WASHINGTON (AP) — Who said the dating game was logical?

An Associated Press-WE tv poll turns up all kinds of contradictions when people lay out their thoughts on dating, especially when it comes to money and gender roles.

Seven in 10 of those surveyed say it's unacceptable to expect a date to pay for everything. But most still say it's a man's job to pay for the first date.

Most say it's OK to ask someone out because he or she seems successful. But even more say it's unacceptable to turn down people because they haven't had much success.

One-third think it's OK to search for online clues about a potential first date's success in life. But very few say daters should pay attention to each other's finances before they are exclusive.

Overall, the traits that men and women rate as important hew to traditional gender roles.

Men and women agree that personality is the most important trait to consider when deciding whether to go on a first date with someone, and very few say money is a top consideration. Yet for men, a sense of humor outweighs intelligence, and they are more apt than women to prioritize looks. Most women place greater emphasis on a suitor's financial situation and career ambitions.

It's not just older people who feel that way. The differences are amplified among younger singles. About half of single men under age 45 say looks are a priority, while 70 percent of single women under 45 call career ambitions key.

There's a clear gender gap on finances.

Men are less likely than women to say they're comfortable dating someone who makes significantly more money than they do. Seventy-one percent of women would be comfortable in that situation, compared with 59 percent of men. Women are more wary of dating someone who earns less. Forty-three percent of men would be OK dating someone with a significantly lower salary, but just 28 percent of women would.

More broadly, uncoupled Americans are squeamish about dating those whose financial situations may not equal their own.

A shaky financial past is generally acceptable, and more say they're comfortable dating someone who grew up in a poor family than in a wealthy one. But a questionable present inspires doubt.

Just 16 percent say they would be comfortable dating someone who is unemployed, and 23 percent say they would be comfortable dating someone with significant student loan debt.

Once dating turns to commitment and love, money is a bigger consideration for women when deciding whether to wed.

Among men who aren't married or living with a partner, 84 percent say they'd marry someone they love regardless of whether she or he could provide financial security. Women are more cautious, with 61 percent would choose marriage for love without regard to financial standing.

Over time, Americans' views on how women ought to balance family and career have shifted in favor of greater choice for women. But the poll also finds a more restrictive view on how men with a family ought to view their career, suggesting the rules many apply to dating continue once families are formed.

A Time/Yankelovich survey conducted in March 1978 found that about three-quarters of Americans felt women ought to put their husbands and children ahead of their careers and felt women with young children shouldn't work outside the home unless it's financially necessary. Now, about half hold those views.

But the AP-WE tv poll also found that half of Americans believe a man with a family has a responsibility to choose a higher-paying job over one that is more satisfying, compared with 42 percent who felt that way in 1978.

The poll was conducted in conjunction with WE tv ahead of the launch of the show "Mystery Millionaire."

The poll was conducted May 16-19 using KnowledgePanel, GfK's probability-based online panel designed to be representative of the U.S. population. It involved online interviews with 1,354 adults, including an oversample of 310 adults who have never been married. Results for all respondents have a margin of sampling error of plus or minus 3 percentage points.

Respondents were first selected randomly using phone or mail survey methods and were later interviewed online. People selected for KnowledgePanel who didn't otherwise have access to the Internet were provided with the ability to access the Internet at no cost to them.

___