Nonetheless, an -test returns a borderline significant -value, which can be taken as support for the breakpoint model ( , ) Two other relative goodness-of-fit measures likewise produce borderline results: the breakpoint model has a slightly better (i.e. lower) Akaike Information Criterion [53] value than the simpler model (642.3 and 644.4, respectively), but a slightly worse (i.e. higher) Bayesian Information Criterion [54] value (651.6 and 651.4, respectively).
← all excerpts
The critical period hypothesis in second language acquisition: a statistical critique and a reanalysis.
1
—
—