Why do sociologists on academic periphery willingly support bibliometric indicators?