A person, trying to pretend how busy he is, answers to your messages right after 3-4 days. Always -- 3 or 4 days. Not $1..2$ days or $5..\infty$ days.
You may get suspicious, because truly busy persons tend to react after more random intervals. This is where variance can be used.
In Wolfram Mathematica (all values are days):
In[]:= Variance[{3.7, 4.8, 3, 4, 4.3, 3.2, 3.1, 4.1}]
Out[]= 0.410714
In[]:= Variance[{1.2, 1, 0.5, 2, 3, 3, 4, 7, 6.1, 5.4, 0.7, 2.1}]
Out[]= 4.81455
The smaller variance value, the more probability you see that the person just answers when he see '3d' or '4d' mark on your message in his inbox, not before or after. The bigger variance value, the more probability this persons acts genuinely.
Yes, I know about these lousy Disqus ads. Please use adblocker. I would consider to subscribe to 'pro' version of Disqus if the signal/noise ratio in comments would be good enough.