Not A Normal Distribution

So, I’m feeling grumpy today. Just am. Not sure why. And then I go and stupidly read an article in on the BBC News website about most babies being born at 4am. You can read the article here and I’ll be taking quotations from it.

Four o’clock in the morning is the time most babies are born spontaneously in England, with the majority arriving between 01:00 and 07:00, a study shows.

So, four o’clock is the “average” of 01:00 and 07:00. Is that how they worked this out? I’m not sure. I’m probably not going to read the original paper but it does seem a strange thing to say. Assuming that spontaneous births are reasonably random the actual number arriving at 04:00 would be quite small and the variation could be little.

While planned C-section births tended to happen on weekday mornings, births after induced labours were more likely to occur around midnight.

This makes sense. C-Sections in the morning to then make sure everyone is fine during the time when the medical cover is best. Inductions probably happen at that time because they are started in the morning and it takes that long for the effects to show. These are not surprising figures and make utter sense.

It’s this next bit that turned me into a rage.

Overall, more than 70% of births took place outside regular working hours.


The researchers said there could be implications for staffing of midwives and doctors, with only 28% of births taking place between 09:00 and 17:00 on weekdays.

So, an event that is reasonably random is spread throughout the day? About thirty percent of births happen in a time frame covering roughly 30% of the amount of time in a day. THIS IS WHAT WE WOULD EXPECT. The working day covers 1/3 of the 24 hours in a day and the number of births in that time period is about 1/3. This sentence made me swear out loud in our staff common room earlier today, much to the ammusment of my colleagues.

I’m going for a walk to calm down.


So, there’s been a shooting in the USA. There are always going to be shootings in the USA while guns are so easy to access. It’s really that simple. If you make guns easy for any idiot to purchase and use then they are going to use them. There should be a cooling off period and background checks. Oh, that and Mr Trump-Cunt removing the rules about the mentally ill buying guns.

It’s important to make sure that when news is presented that it is accurate and (preferably) unbiased. Even BBC R4 annoys me nowadays with John Humphries’s leading questions and stupid un-scientific knowledge. If I can shout at the radio “irrelevant” or “leading” or “biased” then the interviewer isn’t doing their job well.

I screen grabbed this from twitter:

News Issues
News Issues

I am not trying to point out that there haven’t been enough school based shootings, but I do want to point out that if you are trying to make an argument then you should be correct in your figures. I do not know what the actual number is for school shootings. Honestly, the fact that there is a number for JUST THIS YEAR means something is fucked. But all news people surely have a duty to represent the truth as best as they can find it. It doesn’t take much research to get that the 18 figure is misleading and that’s when the NRA will fuck you over.

Testing Heaven And Hell

In my last communication I stated that I thought the difference between my two relative frequencies was not significant. I think I have now performed a test, but whether it is valid or not can be up to you.

If Heavy Metal focuses on the dark side of life then there should be many references to to Hell in the song title. Conversly, there should be only a few references to the term Heaven in song titles. If this is the case then it could be expected that the frequencies of each would be independent of each other, assuming the subject matter for songs is random and the use of words follows this.

My Calculations follow, but essentially I have found that, at a 5% significance level, the distribution of Hell and Heaven is random and that the frequency of hell is close enough to the frequency of heaven for this to be so.



I am fully aware that I have really pushed the limits of my significance test and it probably doesn’t even work properly for this type of problem. But, my happiness about this communication remains about 75% (+- 3% for 95% confidence level).

BBC Headline #5

BBC Headline from the website taken today:

Lagging pupils “don’t catch up”

This headline is lacking and, to be honest, the whole article is shocking. Headline problems are:

Quotation in Headline
No Shit Sherlock
Problematic Assumptions

Quotation in Headline
As long as someone wrote it or said it you can include it as a quotation in any headline or article. Say what you want. There’s always some nutter willing to give their opinion to give your leading headline some weight. “Crystal energies healed me” or “watch out for 23 December 2012! Those Mayans knew a thing or two”.

No Shit Sherlock
Pupils who are lagging behind in their work and understanding don’t then go on to catch up. Really! I need a whole BBC Headline to know this? How about “Some schools do really well!” or “Pupils getting better grades” or “Some schools not as good as others!”. There’s a distribution of schools or pupils, you can’t measure everyone and have everyone above average.

Problematic Assumptions
The biggest issue with the article and what the headline implies is that the bottom few pupils as measured by some arbitrary government test do not proceed to do well as measured by some other government arbitrary process. Have these people never heard of the Gaussian Distribution (the bell curve or normal distribution)? Some pupils will always be behind the others and will probably continue to be behind. Elsewhere in the article it is claimed that the top performers go on to get good grades later on. Holy Cow! This curve needs to be explained to them.

This is a graph of the Gaussian Distribution as everyone sees it:

Bell Curve

The Gaussian Distribution as the government sees it (blue version):

Bollocks curve

No one is allowed to fail or fall behind or not be clever or be too far from the mean.