Spearman correlation is simply the traditional product-moment coeffcient of correlation )"Pearson correlation") computed after both of the variables are transformed to an ordinal form (1, 2, ...). So, the interpretation is parallel to normal Pearson correlation. If the relation is more curvilinear than linear, coefficient eta could be a good alternative. In the non-linear settings it gives a better approximation of the true association (in population) than Spearman or Pearson coefficients. Hope you get out of the "stuck"...
Can you tell us more about what your original measures were and how the items were worded?
All you are doing with a Spearman coefficient is ranking the values and then running a correlation on them. As Jari noted, the interpretation is pretty much the same, but the one caveat is that Pearson's r estimates the linear correlation (in other words, your data should be linear for this to appropriately model the correlation). Spearman's rho measures the monotonic association and therefore does not strictly require linearity (see images in below link to see where the distinction lies).
Reference:
https://statistics.laerd.com/statistical-guides/spearmans-rank-order-correlation-statistical-guide.php#:~:text=Spearman's%20correlation%20measures%20the%20strength,is%20monotonic%2C%20but%20not%20linear.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size