How many women feature in UK film casts? How often are characters who work in high-skilled occupations portrayed by women? Who are our most prolific female actors?
These questions can be answered using a new Filmography by the British Film Institute (BFI). This dataset contains over 10,000 films, stretching back to 1911, and represents every UK feature-length film released in cinemas. The Filmography includes cast and crew lists for each film, and contains over 250,000 unique people. The BFI have inferred the gender of each cast and crew member by using their first name (9% are unresolved).
This article, and another about crews, provides a first glimpse at this new dataset. The aim is to spark ideas amongst film researchers and to provide a set of metrics that can be tracked over time. The Filmography will continue to grow and, with the public's help, it will improve as more contributors are identified. More details on the methodology are provided at the end of this article.
We might expect that the gender mix of film casts has steadily improved over the last half century, matching the rise of women in the wider UK workforce. But the reality is quite different. Since the end of the Second World War there have been no sustained gains in the percentage of cast members who are women in UK films. Instead, this percentage has fluctuated between 25% and 36%, and in 2017 sat at 30% (based on films released up to August 2017). Casts include actors in both main roles and unnamed roles, though extras are excluded.
The year which marked the peak for female representation in film was 100 years ago in 1917, when women made up 41% of the combined casts. And based on the films which have been released so far in 2017, there are signs that the gender mix has slipped backwards recently. These are just 'signs' because the gender of some actors cannot be inferred and more films will be released over the coming months.
The gender mix in film casts varies substantially with the genre of the film. The extremes conform to stereotypes. The casts of romance films have the highest percentage of women (39%), while war films have the lowest (20%). Other genres sit between these two extremes. And most films have been placed into the very broad genres of drama and comedy.
Although the gender balance in children's films is surprisingly low (28% women), since 2000 only 18 films have been placed into this genre and women have comprised a slightly higher percentage of cast members in these films (35%). That said, in only one of these 18 films did women make up more than 50% of the cast. Specifically, the three children's films since 2000 whose casts contain the highest percentage of women are Tooth (2004) (53% women), Molly Moon and the Incredible Book of Hypnotism (2016) (48%) and Heidi (2005) (47%).
Since 2000 just under 230 films have been placed into the action & adventure and crime genres. The overall percentage of cast members who are female has remained virtually unchanged (compared to films in these genres released before 2000). The three films in these two genres with the highest female concentration (released since 2000) have been The Bling Ring (2013) (62%), Bundy (2002) (57%) and Percy Jackson & The Olympians: The Lightning Thief (2010) (54%).
Each year only a small number of films will be nominated for BAFTA's Best Film and Outstanding British Film awards. How do these films compare, on gender balance, to other films released at the same time? In the chart below each dot represents a nominee, and the white line shows the gender balance for all UK films. Only UK nominees are shown because foreign films are not included in the Filmography.
Prior to 1970, UK films nominated for one of these two awards had casts which contained relatively fewer women than compared to all UK films released over this period. This can be seen in the chart, with most of the dots sitting below the white line. Between 1970 and 1990 there are too few nominees to consider. That's because the Outstanding British Film award was not given out between 1968 and 1991, and from 1964 there was a reduction in the number of films nominated for Best Film. Finally, from 1990 onwards, UK films nominated for these two awards have had a relatively similar gender mix to all UK films released over this period. As the chart shows, the dots are more evenly balanced either side of the white line.
The three UK nominees in these two categories whose casts have contained the highest percentage of women have been Ex Machina (2015), Georgy Girl (1966) and Iris (2002). And since 2000, there have been four nominees with casts of at least 50 that have also contained at least 50% women: Made in Dagenham (2010), Vera Drake (2005), Mamma Mia! (2008) and Bend It Like Beckham (2002).
Given the large differences in the gender mix between film casts, it is natural to look for factors that correlate with these differences. The chart below shows how the gender mix of a cast varies with the gender of its senior crew members. The focus is on films released in recent decades, from 1990 onwards.
For casting and production, films that only had women in senior roles within these departments also appear to have casts that contained slightly more women. It is not possible to draw a stronger conclusion because there are a group of actors whose gender cannot be inferred. For writing and direction, the results are clearer. Films that had at least one women in a senior role within these two departments had substantially more women in their casts than films that had no women in these senior roles. At its widest, the gender gap is 13 percentage points. That is, the overall percentage of the cast who were women in films that had all-female writing and directing teams was 13 percentage points higher than in films that had all-male writing and directing teams. A gap of 13 percentage points is large given that the percentage of cast members who were women in 2017 was just 30%.
This result could stem from a number of factors and it is important not to draw a causal link. Unidentified factors may be affecting both the gender mix of the cast and the gender of senior crew members. Furthermore, the sample size is small, as only 155 films since 1990 have had all-female writing and directing teams.
While the spotlight tends to fall on the leading actors in a film, most films will contain many more actors who appear as unnamed characters. In a film's credits these unnamed characters are typically described by their occupation, such as 'nurse' or 'driver', or by their relation to a main character, such as 'aunt' or 'neighbour'.
The chart below shows the most common unnamed characters, and how often these roles have been played by women. Gender specific character names, such as uncle and headmistress, have been removed. Named (or main) characters are also excluded.
The unnamed characters that were most likely to have been played by female, rather than male, actors are prostitute, housekeeper, nurse, secretary and receptionist. That result continues to hold if we only consider films released from 1985 onwards. Other characters which are more likely to have been portrayed by women include dancer, neighbour, singer and shop assistant. At the other extreme, the group of roles that have been played almost exclusively by male actors include several positions in the police and armed forces. This group also includes drunk, prisoner and occupations such as security guard and driver.
Many of these results are not particularly surprising. Of greater interest are those high-skilled occupations such as doctors, solicitors and managers that fall between these two extremes. In UK films these occupations have mostly been portrayed by male actors. There have been some small changes over time. For example, up to 1985 only 3% of doctors were played by women, but that has risen to 15% for films released from 1985 onwards (or 16% from 2005 onwards). For comparison women currently make up 52.2% of doctors on the UK's General Practitioners Register (July 2017). It is important to remember that many films depict historic events and so the gender mix of their characters could reflect past imbalances in the workforce. Even for films set in the present day the gender mix may reflect imbalances that still persist in certain occupations. Despite these limitations the results still shine a light on how UK films have portrayed these occupations over the last 100 years.
The BFI's Filmography can be used to study actors' careers in the UK film industry. This analysis is possible because each contributor has been assigned a unique ID in the Filmography. While unique identification allows for richer analysis, there are likely to be instances where one person's film credits have been split across two unique IDs. These errors occur because it is not always possible to prove that the same name in two different film credits refers to the same person. In these instances, rather than running the risk of erroneously merging two different actors, a new ID will be created. This tendency increases the percentage of contributors who only appear in one film. Another issue is that in the early decades of the Filmography not all cast and crew members were given credits, and it is suspected that women were less likely to be credited than men.
Even after taking into account the bias above, it is still surprising that more than half of all actors, whose first UK films were released in the 1940s and 1950s, are not credited in any other film in the dataset. It raises doubt around the idea of being able to have a 'career' in UK film. Of course many of these actors will have worked on non-UK films, or in television, theatre and other creative fields. Identifying how actors move in and out of the UK film industry over their careers may help to throw light on this curious result.
The charts below focus on actors who have been credited in at least two UK films, and the actors are grouped according to when their first UK film was released. In every decade-group, men have appeared in more films on average than women, and men have a greater number of years between their first and last films (which is a very basic proxy for career length). Interestingly however, these gaps appear to have narrowed over time. For older generations, men were completing more films than women as early as 6-10 years after their first film. But for younger generations that gap is much smaller over the 6-10 year window. Of course, the gender gap will continue to change for these younger generations as they are only part-way through their careers.
The table below shows the actors who have appeared in the greatest number of films in each decade from the 1940s onwards. It details the number of films that an actor released in that decade, as well as the total number of films they have released, up to August 2017. In each decade, the most prolific male actors have starred in more films than their female counterparts. And even in the most recent decade there is still a gender gap, of between four and eight films. The only exception is Marianne Stone, who appeared in many of the Carry On films, and was credited with more films in the 1960s and 1970s than the most prolific male actors.
The table also shows the gender balance across all the film casts which an actor has appeared in. Only two actors in this list have appeared in films whose combined casts have contained at least 40% women; they are Colin Firth and Celia Imrie. At the other extreme are actors whose combined casts have contained less than 25% women: they are Mark Strong, Susannah York and Trevor Howard. These results will reflect the genre of these actors' films, and the gender balance in any one of these actors' films may have been substantially higher or lower. One curious result is that the most prolific female actors tend to be cast in films that contain relatively more women than the most prolific male actors.
While this article focuses on gender imbalances, there is a noticeable lack of ethnic diversity in these lists. Last year, BFI research highlighted the lack of roles for black actors. They found that 59% of films in the last ten years had not contained a single black actor in any named character role.
The BFI's Filmography covers UK films that have been released to cinemas and are 40 minutes or longer. To qualify as a 'UK film', the film must have been funded or produced, at least in part, by a UK-based production company.
The BFI has inferred the gender of cast and crew members by comparing contributors' first names to a database of baby names given to boys and girls in England and Wales. The BFI also searched manually to identify the gender of prolific cast and crew members. And for remaining actors it used the gender implied by their characters' names. In comparison to a survey, this inference method is faster, non-intrusive and does not suffer from non-response bias.
There are two limitations to consider. The first is that inferred genderisation wrongly assumes that gender identity is both a fixed and binary concept. There is also likely to be the odd mistake if a contributor has a first name which is more frequently given to a different gender. As a result, a person's 'inferred gender' may not match the gender (if any) with which they identify. This limitation means that there is still very much a place for other methods, such as surveys, that can more accurately capture the complete range of gender identities.
A second limitation is that the BFI was not able to infer the gender for approximately 9% of all contributors. The first name of a cast or crew member may not be gender specific, such as Sam or Alex. Or their name may not be included in the set of baby names used to perform the genderisation. This issue particularly affects cast and crew members who have non-English-origin names. Omitting this group could bias results if the gender balance among these contributors differs significantly to others. For this reason, all the data visualisations contain upper and lower estimates which show how unidentified contributors could change the gender balance.
The visualisations show how the gender mix of a film's cast (or crew) varies with characteristics such as the film's genre or its year of release. The percentages are calculated by pooling casts for a particular genre or year, and then determining the percentage of actors who were women. An actor who contributed to two films in that year or genre will be counted twice. Thus, the percentage reflects the relative number of opportunities for women compared to men, as opposed to simply the number of women compared to men. Unfortunately, it is not possible to distinguish leading roles from minor roles.
The percentage of a cast (or crew) who are women is a simple metric and it should not be over-interpreted. If one film's cast contains a higher percentage of women than another film, it does not imply that the women in that film had more lines or more screen time. Nor does it imply that their characters were more complex or that the film itself somehow better reflected a women's perspective.