One-page revision sheet

• A statistical question is one whose answer varies, so you need data to answer it.
• Write the question and your prediction first — then you know what data to collect.
• Primary data you collect yourself; secondary data somebody else collected.
• Categorical data is words. Discrete data you count. Continuous data you measure.
• A tally chart counts reliably in fives; the totals become the frequency column.
• Group continuous data into class intervals that do not overlap and leave no gaps: 150 ≤ h < 160, then 160 ≤ h < 170.
• A fair question is neutral, and its answer boxes cover every case exactly once.
• A prediction that the data does not support is still a real result. Report it.

Key words: data  ·  statistical question  ·  prediction  ·  primary data  ·  secondary data  ·  categorical data  ·  discrete data  ·  continuous data  ·  tally chart  ·  frequency  ·  class interval

• The population is everyone you want to know about; the sample is who you actually ask.
• A census asks everyone — most reliable, rarely practical.
• A larger sample size generally gives a more reliable result, because one unusual answer matters less.
• A larger sample does not fix bias. Bias is about who is chosen, not how many.
• Random: everyone has an equal chance. Systematic: a fixed rule such as every 10th. Convenience: whoever is nearest — usually biased.
• Ask where you are sampling from: a bus stop samples bus users, a football club samples football fans.
• To estimate a whole population, find the proportion in the sample and apply it: 24 out of 60 is 0.4, so 0.4 × 900 = 360.
• Always state your sample size when you report a result — it tells the reader how much to trust it.

Key words: population  ·  sample  ·  census  ·  sample size  ·  random sample  ·  systematic sample  ·  convenience sample  ·  bias  ·  representative