Indoor 222Rn concentrations in a probability sample of 43,000 houses across 30 states.

The U.S. Environmental Protection Agency has assisted 30 of the 48 conterminous states in completing statistically designed surveys of indoor 222Rn over the past 4 y. In all states, the lowest livable level of 43,054 randomly selected houses was tested using charcoal canisters exposed for 48 h. The sampled population included owner-occupied ground-level houses having listed telephone numbers. Summary statistics along with the percentage of houses exceeding various concentration levels are given by state and over all states for houses with basements, for houses without basements, and for all houses. As expected, 222Rn concentration varies widely from one state to another and, in every state, basement houses exhibit higher concentrations than nonbasement houses. The lognormal distribution is shown to be a good approximation to the distribution of screening measurements over the 30-state area. There is, however, some evidence that the lognormal distribution underestimates, by a narrow margin, the upper tail of the observed distribution of basement measurements.

Multi-state surveys of indoor 222Rn.

The U.S. Environmental Protection Agency provides assistance to states in conducting surveys of indoor 222Rn. Alabama, Kentucky, Tennessee, Wisconsin, and Wyoming completed statistically designed surveys during the 1986-1987 heating season. In all states, probability-based samples of 5800 houses were tested using charcoal canisters exposed for 48 h. Thus, sample results can be validly extrapolated to the target population of all owner-occupied, ground-level houses having listed telephone numbers. Estimates of population parameters (e.g., median) and their associated confidence intervals are given for each state, for geographic regions within states, and for basement and nonbasement houses. Results confirm that 222Rn concentration varies widely from one state to another and from one geographic region to another within a state; however, the same pattern of differences was not evident in basement or nonbasement houses. Short-term 222Rn readings appear to follow a log-normal distribution.

Effects of an advance telephone call in a personal interview survey.

During October and November 1982, 1,260 Medicare-eligible senior citizens were interviewed in a survey focusing on health care of the elderly. As part of the survey, an experiment was conducted in each of the three survey sites to determine the effects of an advance telephone call to schedule an appointment for a personal interview. One random half sample in each site was sent a lead letter, followed by a telephone call to schedule a personal interview. The other half sample was sent a lead letter followed by a personal contact, with no intervening telephone call. Telephoning to arrange an appointment for a personal interview resulted in a 20 percent saving in data collection costs with only a 1 percent decrease in response rate.