Research your area to find old buildings reputed to be haunted. Keep the findings to yourself, as prior knowledge by your students will ruin the experiment. If you do not find any, set up two rooms with subtly different environments; nominate one as "haunted", the other "controlled". Don't make the "haunted" one too obvious. Use your imagination: dim lighting, barely audible piped sounds, a hidden fan to create a draught.
Create two questionnaires. In the first, students rate the degree to which they know where, in the local area, people had experienced unusual phenomena in the past (definitely yes, probably yes, uncertain, probably no, definitely no).
If you are creating a haunted room, just do the second questionnaire. Ask students to quietly walk around the rooms and write a brief description of any unusual experiences, marking on a floor plan where they were standing, and stating whether they believe their experience(s) were due to a ghost (definitely yes, probably yes, uncertain, probably no, definitely no).
Discuss ways to analyse the data. Should floor plans be divided into squares and percentages calculated? Should data be plotted on a graph or other display? What sort of report should be written? It should include an overview of the area investigated, the method and procedure of the experiment, the results and the conclusion. How do students with varying beliefs come to a collective conclusion? It should make for fiery debate.
The above is based on advice from Dr Caroline Watt