DNA Lab Analysis - The Accuracy is Questioned

Recently I attended a workshop by Dick Eastman on Cloud Computing provided by my local genealogy society.  Dick spoke briefly, a lunch break was given and then the workshop resumed.  Although his information was interesting, it was the side conversations I overheard during lunch that piqued my interest.
I need to offer a disclaimer first - one of my children is employed by a large laboratory in the U.S. and part of the job responsibility is to trouble shoot and then correct problems that individual labs are encountering.   The troubleshooting my child does is regarding equipment and not results.  To my knowledge, none of that organization's business is in DNA analysis.  Even so, this proud momma often hears from family and friends who got results back that there must have been some mistake - how could whatever level that was being measured be so high, etc.  It was with this background that I brought to eavesdropping on the conversation at the next table...
A woman was explaining that she had recently had her DNA results returned and she wasn't matching with anyone in her family.  She is unmarried and has no children so none of them tested.  Her parents are deceased and she had no siblings.  By matching, she was referring to cousins.  A man at the table conjectured the lab had made a mistake and mixed up the samples.  Another attendee reported that his results matched with his children, siblings and first cousins but not with relatives from 3 generations back.  He, too, originally thought the lab had erred.  Then a match occurred with a surname which he was not familiar.  He thought he had somehow missed that line in his research so he went back over his records and low and behold, discovered that the matching surname lived in the same boarding house as his 2x's great grandmother.  Hmm.  And yes, great grandma was married to who he had assumed was his great grandfather at the time.  There went all of his research on that great grandpa's line!
Could a lab make a mistake?  Absolutely!  The likelihood, though, is not as great with the processes and procedures that are in place as is the entanglement of human relationships.
The following day I was reading a list serv to which I belong and an individual had posted how she had inadvertently given a female DNA test kit to a male relative.  The lab caught it and asked for clarification.
My advice if your returned results give you unexpected findings - get the test redone at another site.  Prices are dropping for the holidays so the cost is negligible.  There are "rumors" that Ancestry will run a special beginning November 25th for $69.00 to beat the FTDNA price of $79.00.  I don't have that in writing so check around on the 25th to see what happens.
When the test results are returned, if they're similar, well, you know you need to explore other lines to determine who's the daddy.  If they are not the same, I'd contact the lab and share your findings.  You'd probably get your money refunded if the lab made the error and an offer for another test as a thank you for letting them know there is a quality control problem.  Personally, I'm betting on the relationships and not the lab as the culprit.