The death toll of the Burning Times is a controversial subject -- but not half as controversial as most people think. Historians of Witchcraft generally agree that between 40,000 and 60,000 Witches died in the Burning Times. Some guess as many as 100,000, but that's as high as the experts' estimates go.
What happened in 1977-1981? Historians began systematically studying the verdicts in actual Witch trials, rather than basing their theories on propaganda written by Witch hunters.
There are two types of information we use to estimate the death toll: literary and trial data. Literary sources are the things that people wrote about the Witch trials -- pamphlet accounts of individual cases, Witch hunting manuals, letters, etc. Trial data is the records of the courts that actually did the killing. Typical trial evidence includes verdicts, records of money paid for executions, notes of goods confiscated from the condemned Witches, etc.
Trial data is far more reliable than literary data, at least as far as the death toll is concerned. The chief reason for this is, it was written by people who had access to accurate information. Courts recorded their own verdicts -- they knew whether or not they condemned a specific Witch. Much of the literature was written using second-hand tales and rumors, not by people who knew what had actually happened.
This isn't to say that trial data is completely reliable. Courts twisted Witches' words to make them conform with the theories of the _Malleus Maleficarum_ and other Witch-hunting manuals. They often removed evidence of torture and coercion, so that a Witch appears to have freely confessed to all sorts of nonsense. But the courts had little reason to lie about the number of convictions they handed down, and so a court's recorded verdicts are a fairly reliable guide to how many Witches that court killed.
Literary authors, on the other hand, had good reason to lie about the death toll. These authors were writing to either make money or to highlight the "Witchcraft threat". In both cases, a high death toll was good for the author. Bigger trials were more sensational -- they sold more copy and made the "threat" of Witchcraft seem much greater. Not surprisingly, when historians compared literary accounts of trials to the trials themselves, they found that many authors grossly inflated the number of deaths that had occurred.
We've always known that the best way to find out how many Witches died in the Burning Times was to go out and count the executions. But trial data has one great drawback: it takes a fantastic amount of work to collect it. There are literally millions of court decisions from the Burning Times. It takes almost unimaginable effort to wade through all those records, searching for the cases that involve Witchcraft. A few brave souls tried. In 1929, C.L. Ewen published the first systematic trial record count, a study of the English Witchcraft trials. But for years, few people followed in his footsteps.
Then, between 1977 and 1981, a host of trial record studies appeared, covering countries from Germany to Scotland. More followed since then. Moreover we now have computer data-bases which allow us to do statistical analysis of some countries' trials, detailed studies which earlier scholars could only dream of. For the first time in history, we can begin to base our theories on trial evidence, not Witch hunting propaganda. The impact of this new data has been astounding, revolutionizing our knowledge of historical Witchcraft.