# WITMIT? Neckless Beads : Letter Distribution

For my daughter’s birthday, one of her gifts was a beaded neckless set that includes several beads with letters on (see the picture). The idea is that you can make them with your friends’ names on them.

Naturally, I was curious about the distribution of letters. Does this distribution tie in with the most commonly used letters in the English language? For example, would it tie in nicely with the old Read All About It coursework or an investigation about Scrabble scores?

[[posterous-content:pid___0]]Image and points scores (below) from Wikipedia.

*1 point*:

**E**×12,

**A**×9,

**I**×9,

**O**×8,

**N**×6,

**R**×6,

**T**×6,

**L**×4,

**S**×4,

**U**×4

*2 points*:

**D**×4,

**G**×3

*3 points*:

**B**×2,

**C**×2,

**M**×2,

**P**×2

*4 points*:

**F**×2,

**H**×2,

**V**×2,

**W**×2,

**Y**×2

*5 points*:

**K**×1

*8 points*:

**J**×1,

**X**×1

*10 points*:

**Q**×1,

**Z**×1

More relevant to this specific product is what is the distribution of common letters in common first names.

I can see this being turned into a lesson along the premise that a gift company wants to make a product like this and wants advice on how to select which letters to put in. There’s a fairly obvious survey to be done here. I’d not show them the picture in the first instance and see how closely it matches their answer.

Does it matter that this product is mostly aimed at girls?

Could favourite colours be considered?

As well as surveys it might be an idea to talk nicely to your information systems manager to see if they could give you an excel list of all the first names at your school, connected to the gender to give you a much larger sample. I did this to get an interesting Chi squared goodness of fit example for the days students were born on for A-level statistics.Gary

Having a ‘population’ census seems like a brilliant idea.What were you looking for with your Chi-Squared test? You mike like to click my tag – statistics to see some of the other things I’ve done including A level things.ThanksDave

We were testing the hypothesis that the day of the week you were born on would be uniform. Tied it in with Zeller’s algorithm (nice if you are doing D1 as well) to calculate which day of the week they were born on and I’d assumed that there would be a fairly even spread across the days. Including the whole college’s data it turns out it is far from it (certainly for our students anyway), with statistically significantly fewer on the weekends and peaking on Wednesday. I have yet to find national data, but if you’ve got access to a large list of dates of birth I’d be intrigued if you see the same phenomenon. My students had several theories including hospitals inducing births, and trying to count 9 months back from a Wednesday!Gary

Hi – I had assumed you were expecting uniform distribution but wondered if there was something else. Interestingly, my first child was Caesarean and therefore not scheduled for a weekend.I’ll see what I can do about a large data set…..Thanks againDave

I wonder how long it will be before you ‘steal’ that set to take into school!This prompted me to have a rummage and I found this amazing resource of data on children’s names on the ONS website:http://www.statistics.gov.uk/StatBase/Product.asp?vlnk=15282&Pos=1&ColRank=1&Rank=240I rustled up a quick Excel spreadsheet using the 2010 Girls Names which gave a rather different set of letters required than your daughter received!http://www.singinghedgehog.co.uk/misc/Girls Names Letter Distribution.xlsxWould be reasonable to expect pupils to do much of this themselves?

Aarrgghh!! Link to my file doesn’t work! Try this one with no spaces:http://www.singinghedgehog.co.uk/misc/GirlsNamesLetterDistribution.xlsx

As if I’d steal my daughter’s toys! ;-)Cool spreadsheet – did you type all the data in yourself or is there a quick way of getting Excel to do it for you?A few more things spring to mind: a) Students could look up the list for themselves, then we could build a Google Docs sheet that they could fill in between them.b) The list doesn’t take into account the possibility of knowing multiple people with the same name.These are all things that make the investigation more interesting and ‘real’ in my opinion.thanks for your contributions.Dave

I would agree that they should find the data themselves, with perhaps steering towards the ONS site; they could decide which year would be most relevant to their investigation. I also like Gary’s suggestion to get the school name distribution from ICT support [how much is that an oxymoron!]; secretaries should be able to pull that out very quickly into an excel file.The spreadsheet was very quick to do: I simply opened up the chosen ONS file and copied the names into a blank sheet. I used a cunning formula to do the letter count and copied it along. Basically it counts the word length, takes out the letter you are looking for, counts again and finds the difference. The totals at the bottom again have a simple formula dividing the number of that letter by the total number of letters and multiplying by the number of beads in the box.I have found that the most important trick to show the pupils is the use of the $ sign to ‘fix’ rows or columns in formulae before copying them around a sheet.