they are continuous, as it may be understood from your description, you should rather use multiple linear regression.

If ANOVA is mandatory, then, I think, your only choice will/would be to categorize your IV, with the disadvantage of some loss of information.

Best wishes.

Do not categorize your variables!! ANOVA is just a special case of the linear model, therefore, as Jean suggested, use a multiple regression. In multiple regression you can use dichotomous variables AND continous variables at once.

And by the way, if you have 3 IVs and they would be categorical it would be a 3-way ANOVA and neither oneway or 2-way....

You are wrong in saying this sentence

" and wanted to apply ANOVA as a measurement "

Abolfazl Ghoodjani I would give it the benefit of a doubt and think it's just a language or translation problem.

Thank you @Jean,@Rainer and@ Abolfazl for your quick response. I'm new to research hence seeking your help. Just to reclarify .. my variables do not have any subset or levels under each of them. However ANOVA is supposed to have subsets for every variable. Can I still apply ANOVA that is 3 way as per the suggestion. All variables are continuous. Trying to perform in excel but do not have option for 3 way. Already applied mulitple regression. Looking for another tool for analysis hence thought of ANOVA.

Hi,

if all variables are continuous, you must perform multi-linear regression and not Anova, unless you are constrained to do so by circumstances, then you have to first split your IV in categories. I did not use Excel since years but as far as I remember a multi-linear regression is available, even if it is not very handy.

Regards.

Thanks@Jean.. will do the same

Hey,

If your variables are continuous you should perform regression analysis. However, using Excel for these purposes is a very difficult task. I advise you to use specialized software that will allow you not only to build a graph, but also to evaluate all the regression coefficients, as well as the residuals. In addition, perhaps linear regression will not be enough for you for analysis and you will need tools for building nonlinear models (for example, STATISTICA (StatSoft Inc) contains approximation algorithms for Newton-Gauss and Levenberg-Marquardt). The easiest to learn are the STATISTICA and SPSS software packages. In the future, if you plan to seriously engage in data analysis, use the statistical programming languages ​​like SAS or R (freeware).

I hope I helped a little