# Interpreting the Spearman's correlation test coefficient and p-value

#### AppliedLinguisticsStats

##### New Member
I am currently trying to figure out how to use SPSS and understand the results that are generated by the different tests I'm performing on my data.

I doing an MA in Applied Linguistics and my tentative title is The attitudes of Chinese university students (in a Shanghainese university) to English Medium Instruction (EMI) and if/how these attitudes can be seen to affect in-class behaviour. My first research question (which I'm working on at present) is;

1. What is the link between attitudes to EMI and personal circumstances such as gender, level of English, home province, years spent studying English, number of languages/dialects fluently spoken and undergraduate degree.

At present I'm trying to find a correlation between 'years spent studying English' and attitudes to EMI. I gathered the data by means of a self-report survey and have encoded the responses into SPSS. I've encoded 'years spent studying English' as a scale measure, as the respondents could enter in any number they chose, and I've encoded the responses to the statement 'I like studying my degree through English' as an ordinal measure because students chose their level of agreement to this statement on a Likert scale from 1 - 6 (1 = strongly agree, 6 = strongly disagree).

Having been watching youtube videos about SPSS I believe that the best way to find such a correlation is to run a Spearman's correlation test on these two variables. I've done this and gotten a coefficient value of -0.128 and a p-value of 0.018.
My interpretation of this is that -0.128 indicates a slightly negative correlation between the two variables and the p-value of 0.018 indicates that this is statistically significant (as it is less than 0.05) meaning the hypothesis that "Years spent studying English has an influence (albeit slight) on attitudes to EMI" is correct. Is this a correct interpretation of this data?

If I want to test for a correlation between gender and attitudes how would I do this?
If I want to test for a correlation between home province and attitudes to EMI what kind of measure should I encode home province as? ( I previously encoded it as ordinal and assigned each province a number from 1 to 30, 1 = province with highest average disposable income per capita, 30 = province with the lowest. I'm not sure if that was the best way).

Thank you to anyone who took the time to read this query, I really appreciate any suggestions, advice or insight you might have.

All the best
Rob