The Central Intelligence Agency was created to provide comprehensive, all-source collection and analysis of information so that we might prevent strategic surprises like Pearl Harbor and be forewarned of other developments adverse to American interests. Granted, the effort is immense and complex. Recent press accounts prompt some to ask: Is it any good? Is it honest and objective?

There is little question that maintaining the quality of CIA assessments became much more difficult in the late 1960s and the 1970s. Collection capabilities declined. Our analytical effort on the Third World had been significantly reduced by the early 1970s -- just when problems there were multiplying.

By 1980 the number f analysts working on the Soviet economy (including defense industries) had declined from over 300 to fewer than 50. There was little money for analysts to travel abroad or to meet with nongovernment experts at home. Many academics were unwilling to talk to us and share ideas. From 1973 to 1977, moreover, CIA had five different directors, and from 1975 to January 1982, there were six chiefs of the analysis directorate -- the Directorate of Intelligence.

Much has changed in the past five years. The resource picture began to improve in 1979, thanks initially to the House and Senate oversight committees. We have since made impressive strides toward rebuilding the corps of analysts. New resources for the entire intelligence community have greatly improved the collection of information across the board. We have also undertaken sweeping measures to improve the quality of analysis.

The directorate of intelligence was reorganized in 1981 to bring political, economic and military experts together in regional offices. We have dramatically expanded our contacts outside government, drawing on an extraordinary number of experts in universities, think tanks and business for information and ideas. We require all CIA analysts to have outside training every two years.

CIA has strengthened longer-term analytical research, long put at risk by the pressures of day-to-day reporting. In the first nine months of this year we issued some 700 research studies for nearly every department of government. For the first time there are adequate funds for analysts to travel and work overseas as well as to consult with again-cooperative academic and other experts at home.

CIA assessments now are subjected to more rigorous internal review than ever before. Every manager at every level reviews all substantive assessments that come out of his organization. We often offer drafts for comment (though not consent) to senior military commanders, embassies and experts in other agencies. Many of our assessments are reviewed by nongovernment experts.

We not only offer our best estimate of what will happen in a given situation but also inform our readers of other possible though less likely outcomes -- and the implications of each. I cannot say this approach would have enabled us to predict the fall of the shah in 1978-79, but I believe that that outcome now would certainly be addressed as a possibility. We are more candid now with our readers about the level of our confidence in our judgments and the reliability of our sources. We also make more of an effort to lay out our evidence. Using the example of the fall of the shah, under present practice we would have acknowledged the paucity of information on internal Iranian affairs and the self-serving nature of some of our sources.

We now evaluate past CIA assessments and national estimates to see how they have held up over time. The directorate of intelligence has for the first time its own independent evaluation staff. We voluntarily share these evaluations with the House and Senate oversight committees.

We organize special task forces of agency experts and outside specialists to do competitive analysis and to ensure we are examining all aspects of key problems. We submit our work on important issues, such as the Soviet economy, to panels of outside experts for scrutiny. Finally, the skill and dedication of analysts in CIA and elsewhere in the intelligence community are exceptional -- perhaps never better.

While some of the criticism in the press of our capabilities and acumen is justified, most of it is grossly inaccurate. I urge the reader to consider the access and motives of sources of criticism -- and to be alert to later retractions. Meanwhile, I have hundreds of letters, cables and messages, from the president on down, commending our work. Various news organizations report that policy-makers and members of Congress acknowledge that the quality of assessments has improved markedly.

CIA was created in part to ensure that intelligence assessments would be prepared by people with no stake in approval of weapons programs, defense budgets or particular policies. Perhaps the strongest cultural trait common to all CIA analysts is a very deep sensitivity to the dangers of politicization. Indeed, sometimes we bend perhaps too far toward an adversarial relationship with policy-makers to avoid even the appearance of being suborned.

There is no question that policy-makers have always been intensely interested in the outcome of our assessments, especially on contentious issues. If they were not, it would mean we were working on the wrong problems or were irrelevant. Beyond our natural, visceral independence -- contrariness, some would say -- a number of safeguards exist:

Approval of CIA's assessments rests with intelligence professionals. I have been with CIA nearly 20 years. I am the final approving official for all of CIA's daily production of current intelligence that goes to the president and senior government officials. The director of Central Intelligence (DCI) first sees it at the same time as the policy-maker. I also approve all longer-range assessments.

Our assessments go to the two congressional oversight committees. I am confident they would not hesitate to act promptly if they detected a policy slant. In addition, both Foreign Relations, Armed Services and Appropriations committees receive a great number of our assessments.

A variety of other groups, both independent of CIA (the president's Foreign Intelligence Advisory Board), and inside the Agency, also evaluate CIA assessments and estimates.

Directors of CIA have always played an active role in the preparation and approval of national estimates, which are produced by the entire intelligence community. Similarly, our directors always have had strong views on the major substantive issues we analyze. John McCone, President Kennedy's DCI, believed the Soviets would send missiles to Cuba in 1962 long before the intelligence analysts agreed. However, national estimates also are reviewed by the heads of a dozen other intelligence organizations. The estimate that recently was alleged, in the press, to be slanted went through many drafts and even then nearly half the community's intelligence agencies dissented -- and the dissent was spelled out on the first page.

Finally, perhaps the best guardians of the integrity of the process are the caliber and honesty of the people involved. We are not cowards. We present assessments unwelcome to policy-makers day in, day out on a broad range of issues, and we have for a long time. I believe most policy-makers would attest that, especially on controversial issues, intelligence assessments are more likely to be troublesome than supportive.

Our assessments are not produced in an ivory tower atmosphere. The debates and clash of ideas sometimes are rough. No one's views -- from the director to the newest analyst -- are protected from challenge. It is not a place for delicate egos or mediocrity or people with special agendas.

But, however hot the debate or pointed the questions during the drafting, the final product is as honest and accurate as humanly possible. Despite imperfections, CIA and the intelligence community produce the best, most comprehensive and most objective intelligence reporting in the world. We are working every day to make it better, and however surprising it may be to our critics, we believe they contribute to this process, and so we listen to them.