Day 1

Morning - thinking about sampling

• 10.00 - 10.15 - Are there too few Brexiteers in the survey? How would we decide?
• 10.15 - 10.45 - another problem of proportions. If a family has 4 children, what is the probability that the family has 3 girls? A simulation.
• 10 minute break
• 10.55 - 11.15 - what is the algorithm?
• 11.15 - 12.00 - the machinery we will need: Print, If and For.

Afternoon - probability and simulation

• back for for loops in Print, If and For.
• lists in Lists.
• solving with simulation; Three girls;

Notebooks from the afternoon:

• riffing_on_if_for_and.ipynb;
• riffing_on_lists.ipynb.

An explanation of and and or: and, or in expressions.

Homework

• Download the notebook from the link at top of Three girls - it’s a slightly modified solution to the three girls problem we were working on today. Make a copy, that you are going to modify.

• Modify the notebook to give the probability of 2 girls, instead of 3 girls. What is that?

• Make another copy of the original Notebook for the next problem. The actual probability of having a male child is not 0.5 but 0.513. [1]. To do a better simulation, we need to modify the Notebook to give a 0.513 probability of "B" and a 0.487 probability of "G". You can’t easily use random.choice for this, but you could an expression like this to decide if it is a girl:

  random.uniform(0, 1) <= 0.487
  

  random.uniform(0, 1) returns a random number between 0 and 1 - so the number it returns will be <= 0.487 about 48.7% of the time.

  Modify your new copy of the Notebook to use something like this to give the probability of a girl. You should find that your estimate of the proportion of 3-girl families drops slightly, because girls are slightly less common than boys.

[1]	Official UK government statistics give the birth ratio as 105.3. This the number of boys born for every 100 girls.