6th February 2018
It might be a bit cliché but I love maths. Especially geometry, although I enjoy dabbling in all aspects of mathematics. Currently I am reading a book which involves the mathematics behind statistics and how they are used. This got me thinking about where statistics are used when discussing Autism. This includes news stories, social media feeds, Autistic charities/research websites, books about Autism and Autistics groups to name just a selection of places where you might find statistics relating to Autism and Autistics.
So to begin with what is meant by a statistic? A statistic is a piece of data that is obtained from a source that is used to establish certain facts. This fact can then be used as evidence to bring about change, prove certain truths or determine a direction for a course of action.
For example, a well known Autistic statistic from the National Autistic Society research is that only 16% of Autistic people are in employment. 16% is such a low figure and it proves that more Autistic people should be in employment. Therefore this statistic from the research indicates that change is needed. The National Autistic Society can then produce campaigns around this figure to hopefully change this statistic and improve the situation.
Where do statistics come from? As mentioned above, research. It is important though to be sure that the statistics you are reading are non-biased and the research that has generated the statistic is conducted correctly.
For example, I once attended a talk about Autistics in the workplace. The talk was supported with a lot statistics about Autistic people in various situations. These statistics were taken from a questionnaire. One of the questions asked was “Are you happy in your current employment?” and the statistics showed that a large percentage of Autistics were not.
However, the question was only asked of Autistic people. There was no control group. I know many Neurotypicals that would have answered “no” to that question. The statistic could be the same for both Autistics and Neurotypicals. Without a control group, in this case Neurotypicals, the statistic loses its strength. You can question if it is valid or not.
Therefore this warns us that statistics cannot always be trusted. When asking a group of any people questions for statistical research, the selection of people is important. It is well stated that all Autistic people are different. We exist on a spectrum and are all uniquely different.
If you wanted to know if Autistic people like playing cards, how many do you ask? If you ask 20 and 9 say “yes” then you could use the statment that “45% of Autistic people like playing cards”. If you ask 200 people and only 37 say “yes”, the statement changes to “19% of Autistic people like playing cards”. In summary, the greater range of people questioned, the more accurate the statistics will be.
Averages are another interesting aspect of how statistics can be manipulated. In the book I am reading, it states that when taking an average most people use the mean method. This is when all the numbers are added together and then divided by the amount of numbers. Let’s say we asked 5 Autistic people how many hours they spend a week on their special interest. The results are as follows 3, 12, 23, 18, 70 hours. The mean average is (3+12+23+18+70)/5, so the answer is 25.2 hours.
However, there are other ways to take an average, such as the median average. Here the average is obtained by arranging the results in numerical order and then taking the middle value as the average. So 3, 12, 18, 23, 70 are the hours spent on special interests in numerical order. Here the average is 18 hours.
So as you can see the two results are very different. Therefore you can choose which average figure to use to your advantage. If you want to show that Autistics spend more time in a week on their special interests the figure to use would be 25.2 hours on average. A total of 7.2 hours more than using the other figure. Neither is wrong, but it shows how statistics can be mathematically manipulated.
Visual statistics can also be manipulated too, to show biased results. Graphs for example, normally have two scales. One on the x axis and one on the y axis. When figures are plotted between the two axes a graph is drawn from the data. The scale on one axis can be intervals of 1 unit or intervals of 10 units. Depending on which interval is used the resultant graph could look very different indeed. Yes, it’s confusing, but all you have to remember is to ask yourself “How is this data being presented to me?”
Other things to be wary of is who is conducting the research too. Larger organisations like the National Autistic Society are more likely to have accurate statistics then smaller groups. This is not to say that smaller groups of research are not just as valid though. It depends on how it is done and why. It is up to you as individuals to decide what you make of any statistic presented to you. Just bear in mind some of the points made here and never be afraid to question what is being shown.