This is the first in a series of posts in which I’ll document my quest to find an answer to the question that has for centuries been keeping millions of people awake at night (or maybe not quite): is it possible to build a lie detector on a budget and make it work to some degree of reliability?

First let me make some disclaimers about the terminology I’ll use. Although ‘polygraph’ and ‘lie detector’ are often used interchangeably in everyday life, they’re not the same. A polygraph is an instrument that measures several physical characteristics of the human body such as blood pressure, breathing speed, heart rate etc. Some people argue that these characteristics can then be used to determine whether the person who was measured is lying, on the assumption that when people lie they get nervous and their heart rate and blood pressure will go up, they’ll start breathing heavier (or at least different from the way they breathe in ‘normal’ circumstances) etc. So ‘lie detection’ is merely an application of a polygraph device.

And to confuse things a bit more, I will not be aiming for a polygraph at all - since ‘poly’ means ‘many’ and ‘graph’ is ‘writing’, the term ‘polygraph’ should be used only for devices that can record multiply characteristics. The device I’m envisioning will only measure one characteristic (skin conductivity) so it should be called a ‘monograph’ instead of a ‘polygraph’.

Despite all of that, I will use the terms ‘polygraph’ and ‘lie detector’ and with both of them I will mean ‘a device that is used to measure certain body reactions in an attempt to determine whether the person is lying, even if it only measures and uses one body function or body characteristic’. I will consider other opinions or terminology proposals but I do not guarantee that I will agree with them

With that out of the way, what do I want to do? I’ll try to build a device that can measure the conductivity of the skin, more specifically the skin of the palm of the hand, as this is reputedly (does that make me sound like a scientist or what?) the area where people who are lying start secreting sweat the first and the most. I’ll test the device by giving people a number between 1 and 10 and asking them for each of these number ‘do I give you number x’? They will be instructed to answer ‘no’ to all of the questions, and I hope that the readings of the device will spike at the number they were given at the start of the experiment.

I’m not saying that this will for once and for all prove that polygraphs can be used to detect lies. It will be a simple device, a simple experiment and conducted on a small scale. For a polygraph to be useful in real-life scenarios (like law enforcement), a much higher degree of accuracy will be needed than what I will declare ’successful’ in this experiment (I will post about accuracy in some later installment of this series).

So that’s the plan. There’s a lot more to this of course: how will the device be made? What trade-offs are there in construction? How reliable are the results? How should the tests be done, with how many people, what questions should they be asked, and how should the results be analyzed? Why do this experiment at all? All valid questions which I will try to answer in future posts.