To compare an observed with an expected distribution, you need to know the total number of observations available, the observed values, the expected values and the degrees of freedom. For example, when you toss a coin, you'd expect a 50% chance of the coin landing heads-up compared to tails-up. An experiment would consist of a coin being tossed several times, with the outcomes recorded.
Chi-squared can also be used to assess whether or not two variables are independant, such as age and education. It can't be used, however, to investigate whether or not those who are younger have lower incomes,it simply states whether or not income is related to age.
