NEW YORK — When Burns W. Roper started his job asking people their opinions in 1946, he wasn't employed by a corporation. His boss was his father, Elmo.

As Roper matured, opinion research grew into an industry, and the young interviewer who left Yale before graduation to be a poll taker eventually became chairman of the Roper Organization. By the time he began his retirement April 30, the family business had evolved further, merging with another company and being renamed Roper Starch Worldwide.

But this tidy summary of Burns Roper's career is misleading.

Roper retired only in theory. His was never really a family business and, aside from embracing computer technology, it has not changed that much in four dozen years.

Roper, 69, sat among packing boxes in his Midtown Manhattan office recently to look back on his 48-year career. It was clear that despite criticism of surveys and their methods, he still believes that polling people is a viable and valid process.

"I've said many times that when I see a poll result and it's opposite to my personal feelings, my first instinct is to re-examine my personal feelings," he said.

Pundits sometimes scoff at polls for asking people simple questions about complicated public policy. But Roper argues that the American public is very observant and wise, and its opinion is valuable to marketers and politicians alike--if they study it in enough depth and report the bad with the good.

The media and decision makers have grown more reliant on surveys over the years. That has helped the Roper Organization, and now Roper Starch, grow.

Roper surveys have been used by clients like The Wall Street Journal and Worth magazine, which have used poll findings to produce feature stories; by Good Housekeeping to sell advertising; by Philip Morris to link its Virginia Slims cigarette brand to a major study of women's changing roles in society, and by Warner Lambert to study youth attitudes.

In a highly competitive field, Roper Starch has some distinguishing features, including the fact that it still does a substantial amount of polling door-to-door in an era of telephone interviewing.

Roper's interviewers, 95% of them women, knock on doors in scientifically selected blocks. The main advantage for the personal contact is that it gives a poll taker the ability to show respondents pictures or text that is too long to read over the phone.