TETRA: On/Off Judgements

TETRA study results

1)     What are on/off judgements?

Some sensitive individuals report that they know when they are in the presence of an electromagnetic field signal such as a mobile phone base station.  As part of our investigation into possible health effects from mobile phone base stations we wanted to test whether our participants - in particular our sensitive participants - could tell when the base station was on and when it was off.  

2)     What did participants have to do? 

All participants judged whether they believed the base station was ‘on’ or ‘off’ on six separate occasions. All six judgements were made under double-blind conditions: this means that neither the experimenter nor the participant knew whether the base station was ‘on’ or ‘off’. 

The first 4 judgements were made in the participant’s first session at the laboratory. Each participant sat through four ‘exposures’. Each exposure was five minutes in duration, with a two minute ‘wash out period’ in between each exposure. A wash out period means that the base station is off. 

The four judgements comprised two ‘on’ conditions (TETRA exposure) and two ‘off’ condition (SHAM exposure). SHAM means that the base station is off. The computer selected a random order for the exposures for each participant. So a participant might receive TETRA, TETRA, SHAM, SHAM, or SHAM, TETRA, TETRA, SHAM and so on. Participants made their judgement after each five-minute exposure.

The second set of judgements was made in sessions 2 and 3. Each of these sessions lasted 50 minutes and was separated by a minimum of one week. Again, people knew that over the 2 sessions they would receive one ‘on’ condition (TETRA) and one ‘off’ condition (SHAM).  Thus, during each 50-minute test participants received only ONE exposure condition (e.g. SHAM in Session 2 and TETRA in Session 3 or TETRA in Session 2 and Sham in Session 3). As with the first set of five-minute exposures, the exposure conditions for sessions 2 and 3 were presented in a random order which had been pre-selected by the computer.  People were asked to make their judgements at the end of each 50-minute session.

For all six judgements participants were asked to report whether the base station was ‘on’ or ‘off’. Participants were then asked to rate how confident they were in these judgements. Confidence was expressed as a rating of 1 (low), 2 (moderate) or 3 (high).

3)    What data did we analyse?

We used two main bits of information:  whether participants thought the base station was ‘on’ or ‘off’ and how confident they were about their judgement. 

To find out how accurate participants were in comparison to what we would expect by chance, we calculated how many ‘on’ judgements were correct and how many were incorrect for both groups. The same was done with the ‘off’ judgements. 48 sensitive participants made four judgements in session 1, resulting in a total of 192 judgements. The same 48 sensitive participants made two judgements over sessions 2 and 3, resulting in a second total of 96 judgements. 132 control participants made four judgements each in session 1 and two judgements each over sessions 2 and 3. This resulted in a total of 528 judgements for session 1 and 264 judgements for sessions 2 and 3. The results are displayed in Table 1 below.

Table 1: Showing the number of ‘TETRA’ and ‘Off’ judgements made by sensitives and controls for each exposure condition. A bold, italic typeface indicates correct judgements. ‘N’ refers to the number of people who made judgements. The table shows the judgements made by participants who completed all three sessions and thus made a total of 6 judgements each. Some participants made 4 judgements in session 1 but did not go on to complete judgements 5 and 6. To see a summary of judgements 1 - 4 that includes the data from these participants, please click here.

4)    What would be expected if participants responded at chance level?

Participants made six judgements in total.  The base station was ‘on’ 3 times out of 6 (50%) and ‘off’ (50%) 3 times out of 6. If participants were responding at exactly chance level, then we would expect them to be correct in their judgements 50% of the time (i.e. 3 out of 6 would be correct).

 5)    How did the participants’ judgements compare to what would be expected by chance?

By referring to Table 1 we can see that for the first four judgements made by sensitives there was a 55% accuracy (51 + 54 = 105/192 x 100). Similarly, controls were accurate 45% of the time (101 + 136 = 237/528 x 100).  For judgements 5 and 6, sensitives had an accuracy of 40% (17 + 21 = 38/96 x 100) and controls had an accuracy rate of 53% (60 + 81 = 141/264 x 100).  These percentage rates are broadly consistent with chance.  Although we have not reported them here we also undertook analyses incorporating participants’ subjective ratings of how confident they were in their judgements. This additional dimension had no impact on the results; accuracy remained at chance level.