A parameter is a numerical summary of a population, but because populations are so large, it is impossible to get a true numerical summary of a population (Population, 2004). For example, it would be impossible to know the average earnings of all secretaries in a given year. This would require identifying, locating, and questioning millions of secretaries who worked that year and verifying their earnings. However, it would be practical to select a sample of 1,000 secretaries across the country, using randomization, replication, and blocking techniques, and then compute an average of the earnings of these 1,000 secretaries. Such a sample would constitute a statistic. A statistic is a numerical measure computed from a subset of the population.
In another example, it would be impossible to count accurately how many nails were sold by hardware stores in this country in a given year (a parameter). However, it would be possible to sample several hundred hardware stores across the country and compute an average of the number of nails sold in a year. This would be a statistical average. Political opinion polls are examples of statistics; market research surveys use statistics to assess how many people are likely to buy a certain product. The Census is a parameter because it actually counts (or attempts to count) every person in the country.
Population parameter vs sample statistic. (2004). Retrieved at:
http://www.math.sfu.ca/~schwarz/Stat-01/Handouts/node29.html
...