I began to explore the scientific literature to back up that sentiment.
This from Street & Masip (2015) :
According to Bond & DePaulo (2006) double standard framework, people believe liars are tormented, shameful, and conscience stricken, and so display nervous behavior. Since lies told in laboratory settings are low-stake and most everyday lies are of little consequence and easy to rationalize by the liar, liars normally do not show indicators of anxiety, shame or guilt, and hence are judged to be honest (Bond & DePaulo, 2006).
These authors suggest another explanation
Our starting point here is still another explanation of the truth bias, one that contends that it is a consequence of heuristic processing.
This hypothesis is rooted in the assumption that there are Two models of reasoning.
Regardless, the existence of a truth bias seems to be fairly well accepted. Here is an article from nbc on the subject.