Issue 7 - February 2001
The three experiments were conducted by a total of 1,005 pupils in the UK (694 in primary schools, 311 in secondary schools)
Experiment 1: Does buttered toast tend to land butter-down ?
To find out if this 12 per cent difference is statistically significant, we can calculate its "p-value" - that is, the chances of getting at least as big a difference as that observed just by fluke. This probability turns out to be extremely low (p ~ 10-127) Thus the 62 per cent rate is both substantially, and statistically significantly, different from the rate expected by chance alone.
Experiment 2: Is butter the cause ?
Again, statistical significance tests against the null hypothesis of a 50:50 split gives an extremely low p-value (~ 10-57). Toast with just a "B" on it thus still lands face-down at a rate substantially, and statistically significantly, different from value expected by chance alone.
Experiment 3: Is height the cause ?
Both these differences are statistically significant (p = 0.003 and 10-37 respectively).
So increasing drop height to 2.5 metres gives a substantially and statistically significantly higher probability of producing a butter-up landing.
|page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | credits|