[Gretl-users] Re: First time user of Hansl and creating a matrix of Spearmans correlation

On 30/01/2024 14:19, Artur T. wrote: [...] Here's a slight variation on Artur's script, that doesn't give you tstats and pvalues, but what you get is more similar to a "traditional" correlation matrix: <hansl> function matrix spearman_matrix(list variables) n = nelem(variables) strings names = varnames(variables) matrix SC = I(n) loop i = 1..n loop j = 1..i-1 # Calculate Spearman correlation and p-value scalar rij = npcorr(variables[j], variables[i], spearman)[1] SC[i,j] = rij SC[j,i] = rij endloop endloop # Set row and column names rnameset(SC, names) cnameset(SC, names) return SC end function </hansl> ------------------------------------------------------- Riccardo (Jack) Lucchetti Dipartimento di Scienze Economiche e Sociali (DiSES) Università Politecnica delle Marche (formerly known as Università di Ancona) r.lucchetti(a)univpm.it http://www2.econ.univpm.it/servizi/hpp/lucchetti -------------------------------------------------------