Thirty years ago, an Environmental Protection Agency committee determined that the popular weed killer Roundup might cause cancer. Six years later, in 1991, the agency reversed itself after re-evaluating the mouse study that had been the basis for the original conclusion.

Now the issue is back again, in an even bigger way. An agency of the World Health Organization has declared that glyphosate, the active ingredient in Roundup, “probably” causes cancer in people. One piece of evidence the agency cites is that same mouse study.

The declaration drew an angry response from Monsanto, the maker of Roundup, which has accused the agency of having an “agenda” and “cherry picking” the data to support its case.

The conclusion is “starkly at odds with every credible scientific body that has examined glyphosate safety,” Philip Miller, Monsanto’s vice president for global regulatory affairs, told reporters on Tuesday. That includes a recent review by German government regulators on behalf of the European Union.

The new controversy and the reversal by the E.P.A. decades ago demonstrate how the same data can be interpreted differently and how complicated and politically perilous such a decision can be. But the discrepancy between Monsanto and the health organization can be partly explained by the specific way its agency analyzes the data.

Officials at the agency, the International Agency for Research on Cancer, said they had no agenda other than to inform the World Health Organization. They said the conclusion was based on studies of people, laboratory animals and cells.