Hello, i am unsure how to interpret a correlation if the effect is weak (.17). If i tassume that one variable predicts another variable, and the effect is weak, my hypothesis is confirmed.
But can i explain in my discussion why there could be only a weak correlation? Because maybe the variable isnt important to explain the other variable etc, that the variable could have another effect etc. Or is this thinking completely wrong? And that i dont interpret the effect from a correlation at all? Because a confirmed hypothesis is a confirmed hypothesis.
I hope u understand my problem. Hard to find the right words in english.
The interpretation depends on the sample size. If small, your Pearson's will be unreliable. Or are you using Spearman's?
There may be outliers, too, so you need to check if that's the case (and explain why).
Note that you refer in your first paragraph to a predictive relationship between one variable and another. Correlations are of course of a different ilk.
Hello Antoon, my sample size is 202. I am using Pearson's. I have outliers in my variable: abusive supervision but i didnt delete them from my sample. Cause in my opinion abusive supervision is a sensitive theme and its obvious that there should only be few people who experiences this negative behaviour of their supervisors. I hope i can offer this explanation. Is this a limitation of my study that i didnt delete outlier? abusive supervision ranged from 1-6 and the mean from my sample size was 1.9. Is this a limitation too?
and y sorry my hypothesis is: one variable correlates negatively with another. and based on theoretical assumptions i explain why i think that there is a causality, but of course i cant proof it, cause i only use correlation.
Kim: "I have outliers in my variable: abusive supervision but i didn't delete them from my sample."
Based on what you write, abusive supervision should be one of your predictor (independent) variables. Have you included it?
Have you considered grouping the values together into, say, three categories rather than six? What would the correlation look like then?
- Badges
- Science topic
- Similar topics
- Cognitive Science
- Thinking