Heterogeneity and meta-analyses: do study results truly differ?