IN MID-1977, ATTITUDINAL QUESTIONNAIRES ON FOUR DIFFERENT TOPICS (HEALTH CARE, CURRENT ISSUES, QUALITY OF LIFE, AND TAXES) WERE MAILED TO SAMPLES OF THE POPULATION IN TWO LARGE CITIES AND ONE SMALL TOWN IN OHIO. THESE WERE COLUMBUS, CINCINNATI, AND NEW LEXINGTON. EACH QUESTIONNAIRE WAS PREPARED IN TWO LENGTHS--ONE PAGE AND THREE PAGES. EACH CELL (TOPIC X LENGTH X CITY) CONTAINED 50 NAMES SYSTEMATICALLY DRAWN FROM A LIST OF 400 NAMES IN EACH CITY'S TELEPHONE DIRECTORY. QUESTIONNAIRES WERE SENT WITH A COVER LETTER ENCOURAGING PARTICIPATION AND STRESSING CONFIDENTIALITY AND A SELF-ADDRESSED STAMPED ENVELOPE. RESULTS INDICATED THAT LARGE CITY DWELLERS WERE MORE LIKELY TO RESPOND TO A MAIL SURVEY THAN WERE SMALL TOWN DWELLERS. THE RESPONSE RATE OF AN ALLEGEDLY OVER-SAMPLED CITY WAS NOT SIGNIFICANTLY DIFFERENT FROM THAT OF ANOTHER CITY OF SIMILAR SIZE. THE QUESTIONNAIRE'S TOPIC AND LENGTH DID NOT AFFECT RETURN RATES, AND THERE WERE NO INTERACTIONS. A FOLLOWUP ANALYSIS INDICATED THAT THE EDUCATION AND INCOME OF NEW LEXINGTON RESPONDENTS WERE SIGNIFICANTLY LOWER THAN THOSE OF THE COLUMBUS AND CINCINNATI RESPONDENTS. IF THESE DIFFERENCES IN INCOME AND EDUCATION WERE ALSO CHARACTERISTIC OF NONRESPONDENTS, THEN THESE FACTORS MIGHT PARTLY ACCOUNT FOR DIFFERENCES IN RESPONSE RATES BY CITY. FURTHER RESEARCH IS NEEDED TO ESTABLISH FIRM CONCLUSIONS. TABLES, FOOTNOTES, AND REFERENCES ARE PROVIDED.