Can you elaborate more on your model and your variables? For me, your description does not make much sense. In ANOVA (in the simplest form) you have one metric variable (DV) and one nominal independent variabl(IV). How did you calculate the correlation between your DV and IV?

A correlation test (e.g., Pearson’s, Spearman’s) is used to examine the strength and direction of association between two variables. Analysis of variance (ANOVA), on the other hand, compares the means of two independent groups (or more) to see if the related population means differ statistically. ANOVA will show a statistically significant result if any group differs considerably from the overall group mean. Simply put, a correlation test and ANOVA are very different tests addressing different purposes, based on different statistical assumptions. Their p values could have little in common. Therefore, I agree that you should have rendered more inputs into your inquiry. The following libguide might be helpful.

KSU Libraries. (2021, November 19). LibGuides: SPSS tutorials: Analyzing data. LibGuides at Kent State University. https://libguides.library.kent.edu/SPSS/AnalyzeData

Good luck,

Mohialdeen Alotumi it depends and it is not entirely true that ANOVA and correlation are "very different tests". Both are part of the linear model and in case of only two groups, ANOVA and (biserial-) correlation are identical, but put the emphasize on different aspects of the model. It is the same if you say that there is a relationship between the group membership and the dependent variable (correlation) or group membership leads to differences in the mean values in the dependent variable (ANOVA). You could also view it as a regression model, where the slope describes the mean difference between the groups. All three approaches lead to the same p-values in this case! The explained variance in all models is the same and hence the effect size, r2 of correlation or eta2 for ANOVA.

The situation changes slightly if you have more than two groups, but not the basic principle. With more than two groups it gets more important to describe what you really analyzed.

@Rainer Duesing @Mohialdeen Alotumi

Well, i want to elaborate more about my research.

In ANOVA, i want to check that which category of (education level) is most related to number of hours with( watching satellite television). Because when in Correlation you got a significant association between education and hours spent with satellite television. You want to explore which category is most relevant in watching satellite tv.

That does not explain much M.Adeel Kamran . What kind of variables do you have, nominal/categories, metric etc. Which kind of correlation did you use?