Here’s a long explanation.
This is an example of a 2x2 contingency table, but I can’t show this exactly because there are formatting limitations.
As 1/5 of the guests don’t like coffee, the total number of guests must be a multiple of 5, as we can’t have fractions of guests. Similarly, as 1/4 of the guests don’t like tea, the number of guests must be a multiple of 4. As 4 and 5 don’t have any common factors, this means that the number of guests must be a multiple of 4x5 = 20. Let the number of guests be 20k, where k is an unknown integer.
Then 16k guests like coffee and 4K guests don’t like coffee. 15k guests like tea and 5k guests don’t like tea.
As 3 guests don’t like both tea and coffee, this means that 5k-3 guests don’t like coffee but like tea, and 4K-3 guests like tea but don’t like coffee.
Therefore, (5k-3) + (4k-3) + 3 guests don’t like either tea or coffee or both. Subtract this from the total number of guests 20k and you get 11k + 3, the number of guests who like both tea and coffee.
As this includes k, there is an infinite number of possible solutions. For example, if k=1, there are 20 guests, of whom 14 like both coffee and tea, 2 like coffee but not tea, 1 likes tea but not coffee, and 3 don’t like either.