Two field studies used 2400 “lost letters” to test the hypothesis that return rates would be greater from smaller rural communities (population Ms = 1729) than from a city (population M = 60,591), unless the addressee was affiliated with a socially or politically deviant group. Addresses' affiliations were significantly associated with different rates of return in both studies. In Study A, community size and addressees' affiliations were significantly associated with different rates of return in the city and smaller rural communities, rates generally being greater in the city even when the people in need of help were affiliated with the deviant “For Prostitution, or Female and Male Sex Workers.” Destinations were not associated with return rates in either study. In Study B, community size and addressees' affiliation were significantly associated with different rates of return in only smaller rural communities. Return rates were lower than those from the city except when the person was affiliated with the deviant “NORML or Legalized Prostitution” addressees.

A study using 1,008 "lost letters" was designed to test the hypotheses that returned responses would be greater in smaller rural communities than from cities, that addresses' affiliation with a group opposed to physical education in schools would reduce the return rate, and that a person-positivity effect would influence return rates. Community Size and Person-Positivity conditions were not significantly associated with different rates of return even though returned letters from the smaller rural communities were more frequent than those from the cities across addressee conditions. Findings seem to suggest no community bias in the willingness of people to help a stranger in need by returning a lost letter. Person-Positivity conditions did not seem to improve the technique. Both addressee' affiliations alone and community size and addressee' affiliations were associated with significantly different rates of return in smaller rural communities, with rates of return greater in the smaller rural communities than in the cities (except for the addressee opposed to physical education in schools).

An analysis of letters placed in Mobile, Alabama and Sioux Falls, South Dakota (ns = 500 and 600) showed return rates among the several addresses were essentially uncorrelated (rs = -.11 to .08) so interpoll reliability is lacking.

No abstract

Szcmmay.-Two field studies were conducted in Escambia and Broward Counties in Florida. A total of 900 letters and 726 postal cards mere "lost" in selected voting precincts in these counties. More letters were returned than postal cards. The technique was not effecuve in reflecting gross differences of public opinion toward major and minor presidential candidates. County conditions seemed to influence the return of postal cards but not letters. The technique consistently did not predict trends in the outcome of the 1996 presidential election for major and minor candidates. Limited power may have been a Factor in the return mail associated wirh addressees' affiiadons, but less important when data for the letters and cards were compared with election outcomes. Lost Letter TechniqueIn 1965, Milgram, Mann, and Harter developed the lost letter technique to assess public opinion toward political groups and other institutions.