In small-sample sociophonetics, which formant normalization method (e.g., Lobanov or Bark) has given you the cleanest apparent-time vowel shift trends? Can you point to one plot or statistic that convinced you it was the right choice?

obanov has given me the cleanest apparent time vowel shift trends, especially in small samples, because it pulls speakers into a comparable space without flattening the vowel system too aggressively. The stat that usually convinces me is that the age effect gets stronger after normalization. For example, a clearer linear slope for age in an F1 or F2 mixed model, or visibly tighter within vowel ellipses with less speaker clustering in the plot.