Voice risk analysis, or 'voice stress analysis' does not detect lies, it detects the stress in a person's voice that occurs when there may be negative consequences to their response to a question. This is an important distinction, because some people will become stressed even when they are not lying, ironically, because they fear that the machine will indicate that they are. Also, some people are able to lie confidently and, therefore, will not be detected. A recent study by Kelly Damphouse and his colleagues at the Oklahoma Department of Mental Health and Substance Abuse Services  showed that Voice Stress Analysis only detected 15% of drug users who lied about their recent drug use (actual drug use was measured using urine tests) and 8.5% were incorrectly classified as liars when they were telling the truth. This means that benefit fraudsters need not worry about being convicted on Voice Stress Analysis alone; however, the local authorities using the equipment plan to use it to identify people for greater scrutiny. Many fraudsters, nevertheless, could potentially avoid being caught due to the low detection rates.

While Voice Stress Analysis may not be particularly effective at detecting lies, it may be very effective at getting people to tell the truth; the money saved by Harrow Council was due to people admitting that their circumstances had changed, not because of the detection of lies. This relates to a well known procedure in Psychological research known as the 'bogus pipeline'. The bogus pipeline technique is usually used when participants are asked about their attitudes  towards controversial issues, such as race. Participants are informed that they are connected to a lie detector (which is actually fake) before being questioned. Typically, participants are more likely to reveal controversial attitudes when they believe the machine will be able to detect their lies. In the Damphouse et al. (2007) study it was found that, consistent with bogus pipeline research, the arrested drug users who were interviewed using the voice stress analysis equipment were a great deal less likely to lie than participants in a previous study which did not use Voice Stress Analysis .

So, just as long as fraudsters do not know the system is flawed, these councils should save lots of money.