RICHMOND -- A 5 1/2-day hearing here last week was marked by contradictory estimates by medical and statistical witnesses on the number of people who say they were injured by the Dalkon Shield and have valid claims for compensation, and on the amount of money A.H. Robins Co. must pay them.

The contradictions must be resolved by Robert R. Merhige Jr., the U.S. District judge who presides over the Chapter 11 bankruptcy proceedings of the company.

His conclusions will have substantial consequences. They will determine whether tens of thousands of women -- mostly those lacking full documentation for claims of injury by the intrauterine contraceptive device -- get token payments of as little as $100, or perhaps hundreds of thousands of dollars.

Merhige hopes to announce his estimates by Thanksgiving and to file a written ruling later, which could bring Robins' plan for financial reorganization, and its acquisition by Rorer Group Inc., closer to a vote on confirmation. The company proposed the plan to end its Chapter 11 status, which is nearly 27 months old.

The voters will be the creditors -- principally users of the intrauterine contraceptive device. Robins stopped selling the IUD in mid-1974 at the request of the Food and Drug Administration, which was concerned about an outbreak of fatal and near-fatal spontaneous abortions among its users.

It could have predicted that while each witness would swear to being impartial, rigorous in methodology, historically correct and compassionately fair, each would also provide testimony that would be in accord with the needs of the client.

Thus it was no surprise when, for example, that the statistical experts for Robins and for the outside stockholders weighed in with estimates of $1.2 billion and $1 billion, respectively, for the company's liability.

Payouts of that size would fence out claimants from the equity of the Robins family, which owns about 40 percent of the outstanding common shares, and of the outside stockholders.

Nor was anyone surprised when the expert for the Dalkon Shield Claimants' Committee -- building on the same data base as everyone else -- concluded that adequate compensation would be as much as $7.17 billion -- nearly three times the sum Rorer intends to pay to acquire Robins.

But in closing arguments Robins' Richmond attorney James C. Roberts dwelt on the competing dollar estimates, but ignored the physicians who had testified for the company.

Nor did he cite the testimony with which they upheld Robins' $1.2 billion figure. The medical underpinning was considered critical for Robins' argument for a $1.2 billion limit.

The physicians were Dr. Herbert F. Sandmire, a practicing obstetrician/gynecologist from Green Bay, Wis., and Dr. Brian L. Strom, an assistant professor at the University of Pennsylvania School of Medicine whose specialties include epidemiology. Both were enthusiastic about the shield, rating it every bit as safe and effective as other IUDs.

If the IUD were available now, Strom was asked, would he prescribe it? "Sure," he said. Sandmire's confidence in the device was such that he declared Robins' estimate of the dollar value of the injury claims to be grossly excessive.

But Sandmire and Strom posed thorny tactical problems for Roberts, as the judge signaled from the bench when he said he was conducting the hearing on the presumption of shield-induced injuries, and as implied by Roberts' silence on the medical evidence.

Apparent explanations of why Roberts had problems emerged mainly in cross-examination and closing arguments by Wendell B. Alcorn Jr. and Mark C. Ellenberg, lawyers for the Dalkon Shield Claimants' Committee, and Stanley K. Joynes III, counsel for Dalkon Shield users whose pelvic inflammatory disease (PID) and infertility have or will become manifest by the year 2000.

For example, juries in Dalkon Shield product-liability trials in the 1970s and 1980s repeatedly rejected the kind of defenses offered by Sandmire and Strom. Only last June, the Kansas Supreme court upheld a $9.2 million award to a Dalkon Shield victim, including $7.5 million in punitive damages for conduct such as fraudulent concealment of safety-related information.

It didn't appear to help Robins that Sandmire and Strom disagreed about the relationship between the Dalkon Shield and other IUDs to PID, the infections that spawned most of the serious injuries claimed by Dalkon Shield users.

PID can be life-threatening, has killed the fetuses of women who conceived with IUDs in place, and can cause blockages of the Fallopian tubes and result in sometimes irreversible infertility. Sandmire testified that the IUD could cause neither PID nor tubal pregnancies; Strom testified it could cause both.

The first of the two medical witnesses for the claimants' committee was Dr. Emanuel A. Friedman, professor of OB/GYN at Harvard Medical School, chief of OB/GYN at Beth Israel Hospital in Boston, a practicing OB/GYN, and a former outside adviser to the FDA.

Committee lawyer John J. Walsh asked Friedman to comment on each of more than 20 "propositions" about associations between IUDs, particularly the Dalkon Shield, and PID, infertility and other diseases and injuries.

Friedman's labels for the majority of the propositions were "false," "untrue," "nonsense," although some were merely "misleading" or "wrong."

Afterward, it was learned that the propositions were Walsh's paraphrases of the proposed "findings of fact" Robins has submitted to the judge.

Another issue was a pivotal 1983 study of women who had acute PID and were hospitalized. Dr. Nancy Lee of the Centers for Disease Control in Atlanta reported that the PID rates in Dalkon Shield users were about eight times higher than in users of other IUDs. As a result, the FDA urged women still wearing the IUDs to have them removed.

Last month, a scientific journal printed an abstract of an unpublished article by Lee. The abstract said refinement of her study suggested that the 1983 assessment of PID risks in IUD users "may be overestimated due to confounding factors related to the risk of acquiring sexually transmitted infections."

The abstract also said that among the study participants, "women with only one sexual partner had little appreciable increase in {PID} associated with intrauterine device use, compared with women using no contraception, whereas previously and never married women had {PID} risks of 1.8 and 2.6, respectively."

Sandmire said recent studies he trusts showed that the 1983 data implicating the Dalkon Shield "could not possibly be correct."

On this and other issues, he was supported by Dr. Elton Kessel, the medical witness for the bankruptcy committee representing the outside stockholders.

Lee herself did not testify. But the Sandmire/Kessel version was challenged by Dr. Janet R. Daling, a professor of epidemiology at the University of Washington who did a major study of IUDs and infertility funded by the National Institutes of Health.

Testifying for the claimants' committee, Daling said that in the 1983 study, the married women who had only one sexual partner and no increased risk of PID were a small "subgroup" and consequently could not justify either major conclusions or exoneration of the Dalkon Shield.

A person who spoke to Daling since her return to Seattle said she had seen the full article and Lee had not changed her 1983 conclusions.

Neither Friedman nor Daling budged in cross-examination, which was brief and perfunctory.

Also a problem for Robins was Sandmire's inconsistency about some issues, such as his estimates of the incidence of so-called "silent" or subclinical PID, in which the symptoms may not be evident to the victim or to examining physicians for long periods.

In keeping with his contention that sexually transmitted diseases account for virtually all PID in IUD users, Sandmire testified initially that infections called chlamydia caused "98 percent" of the PID in device wearers.

Within minutes he had lowered the figure to a range of "80 percent to 90 percent," and finally -- in response to a question from the judge -- to 60 percent to 80 percent.

It also turned out that he had lost the office notebooks in which he said he had recorded events associated with Dalkon Shield insertions over a three-year period.