There are many discussions about the risk associated with radiation. I am not going to go into the various arguments about the effects of radiation here but try to explain risk and in particular risk due to exposure to radiation.

There are several different definitions of Risk, but it can be basically understood as the probability of hazard happening times the severity of the event.

An easy way to understand it is to see that, for the same risk, aeroplane or train crashes must be much less frequent than car crashes since many more people will die.

Risky Jobs

All jobs have some risk of serious injury or death. For the moment, let’s talk about the risk of death.

Industry	Deaths per 100,000 per year
Deep sea fishers	116
Taxi drivers and chauffeurs	16.1
Construction Labourers	15.6
Electrical power-line installers and repairers	16.1

Data from: http://www.businessinsider.com/most-dangerous-jobs-2011-9

I pulled this data off the web rather than using completely made up figures. However, when presented with data like this, we must first think about it. Are the figures derived from just one year? If so, a bad deep sea accident could distort the figures. If the industry contains lots of workers (e.g. construction) then the figures would have higher significance than that say of astronauts (see Probability and Significance).

Anyway, on with talking about risk. These figure would allow you to estimate the number of deaths in a year. If we took a sample of 100,000 construction labourers, then we would expect something like 15-16 to die due to an industrial accident. If there were one million then there would be ten times the number of deaths – i.e. 150-160.

It is possible to express this risk as the risk per person per year. If we have a risk of 116 per 100,000 per year, this is 0.00116 (or 1.16×10^-3 – if you don’t understand this, see here) per person per year.

Just because the risk is expressed per person, beware of interpreting this incorrectly. Individuals could have risk that are very different from the average. That risk may change with age or experience. There may be other factors that are important – such as deep sea fishers who can’t swim. What this risk measures an average over many people.

Population Exposed To The Risk

A very important factor is how many people are exposed to a certain risk. Let us say that there is a risk of dying due to a particular cause of 1×10^-3 (0.001) per person per year.

If 10 people were exposed to that risk then the risk for all ten would be ten times this amount i.e. 1×10^-2 per year – that means we could expect one death every 9 years or so (however, only if we average out over many, many years). We might think that was an acceptable risk.

However, what if the number of people exposed to this risk was one million. That would be 1000 deaths every year. If everyone in the UK was exposed to this risk then it would be 62,000 deaths per year.

The point here is that a small individual risk that people would be willing to take in a dangerous job can lead to an unacceptable number of deaths if applied to a large population. Individual risk and collective risk are very different.

How many deaths would you expect in the UK? Well the death rate is 1 in 100 so you would expect about 620,000 deaths per year. So our 62,000 a year from the ‘risk’ would only add about 10%.

OK, now what would happen if the people did not die instantly but the risk caused deaths over the next 10 or even 40 years. We are now looking at a less than 1% increase in the death rate. Would you be able to detect that above the random fluctuations in the normal death rate?

Radiation from Radiation

At high doses the effects of radiation are reasonably well known (see What Is A Sievert?). However, for low radiation doses, the effects are much more difficult to quantify. The reason being that the deaths are not instant and take place over many years. Any increase in the number of deaths is therefore very difficult to quantify. However, if they are over a large population, then they are very significant.

I am not going to go into the controversy surrounding the effects of radiation into any great depth here. It is a very complicated subject. However, if you want to find out more have a look at the Low Level Radiation Campaign. This has a lot of background information on the issue. However, I do not fully endorse everything on this site.

What I am going to do now is to use the ‘official figures’ for the biological effects of radiation. These tend to be based on studies of the effects of the atomic bombs at Hiroshima and Nagasaki. These are limited due to several reasons, including:

• The studies were started many years after the bombing.
• They are based on a short exposure and cannot take into account the effects of exposures over a longer length of time.
• They cannot distinguish between external exposure to radiation and the effects of breathing in or eating radioactive material.

The official figures are based on the following:

• The effects are due to the dose received (measured in Sieverts). This can be a large dose for a short period or a smaller dose over a longer period.
• The increase in risk is 0.05 per person per Sievert. This is from the International Committee on Radiological Protection (ICRP).

A definitely do have problems with these assumptions and figures. However, my main aim here is to give you an understanding of risk.

Let us take an example of 100,000 people exposed to 20mSv. That would result in 100,000 x 20/1000 x 0.05 – that is the number of people times the radiation dose in  (1Sv = 1000mSv) times the risk factor. This leads to excess 100 deaths. Note that this is for a single dose. If the population of 100,000 was exposed to this every year then it would be about 100 deaths per year.

Your individual risk is only 0.1%. I will state again. Individual risk and collective risk are VERY DIFFERENT. These would be 100 real people suffering and dying – they are not just statistics. I think that these lives do matter. Mark Lynas does not (here) – he also gets the science completely wrong.

If you used a similar argument for murder, then you could say that we do not need to worry about it since it does not produce a significant increase in the overall death rate – 619/608,146 = 0.1%.

I will write more on this in future posts.