A confidence interval gives an estimated range of values that is likely to include an unknown populationparameter.
For example suppose a study of planting dates for maize, and the interest is in estimating the upper quartile, i.e. the date by which a farmer will be able to plant in ¾ of the years. Suppose the estimate from the sample is day 332, i.e. 27th November and the 95% confidence interval is from day 325 to 339, i.e. 20th November to 4th December. Then the interpretation is that the true upper quartile is highly likely to be within this period.
The width of the confidence interval gives an idea of how uncertain we are about the unknown parameter (see precision). A very wide interval (in the example it is ± 7 days) may indicate that more data needs to be collected before an effective analysis can be undertaken.