Vona Weishaar of Billings, Mont., worked for 16 years as a grocery clerk, hospital technician and cannery worker, raised 3 children to be a welder, a furniture store clerk and a homemaker, and kept house for her family for almost three decades.

So, some years ago, when she had spine surgery and lung trouble and her marriage broke up, she was embittered to learn that she was not entitled to disability benefits like acquaintances who also had worked in factories and stores. Although she was "fully insured" under Social Security rules, she was 3 months short of the required 5 years of fully paid employment during the 10 years immediately before she became disabled. She had to go on welfare.

Weishaar, now 64, is typical of millions of women who spend most of their lives raising children or keeping house. Many of them never have worked outside the home long enough to qualify for Social Security benefits -- or a check big enough to live on. These women and their supporters are behind a revolutionary new idea called "earnings sharing," that has been the subject of a massive federal study.

Earnings sharing, the Social Security Administration found in the study, would give millions of women the protection of having disability insurance if needed; millions of others would get higher retirement or disability benefits.

But at the same time, the plans that help women the most would cost anywhere from $5.6 billion to $16 billion a year -- and most versions would cut benefits for millions of men.

Earnings sharing does not mean that working husbands would have to give up half their actual cash earnings to their wives. It means that the lifetime earnings records kept by Social Security that make people eligible for full basic Social Security benefits would be split.

The government, in compiling these records, would add the wages earned by a husband and wife and credit half to each. If the husband's actual earnings were $2,000 a month and the wife's were zero, each would have $1,000 on his or her earnings record.

This would ensure that most women -- including those who work in the home -- get basic disability and retirement benefits in their own name, and improve payments to many women as well.

Currently, wives and former wives who have been homemakers are entitled to a free spousal benefit, which is equal to 50 percent of the benefit due the husband or ex-husband. The average paid to divorced women retiring in 1981-82 on the free spousal benefit was $238 a month. Widows get 100 percent of their husband's benefit.

There are many problems with earnings sharing. By giving half their earnings records to their wives, millions of men (and some highly paid women) would have their own earnings records cut in half, and get lower benefits. In many families the wife's new benefit would make up the loss. But if the marriage broke up, the husband would be far worse off than now.

Moreover, without a transition period, millions of women nearing retirement would lose the free spouse's or widow's benefit without having accumulated the shared earnings record to replace it.

On orders of Congress, the SSA evaluated several earnings-sharing proposals. The three major ones were as follows:

* A "no losers" plan, splitting earnings records between husbands and wives but also guaranteeing that no one would ever receive a lower benefit than under current law. The SSA said it would cost $16 billion a year for the next 75 years. The Social Security tax, now 14.1 percent (employers and employes each pay 7.05 percent), would have to be increased by 1 percent, equally split between employers and employes.

* A second plan, called the "generic plan with transition 1," would phase out the free 50-percent spouse's benefit and the widow's benefit and alter other benefits for any losers in a relatively slow and generous manner over the next two generations. It would cost $5.6 billion a year for 75 years -- a 0.35 percent increase in the combined employer-employe payroll tax. While helping millions of women, it would result in large cuts for many men and would not target the improvements to the neediest.

* A third plan, "modified generic with transition 1," would have the same phase-ins as "generic" and cost $11.7 billion a year for 75 years -- a 0.73 percent increase in the combined employer-employe tax. It would raise benefits for a relatively large number of women and cut them for a relatively moderate number of men. In every category, the improvements would help the neediest most.

This plan's high cost results from several exceptions to "pure" earnings sharing. For example, a widow would be allowed to "inherit" her husband's earnings record and add it to her own up to certain limits. The existing "mother's benefit" for young widows with children would be retained, and widows under 62 would be given two years of partial benefits.

The study found that by the year 2030, assuming a phase-in of earnings sharing starting in 1990, this plan would result in 391,000 more women and 116,000 more men receiving disability benefits.

In addition, 154,000 more women and 79,000 more men would be receiving basic retirement benefits; 266,000 more children would be receiving various benefits; and 74,000 more people would be receiving mothers' benefits or the two-year payments.

Most women would gain. In 2030, more than four-fifths of all disabled woman beneficiaries would get higher benefits; 27 percent of retired widows would gain while only 13 percent would lose (the rest would be unaffected); and 54 percent of divorced women would get increases while 11 percent would lose. Among married couples, half would get increases and 34 percent would lose.

Single men would generally lose benefits. Widowers, with two-thirds getting increases, would gain.

Some women's groups are aware that the high budget costs of earnings sharing, coupled with the losses for millions of men, pose enormous obstacles to enactment. But they are trying to develop a version that would help women more and hurt men less than any of those studied so far -- at moderate cost.