A contingency table may arise from an experiment (where one variable is controlled by the experimentor) or a survey (where there is no control over the individuals).
Example (from experiment)
To test whether vitamin C reduces the risk of catching a cold, a 1961 French study involved 279 skiers over two periods of 5-7 days. Skiers in one group of 139 were given 1 gram ascorbic acid (vitamin C) per day whereas those in the other group were given a tablet that looked similar but had no active ingredient (called a placebo). None of the skiers knew which of the treatments they had received.
Cold | No cold | |
---|---|---|
Ascorbic acid | 17 | 122 |
Placebo | 31 | 109 |
Example (from survey)
A health survey was conducted in the UK in 2000. The contingency table below gives the main reason for not using contraception by the 410 women aged 16-49 who were in a sexual relationship, not using contraception and not sterilised.
Age | |||
---|---|---|---|
16-29 | 30-39 | 40-49 | |
Partner sterilised | 6 | 81 | 127 |
Wants to become pregnant | 12 | 28 | 11 |
Pregnant now | 15 | 20 | 2 |
Menopause | 0 | 2 | 11 |
Possibly infertile | 6 | 18 | 19 |
Doesn't like contraception | 3 | 7 | 6 |
Other reason | 15 | 8 | 13 |