SciencesPo (version 1.02.12)

rprob: Pairwise Correlations With Probabilities

Usage

rprob(x, df = nrow(x) - 2)

Arguments

x

A data object

df

An optional value for degrees of freedom. The default assumes you have a common degrees of freedom (n-2) for all correlations in the table.

Value

• A data frame object with the Pearson's product-moment coefficient and its probabilitie. bold{The first diagonal display the correlations, while the probabilities are shown in the second uppper diagonal}. } description{ Compute the pairwise correlations for all valid cases (!NA) in a data frame and find the probabilities for each combination. This function parses the correlations below the diagonal and the significance probabilities above it. } details{ The connexion between correlation and t-statistics is known to be eqn{t = frac{r sqrt{(n-2)}} {sqrt{(1-r^2)}}}. Or then, F = t^2 = r^2 * (n - 2)/(1 - r^2) ~ F(1, n-2). By solving this, it is possible to find the probabilities. }

  ote{ You can print the whole matrix using code{cor(t(x))}. } examples{ data(nerlove63)

  rprob(nerlove63)

  # a stacked up table rstack(rprob(nerlove63)) summary(lm(output ~ plabor + totcost, data = nerlove63)) # The final p-value of the OLS compares to the probabilities in the # intersection of output and plabor and output and totcost in the matrix. } author{ Daniel Marcelino, email{[e-mail]} } references{ Aldrich, John (1995) Correlations Genuine and Spurious in Pearson and Yule. emph{Statistical Science,} bold{10(4),} 364--376.

  Spiegel, M. R. (1992) Correlation Theory. in: emph{Theory and Problems of Probability and Statistics,} 2nd ed. New York: McGraw-Hill, pp. 294--323. } seealso{ code{rstack} } keyword{Descriptive} keyword{Tables}