There are few studies in literature where adiponectin levels are measured and compared among patients and controls or pre and post some drug therapy, however, a reference range is not given.
Attached are some papers mentioning reference range of serum adiponectin.
There is also a meta-analysis and in it you can see the big heterogeneity between studies due to different cohorts (obese, T2DM, PCOS, ect). So it is difficult to stablished a universal reference range as what you could consider "normal values" of it may differ depending on which population are you referring too or comparing it to with (healthy lean subjects vs healthy obese; metabolically healthy obese vs metabolic syndromic ones…..etc).
May be Javier Gomez-Ambrosi or Amaia Rodríguez could be more helpful here as they havea large experience in the analysis of this molecule in their research unit and have published different papers analysing the correlation of adiponectin levels and adiponectin-leptin ratio with cardiometabolic disease. I kindly refer this question to them :D
I think that reference ranges of different variables must be established based on many points like lab methodologies& standards in addition to other categories that may impact sometimes like race, age,... Etc.
You can do this by characterizing the pattern of distribution of your data. So, if your data are Gaussian distributed, the the reference range would be mean+/-2SD, or otherwise mean+/- 3SD if not. Thanks