I identified 6 sampling stations in a village, analyzed the collected water samples for 14 parameters such as pH, EC, TDS, THW, Ca, Mg, Na, K, Cl, Fe, Nitrite, phosphate, total alkalinity... I wish to calculate correlation matrix.. It was found that out of all the six stations analyzed, the analysis values obtained for certain parameters are the same. Do I violate that parameter while calculating correlation matrix?
Please find the attachment.
Having identical values for some observations is not a problem. Remember for correlation analysis that variables should be approximately normally distributed, and should be appropriately transformed if they're not. You have only six observations for each variable, which is not really enough to calculate meaningful correlations. In your dataset, I presume that "nitrite" is actually nitrate?
I know this is late but, why would you create a correlation matrix? If you have more than a few things to correlate, some of the positive correlations will be false.
If all of your samples come from a single village, I would expect most of the samples to be similar. If you collected samples from several different villages, I would use cluster analysis.
- Badges
- Science topic