A detailed discussion would be appreciated about spearmon and pearson correlations. Can we use any of them whether the smapling is probability or non probability?
I want to make it clear that my data is normal (normally distributed)
Spearman is used for non-parametric data, which make no assumptions about the probability distributions of the variables being assessed. Pearson correlations is a measure of the linear correlation between two variables, X and Y, and doesn't assume the distribution either :)
I totally agree that Spearman is a nonparametric test for the correlation between two variables. The variables have to have at least rank order (i.e. they can be described on an ordinal scale). I also agree that Pearson correlation is for linearly related variables. The following website provides further details and examples: www.statisticssolutions.com
On this website Pearson, Kendall and Spearman correlation are compared to each other.
Best wishes and kind regards,
Rainer
Measurement scales are nominal, ordinal, interval and ratio in order of strength. Spearman's correlation coefficient being for the nominal and ordinal scales is based on ranking which is associated with loss of information. It is non-parametric. If applied to interval or ratio variables it cannot compare favourably with the parametric Pearson's correlation coefficient in terms of power. In fact, whereas Pearson's coefficient can be applied to any variable, Spearman's coefficient should not be applied to interval and ratio variables. This position is irrespective of whether sampling is probability or non-probability.
If you want to test the significance of the correlation coefficient for the entire population (this is true whether pearson or spearman is used ) then the sample should be drawn using probability sampling.
The Pearson correlation evaluates the linear relationship between two continuous variables. ... The Spearman correlation coefficient is based on the ranked values for each variable rather than the raw data. Spearman correlation is often used to evaluate relationships involving ordinal variables
and i agree with answer of Anthony Capili
and if you want to test normality of the data then do test of normality . you can find normality test from SPSS Analysis - Explore
see this youtube link to learn it
https://www.youtube.com/watch?v=sQkB-AlJgPI
Spearman is a non-parametric test which is is based on the ranked values (least should have rank order). But Pearson correlation is used to to identify he linear relationship between two continuous variables.
I use 30s bioclim global layers for maxent modeling. I want to remove correlated variables, none of the tools worked as the layers are too big. But when i used 10m layers they worked well.
Can some body help me to resolve this issue?
If you need to remove correlated variables , you can drop one of them arbitrary. But you can use your experience to keep the most suitable variable which can explain your outcome variable better. or if you replicate a previous model, go through it. .
Thank you all.
I have not used VIF previously. I have my environmntal predictors as raster asc files. So what file format needed? Is it done in R ?
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Probability