Statistics are interesting and help establish relationships among diverse phenomena, such as between parental income and a child's likelihood to go to college. So, statistics have value. However, the social and natural sciences have come to revere statistics as an a oracle, a truth telling device.
All other modes of inquiry are denigrated in comparison to the "truth" (however contingent) revealed by statistics.
Consequently, research inquiry that fails to adopt statistics is rejected.
I enjoyed learning statistics but often found problematic the research models that were required to be adopted in order to use statistics to study social interaction.
I was, frankly, appalled when I learned about how statistics were being used by behavioral geneticists to explain human behavior in terms of gene alleles.
The following link clearly outlines some of the problems inherent in this blind faith in statistical analysis