Chapter 10 | Exercise 1
Compute the other associated measures from Table 10.4 of the textbook for the Russian verbs used in the ditransitive construction and examine the relationships between these measures. Which ones are most strongly correlated with Attraction, and which ones with Reliance (use Spearman's rho > 0.9 as the criterion)?
Use the code in Table 10.4 in the textbook to compute the remaining association measures for the Russian ditransitive study. Next, combine all scores as columns in a matrix:
> cscores <- cbind(Attr, Rel, dP.cueCx, dP.cueVerb, logpvF, LL1, PMI, MI2, MI3, MIloc, nPMI, z.score, t.score, chisq, MS, Jaccard, Dice, logOR, logOR.disc, gmean, Liddell)
Compute Spearman’s correlation coefficients for pairwise comparisons.
> cor(cscores, method = "spearman")
Attraction is very strongly (ρ > 0.90) correlated with MS, Jaccard, Dice, dP.cueCx, MI3, Liddell and t-score. Reliance is very strongly correlated with dP.cueVerb, PMI, MI2, nPMI, z-score, logOR, logOR.disc and gmean.