My problem is tha I have 40 teachers as participants in likert scale and 90 student scores. How can I find the correlation between teachers' leadership skills and student scores ( ordinal independent variable and continuous dependent variable)?
Most people compute correlations between rating scales and continuous variables. Another issue is the nested nature of the data--there is more than one student/teacher, so students are nested in teachers. Depending on your question, you could average the student scores within each teacher (n = 40) to see if the teacher variable relates to their mean student scores. If your focus is on students, you don't have enough students/teacher to do multi-level modeling. You could associate each student with their teacher's data (n = 90), but there are dependencies in the data that could affect results. One way to deal with those is to compute pooled within-group correlations that adjusts the sums of products and sums of squares for each group (students within teacher). But again, your number of students is small given the number of teachers. I imagine some teachers might even have only 1 student.
I'm a beginner with SPSS. I'm free to choose the number of students, can I take two groups of students (40 each) make grouping and apply spearman correlation with teachers' likert scale? (teachers n=40)
Do you know who are the students under a particular teacher? If you do, you can take mean of those students' score as the variable against the leadership score of that that teacher