Putting aside the meaning of "communications," I think LE could have been mistaken early on.
Here's a scenario: Mollie may have logged into various programs on her computer before she went for her run, and when LE spoke to different people they may have said they logged into the programs at XYZ time and saw Mollie was online. For example, Mollie says she can't go out with her friend because she has to write a paper. Mollie is on Facebook before her run, and her chat status is active. She leaves to run, never comes back. Her friend logs on to FB at 11pm and notices Mollie's chat status is "active." That friend might have honestly and truthfully testified to LE that Mollie said she was staying home to do homework and she was online at 11pm.
Later, when Mollie's computer is forensically investigated, they may have seen that the last activity was actually at 7pm. They believed the initial reports, and the person was being truthful, but the digital evidence did not support it. Based on the evidence, LE may not be able to identify any verifiable message sent or received after she was seen by witnesses (or on camera) on her run.