I don't know which methods I can use here - t-test, ANOVA, Pearson or something else

Hello,

I was wondering if you might help me with my work, I have six categorial variables: gender (m/f), education (high school/university students), time spent on twitter, number of folowers, frequency of tweeting and frequency of posting selfies (these four have five categories), and on other side I have two continual variables: narcissism measured by NPI and self esteem measured by RSE.

So, my questions are what method shoud I use? I used two t-test for gender and educational differences on nacissism, but I want to know if there are differences between male high-school and university students on narcissism and self esteemand the same for female, and if there are differences between male and female high-school students on narcissism and self-esteem and the same for university students. I can use ANOVA, but than I don't get significant difference between high-school and university students on self-esteem, and when I do t-test, there is a significant difference, I'm so confused.

Also, which method shoud I use to find a corelation between other four categorial variables and narcissism, and the same with self-esteem. I cannot use Pearson's corelation test, right?

If you can help me, I would be thankful. Also, I'm very sorry for my bad English.