The National Aeronautics and Space Administration's process of deciding whether it is safe to launch a space shuttle "is flawed" and must be improved, William P. Rogers, chairman of the presidential commission investigating the Challenger explosion, said yesterday.

Rogers, ending three days of public hearings that brought to light a web of misunderstandings and failures to relay warnings, issued what in effect is the panel's first stated conclusion.

"You'll remember," Rogers told the top space agency officials who ultimately decided to launch Challenger a month ago, "that I did say at one point that we thought the decision-making process may be flawed. I believe I'm speaking for the whole commission when I say that we think it is flawed.

"And I think probably you . . . will agree with us on that -- that the process as it worked in this case was clearly flawed because recommendations were made that were not fully understood by you or not conveyed to you. We assume that you are working on improvement of the decision-making process."

The NASA officials present did not reply. Acting administrator William R. Graham, reached later, indicated that no improvements are under way because "at the moment we are focusing all our efforts on supporting the commission and at the appropriate time we will take their recommendations under careful consideration."

Although the commission's mandate from President Reagan was to find the cause of the explosion that killed the crew of seven and recommend a remedy to protect future shuttle flights, Rogers very quickly enlarged the scope of the investigation when it became apparent that several warnings not to launch had been ignored or overridden by mid-level decision-makers.

The most relevant warning appears to have come from engineers at Morton Thiokol Inc., manufacturer of solid rocket boosters used to push shuttles into orbit.

The engineers told NASA officials of their unanimous concern that unusually cold temperatures at the Cape Canaveral launch area threatened rubber O-rings designed to seal booster joints against burn-through.

Just such a burn-through is thought to have occurred.

The night before the Jan. 28 launch, when Thiokol engineers recommended that it be postponed until a warmer day, mid-level NASA officials responded by arranging an unusual late-night telephone conference call with the engineers in Utah. The engineers later told the commission that they felt pressured in that call to approve the launch.

Eventually, Thiokol management overrode the engineers' recommendation that night and gave NASA the go-ahead. The ranking NASA official involved in the conference call told the commission that he decided the problem was resolved and that it was not necessary to tell his superiors, the officials who would make the final launch decision, about the engineers' fears.

Under intense panel questioning, led by Rogers, NASA officials said that, if Thiokol management had not reversed the recommendation, Challenger would not have been launched.

Every higher-ranking NASA official who testified said that, had he been told of the engineers' opposition, he would have stopped the launch.

In other words, Rogers told the top NASA officials before him, the launch decision was made not by NASA or Thiokol engineers but "by just a couple of people [Thiokol management], and apparently none of you gentlemen knew about it."

Stanley Reinartz, the mid-level NASA official who testified Wednesday that space agency procedures did not require him to inform his superiors, "lived up to the book," Rogers said, "but there was no application, as far as I can tell, of common sense."

Another example of blocked warnings emerged yesterday when officials of Rockwell International Inc., the shuttle's prime contractor, said they recommended against launch because too much ice was on the shuttle and the tower next to it.

Arnold D. Aldrich, the second highest decision-maker on the Challenger launch, told the commission that he heard Rockwell's comments but did not interpret them as a no-go recommendation.

Jesse Moore, the top NASA official, said Aldrich told him of Rockwell's comments but did not convey them as a definite recommendation not to launch.

Several top NASA officials said they were surprised that none of the warnings reached them. Moore said he assumed that all concerns about flight safety would have been relayed to higher levels.

"Issues that reflect flight safety should be properly brought up to the appropriate level," Moore said. Referring to the engineers' objections, he said, "I would have thought it would have been brought to Level 2, if you want my honest opinion." The issue was disposed of at what NASA calls Level 3.

Launch director Gene Thomas, referring to all of the unheard warnings, concurred, saying:

"I can assure you that, if we had had that information, we would not have launched."