Univariate analysis, anova test or ttest when comparing three groups

[Herman Bloemhof]

I am researching retail investors where I look at three different groups based on the stock price. Low includes deciles 1-3, medium contains 4-7 and Hgih includes 8-10. Then I do a univariate analysis per group and analyze different variables such as stock price, assets per investor, leverage, number of investors etc. By testing this difference, I can provide evidence that supports or refutes the idea that different stock price groups have distinct investor profiles or investment behaviours.

My question is should I do an ANOVA to test the differences across groups? or would it be better to compare low to high groups with a normal ttest.
What would be the preferred method?

[Carlo Lazzaro]

Herman:
given the number of your predictors, why not going -regress-?

[Felix Bittmann]

I think Carlo`s idea is great if you want to use raw deciles and go for a linear trend. Then, even a simple correlation analysis, maybe with spearman, might be fine. However, comparing extreme terciles is a rather different research question. You need to decide what is the main goal of your study.

[Herman Bloemhof]

Thanks for the fast replies. I do a univariate analysis to see how the data is constructed and will regress later. But first I want to see how the data is spread over these different deciles.