While working with an SPSS file which produces a Spearman analysis on a pair of column I observed that the Value of N in the output is not the same. How is that possible?
If I have 2 vectors X and Y of length 20 and I do a spearman analysis on it but Y vector has some 0's and Empty values, will that effect the N value at the end?
Assuming that the two random vectors are negative dependent. Then, your proposed tests are more advantageous in these two points. They perform much better in independent tests for two random vectors in investigated cases. The closer rs is to zero, the weaker the association between the ranks.
Hello Harish,
Correlation implies assessing the strength of relationship between scores on two (or more) variables. Therefore, if one vector has missing data points, you will have fewer paired scores with which to work than the length of the two vectors.
So, the answer to your final question is, "Yes," missing or unpaired points are omitted from the computation of a correlation.
Good luck with your work.
The results of the use of an algorithm in statistical analyses cannot vary if the input values are constant. However, they can vary if different algorithms were applied for the treatment of data. It is possible that, for example, missing values were treated differently and in that case, the value of N can be different in the output.
- Badges
- Science topic
- Similar topics
- Vectorization