Panel Parity – what we achieved, and how we did it

(This was originally posted on 6th May, and an amended and extended version was posted on 8th May).

Just over a year ago I wrote a blog post explaining why Eight Squared Con was committing to gender parity on panels and what we meant by that. We said that we would seek to avoid panels where one gender was in a minority, and would do so by aiming to have a sufficiently large pool of programme volunteers that itself represented the broad gender parity seen at Eastercons in recent years.

So, how did we do?

This graphic depicts all our multi-person panel items. In other words, it doesn’t include single-speaker talks, two-person interviews, or ‘event’ items such as opening and closing ceremonies. That leaves us 73 panels that we consider we were seeking to achieve gender parity on.


In numerical terms:

  • Of the 3 6-person panels, one was balanced; the others were predominantly female.
  • Of the 46 5-person panels, 32 were balanced. 6 had one man and 7 had one woman; one was all female.
  • Of the 21 4-person panels, 13 were balanced. 4 had one man and 4 had one woman.
  • All 3 3-person panels were balanced.

Now, one of the first things to observe is that although we achieved gender parity, we didn’t quite do it the way we originally envisaged, i.e. by having as equal as possible a gender split on each item. (By ‘equal as possible’ I mean that changing the gender of one participant would not affect the overall balance. So a 5-person panel is balanced if it has a 2:3 or 3:2 split; a 4-person panel only if it has a 2:2 split.)

In practical terms, it became clear that this would be extremely difficult. Even with a very large pool of participants, as discussed below, there are some topics where you have more prospective panelists of one gender than of the other. Equally, there are practical constraints such as avoiding programme clashes or times when a particular person isn’t available that have to be taken into account. Some sorts of items are harder to assemble a panel for, either because there are few people attending the convention with those particular interests, or because certain sorts of item (in particular, those with an element of performance, such as panel games) attract fewer volunteers.

However, there is another way of looking at gender parity that is, I suggest, as valid or perhaps even more valid than parity on each item. That is striving instead for a programme that is no less gender-equal than one where the panelists had been selected from an equal pool of men and women without any consideration to gender at all.

One can calculate the statistics for this (hands up if you remember the binomial theorem). If we’d allocated in this manner, we would have expected 29 of our 46 5-person panels to be balanced, and 2 or 3 of them to be all-male or all-female. Instead, as noted above, we did rather better than this, with 32 balanced panels and only one single-gender one. For our 21 4-person panels, we’d have expected a random allocation from an equal pool of men and women to result in only 8 balanced panels, and 2 or 3 to be all-male or all-female. In fact we had 13 balanced panels and no single-gender ones.

Moreover, if you count panel slots (i.e. seats on panels) we had a total of 341; this is a bigger number than our number of programme participants (190) because many people were on more than one item. Of those 341 slots, 169 were filled by women and 172 by men.

So, in our view we succeeded in our overall aim. The programme was balanced in terms of the number of male and female participants, and the majority of individual panels (49 of 73) had parity. We could not always avoid panels with a single man or woman, but we only had one single-gender panel. Above all, we had more gender-equal panels than would have been expected by a truly gender-blind allocation from equal numbers of men and women.

How did we achieve this? In large part by doing what we intended and getting a wider pool of participants than has tended to be the case. With 190 people on programme, we had about a quarter of the adult attending membership taking part. Given that the overall attendance at Eight Squared Con was about 40% female, our programme participant pool was also fairly evenly split. We made every effort to reach out to potential programme participants, and used a detailed volunteer form to ensure that we had as much information as we could about our volunteer’s interests and the sort of items they wanted to be on.

Now, this discussion has all been in the context of panel parity. I said earlier that I’d excluded programme items with only one or two participants. But there is one area outside panels that we should acknowledge we could have done better on, and that was on individual talks. Setting aside those talks arranged by other bodies, and Guest of Honour items, we had four items presented by a single speaker. All four of those speakers were male, and if I was looking to move beyond just panel parity at conventions I’d seek to ensure that we made sure that men and women were more equally represented on talks too.

Overall though we had a lot of feedback at and after the convention suggesting that we’d managed to run a varied, full and diverse programme that attained the high quality standard we’d set ourselves. Indeed, by making us look more carefully at the composition of each programme item, and by leading us to go out and find as wide a possible pool of participants, we believe that striving for panel parity actually improved the quality of the programme as a whole.


Why we are committing to Gender Parity on Panels

A question asked at the bid session for the 2013 Eastercon was whether Eight Squared Con would apply a policy of gender parity for panels. On behalf of the bid committee, having discussed this issue in advance, I stated that we would. This blog post is aimed at explaining a bit about what we mean by this.

In short, gender parity means that for those programme items that are discussion panels (i.e. four to six panel members discussing a topic, usually facilitated by a moderator) we will, so far as is practical, avoid having panels where there is a gender imbalance.

Firstly, I will say that we have to be careful what we mean by gender. ‘Sex’ and ‘gender’ are words with particular meanings that are not always understood the same way, especially in their technical as distinct from cultural senses. For the purpose of ‘gender parity’, we mean the male/female identity that a person identifies as in cultural and social terms. (Yes, we are aware that this is not always straightforward, but we are aiming to set a general policy in this respect.)

Secondly, by ‘avoiding a gender imbalance’ we mean that we would generally avoid not just all-male (or all-female) panels but also panels where one gender was in a small minority. Having a single woman on a panel of four, or two on a panel of six, would be imbalanced in this respect.

We are setting out our policy on this because in the run-up to the 2012 Eastercon there was significant discussion in many fan forums about the lack of gender parity on panels and the steps that might be taken to deal with this. We have thus given this issue considerable thought and have decided that we will commit to gender parity. Our reasons for doing so may be summed up as follows:

1) It is one of the defining features of fan-run conventions such as Eastercon that the bulk of panel participants are recruited from the convention membership.

2) Eastercon membership now shows broad gender parity; fandom is no longer male-dominated (at least not in terms of numbers, and arguably not in terms of active participation.)

3) The Eastercon membership, taken collectively, constitutes an exceptionally interesting, experienced and well-qualified group of people. In my own expericne I recall one prominent visiting science lecturer saying that of all the audiences he’d given talks to, including major science festivals, the Eastercon membership was the most engaged and erudite he’d ever seen.

4) Points (2) and (3) work together: the Eastercon membership is comprised of interesting, experienced and well-qualified people, evenly split across gender. Our pool of potential panel members should thus reflect this and should itself be evenly split across gender.

5) If we draw our panels from this pool of panel members, then panels themselves, on average, ought to come out evenly split on gender. Some might have more men than women, some more women than men, but these would be the exception rather than the rule and the programme as a whole should contain broadly equal numbers of men and women.

Experience shows that this has not generally been the case. Historically the bias was because (2) was not true, but that’s not really valid any more. It is more likely that what happens is that (3) is not fully exploited and that in practice the pool of programme participants is drawn from a relatively small subset of convention membership. That pool is small in part due to inertia (“Andy, Bob and Charles always do well on panels”) and partly because women may feel disinclined to volunteer for programme if they see that panels are male-dominated.

Is it a bad thing that panels are male-dominated? We suggest that it is, for several reasons. If this argument is correct, then we are missing out on many really good panel members. Male-dominated panels imply that fandom is a male-dominated community, which it shouldn’t be and indeed arguably isn’t. And if we restrict the pool of panel participants, even without intending to, then it makes it harder to put together a programme. As someone who has organised programme for many conventions, one of the biggest headaches is trying to avoid using the same small pool of people to populate panels with.

So how do we avoid this problem? Firstly, we are going to do our best to be very inclusive in getting people to volunteer for programme. Not everyone makes a good panel participant, but very many of the people attending an Eastercon would, and members should be encouraged to feel comfortable in putting themselves forward. (Something I will stress is that this in now way implies that women should be or will be pressured into being on programme; rather, they should feel that there is no pressure against them participating.) We will also encourage members to suggest good panel members; again, while we would never pressure someone to be on programme, we are keen to invite people who might not otherwise come forward.

Secondly, we are going to apply our argument as set out above. If we have a panel of four panel members and a moderator, then we will assume that we would expect to end up most of the time with two male members and two female members, and that it’s equally likely across the programme as a whole that the moderator be male or female. We will therefore, as a default, look for two good female and two good male panel members, and will keep track of moderators to ensure that we are not ending up with (as all too often happens) mainly male moderators.

One criticism I hear is that programme organisers ought to focus on putting together the best possible panel, and that if such a panel is all-male, then so be it. My response to this is as follows:

  • If our argument is correct, the ‘best possible’ panel is in fact unlikely to be all-male.
  • ‘Best’ is a very open-ended concept anyway; do we mean most experienced, most qualified, most entertaining, most informative (some very well-qualified panel members have actually been very bad at actually expressing themselves) or some other criteria?
  • This is a convention, not a UN conference: a panel does not have to aspire to some Platonic ideal but rather has to meet the standard of being the appropriate mix of interesting, entertaining and informative so that the audience come away feeling that they got good value from it.

Now, there will be some circumstances where it is difficult to achieve gender parity because the pool of relevant expertise is too small for it to itself be likely to be gender-balanced. There may also be panels discussing topics (such as the experience of female writers) where it would be appropriate to have a predominantly or entirely single-gender panel. But such items are likely to be rare and we will still seek to balance overall male/female programme participation.

To conclude, I would note that the single biggest factor that will result in gender parity on panels is if as many women volunteer to take part in programme as men. Historically this has not happened; Eight Squared Con aims to be as inclusive and open as possible in this respect, and we are keen to have as wide and diverse a selection of programme volunteers as possible.

Simon Bradshaw

Programme Coordinator, Eight Squared Con