Guide to Navigating the Visualizations

Our visualizations are created in Tableau, many of which are interactive. We compiled different graphs together that can be linked. Dashboards in Tableau allow us to use filter actions to make one view the main lens. For example, by clicking on the “Female” bar of a bar chart, the rest of the dashboard will be limited to data associated with females. In the legends, users can click on a specific category to have the option of keeping only that category or excluding it altogether.  Also, hovering over parts of the visualization will give more details about that data point.


Is there a relationship between gender and frequency of witchcraft cases in Scotland through time?

The trend in frequency of witchcraft cases in Scotland from 1563 to 1736 does not seem to differ greatly between men and women. For both men and women, there are spikes around 1629, 1649, and 1662. Where they differ is that for women, there are also large spikes in witchcraft cases around 1591, 1597, 1643, 1659, and 1678. These spikes are consistent with what is presented in our timeline. As a whole, the proportion of women accused (85.63%) is much greater than the proportion of men accused (14.37%), so the spikes in frequency for men are not as dramatic. We consider the proportion of women to men accused rather than the raw count when examining gender in relation to other variables because we want to see if there is disproportionate treatment for different genders.

What types of case attributes were used to classify witchcraft? How frequently were witches accused of these case attributes and how did they differ with gender?

There were various attributes used to accuse people of witchcraft during this period. This bar graph shows how these attributes were impacted by gender, with the largest categories being  Demon Pact and Other Charges. However, when we take the gender value for each category and divide it by the overall category total we can see which attribute has the highest percentage for men and for women. We found that the highest three category percentages for men are  Other Charges (25.40%), White Magic (21.43%), and Musical Instruments (19.61%), while the highest three categories for women are Shape Changing (97.56%), Malice (93.48%), and Demon Pact (90.14%). Based on these results, we see that men are commonly accused of crimes other than witchcraft such as Sorcery, Charming, and Murder. Furthermore, if men are being accused of witchcraft they are associated with the categories that would encompass “good” or positive forms magic or witchcraft, while women are associated with the categories with a negative connotation. In fact, we see that the highest percentage for women comes from Shape Changing, an attribute which is typically associated with non-human forms, such as animals. The fact that women are predominantly being linked with this attribute tells us that the female form is more likely to be dehumanized than the male form. By analyzing attributes by gender, we can see potential differences in how women and men are viewed in society, and while the men were also targets of the witch hunts, there is an underlying sexist bias that contributes to the accusation of women in this period. 

Is there a relationship between different attributes and residence counties? 
Definitions of Attribution Cases

Demon Pact: Accused of making a pact with the devil 

Devil Appearance: Accused of having met with non-natural or demonic being(s)

Elf/Fairy Elements: Accused of having contact with fairies or using power from fairies 

Malice: Accused of using magic to enact evil acts upon an individual 

Musical Instrument: Accused of using a musical instrument for witches meeting

Other Charges: Accused of crimes other than witchcraft (Sorcery, Charming, Murder, etc…) 

Property Damage: Accused of property damage allegedly committed by the suspect

Religious Motif: Accused of using motifs associated with prohibited religious or spiritual expression

Ritual Object: Accused of using objects in a ritualised way or for a ritualistic purpose

Shape Changing: Accused of changing their shape or form (Animal, Apparition,etc…)

Weather Modification: Accused of modifying the weather

White Magic: Accused of using white magic (Healing, love spells, etc…) 

Witches Meeting Place: Accused of being at a witches’ meeting

These tree maps provide insight into how geography factored into the attribute types people were accused of. By separating the treemap into Lowlands and Highlands, we can easily visualize which counties have the most accusation attributes, and what the most prominent attributes are. By comparing the largest county clusters in both maps, we found that for the attribute types, Demon Pact and Devil Appearance, Edinburgh and Haddington had the largest amount of people accused. These results are not unusual because as shown in our maps, Edinburgh and Haddingtion also have large amounts of witch trials. Additionally, witch hunts occured more often in Lowlands compared to the Highlands, due to the presence of the church and state (Goodare, 1998). However, theses two attribute types are interesting because almost all of Lowlands counties have these categories very high compared to the Highlands. In the Highlands, we found that Ritual Object, Property Damage, and Other Charges are the most common attribute types across the region. In comparison to the Lowlands’ attributes,  the Highlands’ attributes are less “witchcraft” focused. This difference may be because the Lowlands viewed witchcraft as more heinous, while the Highlands overall did not view witchcraft as an inherent force of evil (Hutton, 2011). 

What types of torture were used to extract information from and punish accused witches? How frequently were they used in various counties?

To navigate the visualization above, click on different counties listed in the bar graph and hover over each packed bubble to view the number of torture type records. You can also view the gender distribution of the tortured victims in the bar graph.

Compared to English law, Scottish law more readily recognized the use of torture  (Melville, 1905). Torture was routinely used in Scotland as a means of extracting confessions in order to apply to the Privy Council for a trial commission. The Privy Council of Scotland advised the monarch and also had the power to deal with the implementation of torture. In some cases, the justiciary of Scotland and Acts of Parliament authorized torture. 

Means of torture included rope binding, bow strings, feet burning, cashielaws, haircloths, thumb hanging, irons, sleep deprivation, stocks, pole binding, wedges on the shins, and whipping. Women make up 93.33% of those tortured, but they do not seem disproportionately tortured because they already make up 85.63% of those accused. 

Inverness is clearly the county with the highest number of tortured witches (60), despite it not being the county with the greatest number of accused. However, it should be noted that there are only 125 records that indicate that torture exists in our data set out of 3212 total possible rows, so this may not be representative of what actually happened. Though the scope is narrow, we still glean glimpses of what kinds of torture were used in various regions of Scotland. 

We found that a majority of the torture cases were recorded in Inverness, which is unanticipated if we examine the ratio of torture cases to accusations and note its location to be in the Highlands, where there are less trials. The lack of torture data in other counties is a case of data silence, which might indicate that torture was not always legally or officially allowed in the trial process. A possible explanation is that Inverness’s more isolated location in the Highlands allowed for torture to occur more frequently than in more central counties where central supervision was prominent (Levack, 2013).

Torture Devices
Scottish bridle

From our data set, the most common forms of torture are sleep deprivation, burning feet, being bound with rope, and being whipped. However, since our data set is limited, we note that there are many other forms of torture that have been shown to be used in Scotland (Melville, 1905). Witchcraft was considered a crime of great magnitude in Scotland, so torture methods were very harsh. Both sexes were subject to various torture methods in order to extract confessions and to punish accused witches. For instance, accused witches would go through the ‘waking,’ which involved fitting witches bridles on the accused’s head (Melville, 1905). These bridles are iron muzzles and there is a piece of iron to forcibly compress or raise the wearer’s tongue, or else their tongue would be pierced by the iron. The person was rendered humiliated and unable to speak. The accused would be watched and constantly awoken day and night until they could not take it any more and confessed. 

Is there a relationship between gender and verdict? How does verdict differ between residential counties? Is there a relationship between confession and verdict?

Explore the verdicts of each gender by clicking on the “female” or “male” bar at the bottom. By clicking on the “true” button that indicates a confession was recorded, the frequency of verdicts and gender will change accordingly. By clicking on the verdict bar, you can see the gender distribution and if a confession was recorded. 

There are four possible verdicts: guilty, half guilty, not guilty, and not proven. Guilty means that the accused was found guilty of the crime of witchcraft, regardless of whether they were found innocent of some other specific indictments. Half guilty refers to cases where there was not enough evidence to find the suspect guilty, but there was still enough evidence to presume some culpability for the crime. Not guilty means the accused was found innocent of the charge. Not proven means there was not enough evidence to proceed with the trial, but there was still some doubt about the accused. 

From the graphs, there does not seem to be a difference in the distribution of verdicts for women and men. Given that the ratio of women to men who are accused is already high, women do not seem to be found guilty at a disproportionate rate. For cases with confession recorded, 87.83% are women and 12.17% are male, so women do not seem to disproportionately confess. In our data set, there are 946 records with confession recorded, but only 235 of them also have a verdict recorded. Of the verdicts with confessions, all but 9 of the accused were found guilty, but the overall proportion of guilty cases is high regardless of confessions.  

How does verdict differ between residential counties?

This map provides the frequency of verdict type for each county. Generally, most accused individuals were deemed guilty. Edinburgh has the highest number of cases with a guilty verdict, and it is one of the Lowland counties closest to the central government. Ayr, another Lowland county, has the highest number of cases with a not guilty verdict. Aberdeen has the highest number of guilty verdicts in the Highlands. Along with Ayr, Aberdeen is one of the two counties that have half-guilty verdicts. Half-guilty and not proven verdicts were rarely given. Knowing that torture was used to extract confessions, it is not surprising that there was a high rate of guilty verdicts compared to other verdicts.

This visualization does not contain the original number of accusation cases from the dataset due to the complicated data organization. There are only about one hundred cases where we could extract both the county and verdict together due to complications with missing data values. Keep in mind, this is not a full or accurate representation, but given our data, it provides the best possible view of verdict per county.


Get 30% off your first purchase