HAS THE American family collapsed? To read some commentators, you would think so.

A few months ago a review in Book World began with the assertion, "The family is dissolving . . . two-parent homes are increasingly scarce."

Two years ago a respected nonprofit educational organization published a fact book on demographics which, as a central point, suggested that the traditional mom-and-pop American family had all but disappeared in the brief space of 30 years.

"In 1955," the publication's author, former National Institute of Education Director Harold L. Hodgkinson wrote, "60 percent of the households in the U.S. consisted of a working father, a housewife mother and two or more school-age children . . . . In 1985 it is 7 percent, an astonishing change."

The figures have been widely cited at conferences on current social problems. If correct, the numbers would signal that the family is already finished as a major American institution -- the victim of divorce, the rise of the one-parent family and a growing tendency among mothers to take jobs outside the home.

These numbers, however, are wrong. True, the U.S. trend is toward fewer two-parent families. But the sentimental notion that most households a generation ago consisted of a married couple with two or more school-age children and a wife who did not work outside the home is simply incorrect.

The "traditional" family cited did not come even remotely close to constituting 60 percent of U.S. households in 1955, presumably its heyday. Even then, there were millions of single people -- both young and old -- living alone in one-person households. More than 5 million married mothers were working. Many married couples had no children under 18 or only one child.

According to Census Bureau figures, married couples with two or more children and a non-working mother constituted at most 22 to 23 percent of all households in 1955. A slide from 23 percent to only 7 percent in the proportion of households made up of "traditional" families is noteworthy enough, but it obviously doesn't imply the same social disintegration as a slide from 60 to 7 percent.

When myths like this gain currency they can interfere with informed policy decisions by public institutions and the public at large. Unfortunately, quite a number of false notions about poverty, income changes, the family structure, infant mortality are frequently heard or seen in the media, at policy conferences or in advertising.

Here are some social myths that this reporter frequently encounters, followed by the facts as provided by the Census Bureau and the Labor Department's Bureau of Labor Statistics:

Myth: "The old-style family with ahusband and wife is an endangered species."

Fact: More than 165 million people, or 71 percent of the entire population of the United States in 1985, lived in households headed by a married couple. That is down from 84 percent in 1955, but still a big majority. These households include: about 50 million married couples, some young, some old, some with children, some without; about 47 million children under 18; plus older children and other relatives living in the same home. Paul Glick, former Census Bureau expert on the family once said, "Americans like to live in the married way." That's still true.

Myth: "Most kids live in broken homes."

Fact: In 1985, 78 percent of all families with children of their own were still two-parent families. Almost 75 percent of all children under 18 lived with both parents (in most cases their own biological parents) in 1985. That is down from 91 percent in 1955 but it is still three-quarters of all families with children.

Myth: "Nearly all women are working these days, even mothers of young children."

Fact: Close to the truth but overstated. There has been a sharp increase in working mothers over the past generation. In March 1986 about 70 percent of all mothers with children ages 6-17 were in the labor force and about 54 percent of women with children under 6.

But in many cases, the work was only part-time or part-year, not full-time all-year work. According to statistics developed by Child Trends, a nonprofit research organization, in 1984 only 27 percent of children under 18 had mothers who worked full-time all-year-round. Another 37 percent had mothers who worked part-time or part year. The remainder had mothers who didn't work at all.

Myth: "People are worse off economically than 10 or 20 years ago."

Fact: Economic growth has slowed very considerably over the past decade or so, but the statistics suggest that Americans are much better off economically than 20 years ago and at least as well off as in the mid-1970s.

Census figures put the money income of Americans at $7,939 per person in 1967, measured in constant 1985 dollars. It rose to a high of $10,644 by 1978 before declining in the recession of the early 1980s. But with recovery, it climbed back to $11,013 in 1985. Within this overall average, the 1985 figures for whites ($11,671) and blacks ($6,840) were both at their top levels of the past 20 years.

Some economists prefer to use family income rather than per capita income in measuring economic gains or losses. Family figures do not include the income of single people living alone or with other unrelated people, including many aged persons whose Social Security benefits improved markedly over this period. These family data do show some losses. For 1970, Census figures measure average family income in inflation-adjusted 1985 dollars at $30,768. By the peak year 1978 it had jumped to $33,129 before dipping in the early 1980s, and then recovering by 1985 to $32,944.

However, the size of families was declining over this period. In 1970 the average family had 3.57 members. By 1985, it had only 3.21 members. This means that income within families was $8,618 per person in 1970, $10,008 in 1978 and $10,262 in 1985 -- statistically a slight gain. These figures, of course, are an average of all families. Some (older, more well-to-do) families had larger gains, while many younger families with children fell below the 1978 level.

Myth: "Poverty is worse than ever."

Fact: The proportion of Americans with cash income below the government's official poverty line was 22.4 percent in 1959, the first year for which figures were computed. The percentage dropped to an all-time low of 11.1 percent in 1973, went up a few tenths in the next few years, then rose sharply in the two recessions of the 1980s, reaching a high of 15.2 percent in 1983. But with recovery the poverty rate edged down again to 14 percent in 1985. Thus, poverty is clearly not worse than ever.

Still, more Americans are in poverty today than in the mid-1970s. The reason is that income gains in recent years have mostly gone to people in middle and upper income brackets, while the proportion of income received by the lowest fifth of the population has dropped.

Myth: "Black female-headed families with small children make up the bulk of the poor."

Fact: In 1985, about 33 million people had cash incomes below the government's official poverty line ($10,989 for a family of four). Two thirds were white. Of all people in poverty, about 13 million were in married-couple families. About 6 million people were in white female-headed families, some of which had no children, and only about 5.3 million people were in black female-headed families, some of which had no children.

Myth: "Infant mortality has gone up."

Fact: Infant mortality rates in the U.S., while still higher than in most industrialized countries, have declined steadily from 29.2 per 1,000 live births in 1950 to 10.4 in 1986. The rate of decline, however, has slowed considerably in the past few years.

What causes these misconceptions? In some cases, they originate in the misuse or misunderstanding of statistics; in other cases, they arise from simple statistical errors -- the Institute for Educational Leadership's error seems to have come from using the wrong denominator for its calculations. Once in currency, however, faulty statistics seem to take on a life of their own, gaining momentum from the tendency of the public, once it becomes aware of a trend, to imagine the trend is moving much faster than it really is. For example, a movement away from the traditional family structure is soon interpreted to mean that the traditional family has almost ceased to exist.

That's why opinion makers, as well as policymakers, ought to be sure that the numbers they cite are true, not just sensational.

Spencer Rich covers social policy for The Washington Post.