Left handed birthday dates – Results
We recently asked Left Handers Club members to let us know their date of birth, sex and country so we could do some analysis to see if left-handers differ from right-handers in their birth month.
Thank you to all who contributed. After removing duplicate entries and invalid data, we had over 3,100 entries which makes this by far the largest survey of its type ever analysed!
I think this can be considered a pretty random sample of birth date data for lefthanders (though it is skewed towards females because that is the balance of our customers and newsletter readers). I have done some initial analysis that DOES seem to show a difference in seasonality for left-handers. I have compared my left-handed results with overall birth rate statistics for UK, USA and Australia for 2010 (which do all seem to follow a pretty similar pattern) and charted the results.
The graph below shows the percentage of the annual total births that occur in each month. There do seem to be significantly higher births of lefthanders in July and August than expected from the normal overall trends. There is also a smaller but still significant-looking peak for lefthanders in January and February.
The patterns for male and female left-handers are very similar except for two marked differences:
- Less males in Jun
- More males in August
I did try to analyse by decade but there seems to be a pretty wide variation in the data and as the sets for each decade are much smaller I am not sure how valid it is. The same goes for analysis by region and North /South Hemisphere.
To get some expert input into the analysis and a review of the data, I sent the information and my analysis to our friend Chris McManus, Professor of Psychology and Medical Education, University College London and author of the excellent book Right Hand, Left Hand, and this is what he had to say:
Suddenly seasonality [of left handed births] seems to be of interest to people again! I've had two other things recently on it. The data are all a bit variable, which is slightly worrying.
Your data certainly seem to show some non-randomnesses (and I did the chi-square test on your data to check on that), and you were right to compare with population birth rates (and there is some seasonality in that). The problem, potentially, is that on the website you flag up the dates you are interested in (and you particularly mention August which then is way above all of the others…). In psychology we would probably regard that as a leading question, and we would worry about it. Ideally one wants data from people who don't know what the hypothesis is. The risk otherwise is that people who think they are interesting tend to reply preferentially. (Many twin studies have that problem as identical twins who are more similar e.g. same handedness, are more likely to reply than, say, non-identical twins with different handedness).
I did try removing the August data, and then the effect is of very marginal significance (not at all against flat, and only just there in the other groups).
Looking at the sex ratios by month was a good idea, as these should really be pretty stable throughout the year. They don't seem to be, though, with that funny bump of males in August, which doesn't seem very likely biologically. Maybe males are more likely to try and help by responding if their birthday is in August.
So, difficult to know what to make of it, but very impressive that you can get so many participants so quickly. What we really need is a question on handedness in the census. That would answer it once and for all!
All best wishes as ever, Chris
Following Chris's comments, we did look more closely at the data and there is indeed an unnatural looking peak of birthdays on 13 and 14 August (more than twice the normal peak daily levels). We do not have any explanation for that apart from Chris's suggestion that we were ” Leading the witnesses” and it is more likely that people with birthdays on 13 August would respond to the survey as we had mentioned that as a significant date.
Our analysis of the data DOES seem to show some difference between the month of birth for lefthanders and the general population. However, as is often the case with research and analysis, it could well be that our data is skewed rather than this being a real underlying effect. But with such a large sample (compared to all the other samples that have ever been analysed) and with some clear differences that are NOT related to 13th August as well as the “Left Handers Day” effect, there may really be something to this.
That said, we cannot offer any theories as to WHY that might be the case so, although it has been an interesting exercise, we do not really feel able to draw and stong conclusions from it.
Do you know better?
If this is you area of expertise and you have any comments, or you would like to do your own analysis of the data, we would be very pleased to hear from you. You can add any thoughts as a comment below or you can find an anonymised version of the cleaned data here (it is in an Excel spreadsheet with a Pivot Table set up and also a linked sheet of country / region data and national birth rates data that you may find useful).
Thanks again to all who took part and we hope you found this interesting, if inconclusive!