Trial of George Jacobs
Trial of George Jacobs


Exploring Salem Witchcraft by means of the data sets provided on this website reveals the advantages and limitations of using quantitative analysis to understand the past. On the one hand, working with quantitative data stimulates new ways of thinking about events. It encourages a search for patterns in events, and it can yield greater precision in detailing economic relationships that otherwise might escape detection or be misinterpreted. At the same time, only certain dimensions of historical experience are open to quantitative analysis, and patterns found by such analysis must often be explained by other means. The data sets will not directly help us understand what went on in Samuel Parris's household in the early months of 1692. Nor will they explain the motives and thoughts of judges, juries, and accusers during the outbreak. But they are a useful tool for uncovering aspects of the Salem event that might remain hidden by a sole reliance on traditional historical sources and methods. Quantitative analysis is a complementary methodology that can stimulate and test ideas, and it is best used in conjunction with, not as a substitute for, contemporary documents and historical accounts.

What aspects of the Salem outbreak does this website highlight?

Chronology and Geography

Working with the data sets focuses attention on Salem Witchcraft's chronology and geographic scope. Analysis of accusations reveals, for example, that the initial three accusations in February 1692 did not cause an immediate and vast witch hunt. Instead, much like traditional limited-scale witchcraft events, the Salem episode was first contained to a relatively small number of cases in Salem Village and Town. However, in April and May a genuine witch hunt got underway: the number of accusations dramatically increased and spread to numerous communities beyond Salem. Contemporary sources and historical research have suggested a variety of reasons as to why the episode eventually burst into a full-fledged outbreak: among them, the continued fits of the afflicted; the credence given by influential adults, including religious and secular leaders, to the notion that the Devil was, in fact, attacking their colony; the accusation and arrest of a former Salem Village minister, George Burroughs, which was interpreted as a sign of the Devil's power to subvert the church; and the confession of some of the accused, who then accused others. But exactly how, why, and in what manner the initial afflictions within Samuel Parris's household turned into what we know as Salem Witchcraft is a fascinating subject open to further investigation.

Accusation data also reveal a significant wave pattern to the Salem event. The wave of arrest warrants in the spring suddenly reversed itself in June. Accusations dramatically declined after the first trial, conviction, and hanging of a witch (Bridget Bishop) by the newly-established Court of Oyer and Terminer. However, with the resumption of legal proceedings at the end of June, a second wave of accusations was underway. It peaked in September 1692, and coincided with a rising number of executions resulting from trials.

This second wave differed significantly from the first. Its epicenter was not Salem Town or Village but Andover, and almost all of the communities involved in second wave differed from those involved prior to mid-June. The Salem witch hunt was, therefore, extensive in its geographic reach, involving some two dozen Massachusetts communities. At the same time, the evidence indicates that in most communities the outbreak struck briefly and was directed at only a small number of residents. Even in those places with many victims, such as Salem Village and Town, Andover, and Gloucester, the witch hunt was contained within a few months. It seems that the longevity of Salem Witchcraft was, ironically, the product of its rapid movement from town to town rather than its persistence within communities.

By the end of September, however, the witch hunt again subsided. In October, only one arrest occurred and the Court of Oyer and Terminer was disbanded by the Governor Phips. Three final accusations were recorded in November. The outbreak was over. There would be no third wave of accusations.

Historical sources provide some guidance regarding the witch hunt's second wave. The pause in activity in June was likely owing to the temporary suspension of trials following the Bishop case. Concerns about the behavior of the afflicted accusers as well as the Court's reliance of spectral evidence surfaced, and not just by families of the accused. One of the Court's judges, for example, resigned after Bishop's trial. The Court halted its deliberations while Governor Phips and civil leaders sought advice from the colony's leading ministers about how best to proceed. The minister's "Return," though strongly cautionary about reliance on spectral evidence, was clear as to the need to vigorously prosecute witches. When the Court resumed deliberations at the end of June, the legal machinery again fully supported a resurgence of afflictions and witchcraft accusations.

The second wave focused on Andover and other new communities, little touching Salem or communities involved in the earlier wave. Comparatively little has been written about the geographical spread of accusations in 1692, but the different pattern of accusations after June suggests that in many communities there were only a limited number of likely witchcraft suspects. Once they were arrested, the witch hunt moved elsewhere. In any event, the attempt to detect and punish witches during the summer of 1692 was as vigorous as in the spring. Afflicted accusers grew in number; those living in or near the newly involved communities like Andover joined the ranks of the earlier afflicted from Salem Village. Most dramatically, the number of confessors soared, driving the witch hunt forward as it appeared more certain than ever that the Devil was engaged in a campaign to destroy the Puritan commonwealth.

The precipitous drop in arrests in October, which signaled the end of the Salem episode, resulted from a complex mix of considerations. The witch hunt assumed aspects of mass hysteria, but it never lacked skeptics and dissenters. The dissatisfaction gained traction as the number of both afflicted and accused grew unrealistically large, particularly when prominent and respected members of communities were charged or named. Many ministers, like the prominent Increase Mather, grew increasingly uneasy about the continued prominence of spectral evidence in examinations and trials; might not the Devil use images of innocent people to deceive the afflicted? Clearly, something was wrong when a legal system, which had been instituted to contain and eliminate the threat, seemed instead to be endlessly validating new revelations. Opinions differed as to what had gone wrong. Some, for example, thought the afflicted girls had simply engaged in a campaign of lying; others thought the Devil had been too clever in spinning a web of deception to undermine the social order. By October, widespread disapproval with the Court was being expressed by civil and religious leaders and lay people. Governor Phips both recognized and shared this reaction, and beginning in October, he dismissed the Court and instituted other measures to quiet the province. Although his own Lieutenant Governor William Stoughton, who served as chief judge of the Court, complained, few others wished the witch hunt to continue. A period of time was required for the colony's new superior court to hear cases of those still imprisoned. Virtually all these trials resulted in decisions setting the defendant free. For the few found guilty, as well as for those condemned but not executed by the previous court, the governor provided reprieves. In May 1693, Phips issued a general pardon, which released those remaining in jail who paid their jail fees.

The data also demonstrate that the 1692 witchcraft outbreak was geographically extensive. Although Salem Witchcraft is popularly associated primarily, if not exclusively, with Salem Town and Village, the outbreak claimed victims from two dozen Massachusetts communities (including Maine). From its initial locus in Salem Village, it expanded in all directions. Because the data sets are based upon legal proceedings, they likely undercount the communities that were affected. But these communities were not all involved at the same time. In the period from February through early June 1692, activity centered on Salem Village and Town. But in the period from the end of June through September, activity shifted northward, especially to Andover and Gloucester. At least as regards towns with at least five accused witches, there was very little overlap between the two periods; communities involved in one period were not involved in the other. Since the witch hunt had all but ended in Salem Town and Village by June, its geographic spread was essential to its duration, although the way in which it moved from place to place and singled out its victims remains obscure.

Salem Village Discord

Analyzing the chronological and geographic dimensions of Salem Witchcraft provides many insights into the outbreak. But much scholarly attention has been directed at Salem Village itself. It was here that the initial afflictions began at the home of the Village's minister, Rev. Samuel Parris. Village residents were the initial targets of witchcraft charges, the village was the site of the earliest examinations, and the village's afflicted girls remained active throughout the outbreak, accusing new victims and providing court testimony against defendants.

The question of "Why Salem?" has received a variety of answers. Among other ideas, studies of Salem have emphasized property and other local quarrels; the lying antics of the afflicted girls; conflict between rival settlement groups; the religious enthusiasm of Samuel Parris; the existence of residents actually practicing witchcraft and magic; generational tensions; and the effects of hallucinatory-inducing ergotism or of encephalitis letharica. While some arguments are unconvincing to scholars, particularly the medical conditions of ergotism poisoning or encephalitis, others offer a useful, though partial, understanding of an event that defies a single explanation.

Among the numerous explanations of Salem, one has a special prominence, and the data sets in Section Three relate to it. In a now-classic study, Salem Possessed, Paul Boyer and Stephen Nissenbaum examine how social and economic tensions within Salem Village resulted in the witch hunt of 1692. Boyer and Nissenbaum uncovered a deep rift in Salem Village between those who supported the ministry of Samuel Parris as well as the witch hunt and those who objected to or withheld support from Parris and the pursuit of witches. Behind the conflict between the village's pro- and anti-Parris factions was a momentous clash over economic modernization. The anti-Parris group was associated with the forces of market capitalism and were its beneficiaries in terms of their greater wealth. The pro-Parris group was associated with traditional agrarian values rooted in support for the village church; they were considerably poorer than their rivals. Much, though by no means all, of the evidence for the economic condition of each faction came from an analysis of the village's 1695 tax assessment, taken about three years after the witchcraft episode.

The website's tax data sets permit a more thorough investigation of the relative position of Salem Village's two factions by extending the economic analysis over a period from 1681-1700 and by examining both group and individual mobility. They allow users to think analytically about categories of economic standing (e.g., the "rich," "poor," "middling") and about ways to measure changes in economic position. The analysis does not always yield clear-cut results, and users may interpret the findings somewhat differently. But it appears that while the village's tax data provides some support for Salem Possessed, they generally casts doubt on the book's thesis. For example, investigation confirms relatively large wealth disparities between the pro- and anti-Parris groups in 1695. But much other data undermines that evidence. The tax list for the year most proximate to the witchcraft outbreak, 1690, reveals little difference between the two factions. Furthermore, when placed in a longer perspective, the 1695 assessment proves to be an aberration. Over the period of time between 1681 and 1700, the pro-Parris group was not falling behind the anti-Parris group. If anything, the pro-Parris group was advancing relative to its opponent and the factions were drawing closer together. This was especially true for the period leading up to the outbreak: in the decade from 1681-1690, disparities between the two groups virtually disappeared. The median assessments for the pro- and anti-Parris factions in 1690 were identical. The downturn in the fortunes of the pro-Parris contingent evident in 1695 were largely corrected by 1700.

If the claim that changes in the economic fortunes of Salem Villagers provoked the witch hunt seems unpersuasive, there may be better evidence that Salem Village's factional strains were rooted in religious tensions. Analysis of both the pro- and anti-Parris petitions as well as the village's committee membership discloses a sharp division in Salem Village between church members, strongly associated with their minister, Samuel Parris, and non-church members, who dominated the anti-Parris committees of 1691-1693 and signed the petition of 1695 requesting that Parris be removed from his position. Church members composed a much greater proportion of pro-Parris petition signers as well as of pro-Parris committee members.

Non-membership in the Salem Village church does not necessarily signify a secular mindset or opposition to religion. Many of Parris's supporters were not church members, and many anti-Parris non-members likely attended services elsewhere. Church members and non-members alike might object to a particular minister without disapproving religion or desiring the elimination of the village church. Such seems to be the case in Salem Village. When the village eventually replaced Parris, his successor, Joseph Green, brought an end to the divisions that had wracked the community for years. Although Salem Village's turmoil over its church and minister might not account for events elsewhere, Salem Village's longstanding religious tensions seem likely to have significantly contributed to the initial afflictions in Parris's own household and to the conclusion that they were caused by agents of the Devil. Once evidence came to light that the Devil's agency was not limited to Parris's household or to Salem but was designed to destroy the entire Puritan community, some religious and secular leaders determined to root out the Devil's minions, and the witch hunt became widespread.

To inquire further into Salem Witchcraft, consult the historical sources in the Bibliography. Click Next.