One of the rooms was supposed to be genuinely haunted, whereas one of the others had concealed speakers generating infrasound. I think it was the latter which people claimed gave them a spooky feeling.
Precisely why it would be fascinating to see the non-supernatural vs supernatural based explanations and behaviours for induced phenomena.
Get yourself a copy (it's out there, try google scholar) of
"The “Haunt” Project: An attempt to build a “haunted” room
by manipulating complex electromagnetic fields and infrasound
Christopher C. French, Usman Haque, Rosie Bunton-Stasyshyn & Rob Davis"
Although the paper is little light on the exact nature of the infrasound and electromagnetic fields used (I would say that), their conclusion was "that people who believe in ghosts experience ghostly phenomena and the room made no noticable difference"
It also discusses the problems with experiments of this type. How does one create a control condition for example and ethically it's not always appropriate to completely hoodwink people about the nature of the experiment (thus some priming occurs) and so on.
If you ferret about Prof. French's work, a lot of the stuff that seems to show some physical phenomena responsible for 'haunting' hasn't been verified as the cause, only a coincidental condition. So case where large EMF were detected at the site of haunting hasn't been show to cause an experience, only that it was there at the same time.
Worse, if you do it as a TV show you'll get a higher than average number of folk with narcissistic traits and that'd skew the results again. You'd have to select using (probably several) personality trait questionnaires for a start, and one of the issues is the questionnaire you offer can prime enough people as to the nature of the experiment to skew the results again (in an ideal world you'd counterbalance with some people 'room first' and some 'questionnaires first' but sample sizes are starting to grow...). Simples...