Why? "County level data was not complete. We do not have data based on licensed, actual records in some states. In some states data was sampled," says Stephanie Ventura of NCHS. Accurate marriage data was coming only from 42 states; and complete divorce data from only 31 states.

This is shocking. Therefore the Census Bureau added questions to its American Community Survey (ACS), a huge sample of 2.2 million households. Beginning in 2007, people were asked if they married, divorced or were widowed in the past year.

However, the results strike me as very odd.

For the first time since 1990, we know how many divorces there were in California. Census reports in 2008, there were an estimated 119,300 men who divorced in California and 145,300 women who divorced. That compares with 225,900 men who married in that year and 259,200 women.

That's a state divorce rate of 48.8%, which is surprisingly below the US average of 54.6% according to Census.

However, how could there be more men getting married than women?

How could there be more women getting divorced than men?

Is this something peculiar to California? No.

Census actually reports that in 2009, 2,286,000 men married 2,208,000 women. So there were 78,000 more men getting married than women.

How absurd. This data refers to only to heterosexual marriages. Did 78,000 women forget that they got married?

Not likely. A marriage is the most important event in the life of a woman, or a man. The only experience of equal importance is a divorce.

Yet Census is claiming that 1,178,000 men divorced in 2009 and 1,309,000 women.

Again, how could 131,000 more women get divorced than men?

The ACS report states: "Variation in rates between men and women can be attributed to gender differences in marriage. Women tend to live longer than men... Men also remarry more than women do, so men's marriage rates were higher than women's."

These explanations are silly. So what if women live longer than men? The issue is how many married in one year. While men marry more than women, that is over a lifetime. In a single year, not one man in 10,000 would marry twice.

In an interview, Diana Elliott, a co-author, explained, "Men tend to underreport their divorces. They are also harder to find." That is a somewhat plausible explanation. Some divorced men disappear on to the streets, in liquor-fed depression. But that does not happen to a tenth of men who divorce.

Her explanations why, in 2007, there were 100,000 more men who married than women: "Men have higher remarriage rates than women." Yes, but not in a single year.

What's most puzzling to me is that Gallup can predict with remarkable accuracy who will win the Presidency by interviewing only 1,000 people. The ACS sample was 2.2 million households, 2,200 times the Gallup sample. It should report almost exactly the same number of men as women who marry or divorce.

This is elementary. Men marry women in equal numbers in one year. Equal numbers of both genders divorce in a single year.

"It is definitely weird," said one national marriage expert who didn't want to be quoted by name.

More than that, it is tragic. In part due to sloppy local and state record-keeping, and due to almost laughable Census data, we will never know how many are marrying or divorcing with confidence.

Census appears to be embarrassed by its own data.

"Marital Events of Americans: 2009" does not even contain state-by-state numbers of marriages and divorces. Instead, it gives marriage or divorce rates such as 19.1 marriages per 1,000 men in California and 17.5 marriages per 1,000 women.

Earlier in this column, I gave actual numbers by Census for California, but they were for 2008 in an earlier report. Apparently, the agency was criticized for data lacking credibility.

Census can't count. To cover that up, we are no longer given actual numbers in a 24-page report.

It is a sad, shabby story.

----Mike McManus is President of Marriage Savers and a syndicated columnist.