Gender Bias: A Twitter Folly?
There’s a claim made by researchers at Harvard Business School that men are followed disproportionately on Twitter. That may be true on a straightline basis. But there may be more – or less — here than the authors make out. The fact is, we can’t tell yet.
A first order question is, “What is the correct denominator in this rate? What’s the expected value of the rate of male-male follow?” Then, “What’s the observed deviation and to what extent is it attributable to gender?” We don’t think the authors are in a position to answer that yet, based on the data they’ve offered.
How to sort this through? Best thing to do here is find a real world use case where we can test our intuitions about what might be going on in the authors’ data, and the claims made for it.
As we think about how we come to follow others on Twitter, there are three or four obvious vectors. To explain, we’ll choose one, and use a man named “Harry” as an example.
Here’s the vector: Harry wakes up to find that someone is now following him on Twitter. He could (but doesn’t always) follow back. Harry faces a decision: should he follow back? In this case, we’d want to know, when making his choice, does Harry show a bias in favor of following males vs. females? Secondly, if he does appear to bias towards male or female, is it due to maleness/femaleness, or to some underlying trait?
The first thing we’d want to know is what the proportion of males-to-females is in the group of people newly following Harry – this is the universe of his potential choice. If it’s 30-70, then all things equal we should expect Harry to follow back at that rate if he’s gender blind.
That gives us the expected value of Harry’s follow-back rate for males vs females: it’s 30-70.
Anything other than that is a deviation that may or may not be attributable to chance – and if not to chance, then to some other factors, including (possibly) gender bias.
But the authors of the HBS note seem to suggest that the 30-70 (or whatever its equivalent) is in fact prima facie good evidence of a gender-based selection bias on Harry’s part – when in fact, for a 30-70 population it’s exactly perfect. So we need to know what the expected value is, and what the observed deviation from this expected value is – if any. If Harry should follow 30-70 and does, he’s perfect: zero observed bias. But if Harry ends up following 50-50, ok – something is possibly going on.
But we can’t stop there. Just because it’s 50-50 doesn’t mean there’s gender bias. Because when someone follows Harry, Harry actually *reads* what these folks do – and Harry won’t follow folks whose profile says they do Internet marketing, for example. It may turn out that the ranks of Internet marketers are disproportionately male, or female. So when Harry elects not to follow, what might at first blush look like a gender bias is actually a bias against a profession.
So, Harry might have plenty of biases, but gender might not be one of them. What looks like a gender bias is in fact first and foremost an expected value; and after than, much of the deviation could be explained by non-gender factors like profession.
This points to the importance of understanding the vectors and dynamics of the follow/don’t follow decision in Twitter. There are more vectors than this one use case, and plenty of dynamics. In the current analysis, the data offered could be explained by a host of factors, most not explicated by the authors.
Until they are it’s a little early to point the fickle finger of gender bias at Twitter.
We’ll save for another post taking a look at these important questions, too:
– How does the fact that 80% of users follow or are followed by one or more in fact test the capacity of a user base to understand the service? Do we have some expectation about the probability of an occurrence of a tie, and if so, why?
– A large proportion of Twitter users keep their gender identification ambiguous. To what extent does this alter the authors’ conclusions – did they adjust for this?
– As the authors assert, all well developed online social media services have a contribution pattern that roughly follows power-law or exponential distributions. Does the fact that Twitter falls within the extreme bounds of these distributions point to the fact that it is still settling into an equilibrium?
(With Andrew Conway/cross-posted to http://www.drewconway.com/zia/)
Interesting article. There is a study here (http://blogs.harvardbusiness.org/cs/2009/06/new_twitter_research_men_follo.html) on gender trends on twitter. Worth looking at..
I have always been intrigued by these same questions, Shuying, and esp #3. We do get cues from the photos and names, and stripped of those we may have different proclivities to follow. That would be a great experiment to run – couple of groups, same twitter content, different avatars. That would give you insight into how gender, race, age play into the decision to follow. But you’d still need to parse it into a variety of other follow behaviors as those factors will have greater or lesser salience, I’m guessing, depending on the vectors – that is, how the decision to follow is framed/the choice as presented initially.
Laura – That’s an interesting suggestion – I’m betting there’s good research on how people perceive the “gender” of brands. And your suggestion that the behavior of women vs men could be explained by underlying rational vs (implicitly) biased gender-based behavior is a great insight. Many thanks for that.
That’s interesting, Bill, and another factor worth noting – how “newbie” vs. “veteran” status might affect the follow/no-follow decision. It may have little or a lot of salience – I notice, for example, some veterans following everyone who follow them; and some who follow no one; but of course a “newbie” strategy is to follow as many as possible in the hopes of being followed back. These “follow vectors” have many factors going on; gender could be one, but the decision to follow is complicated and not easily explained by appealing to it. Thanks for your note.
I would offer a one feed anecdotal rejoinder: my sense is followers are more balanced after the initial start. This is based on operating the official University of Arkansas athletic department feed (@ArkRazorbacks). Once we got out of the first 200-300 followers — which were decidedly male — I have noticed as we have broaden into a “mainstream” of more Razorback fans joining I am seeing at least 50-50, if not slightly more female, followers joining the list.
This may be my own bias in trying to find a vehicle more open to both gender’s participation to broaden our base, but when compared to the traditional participatory media (message boards, old school full blogs with comment) there is a vastly higher number of women willing to follow. A typical message board might be 90-10 (even 95-5) on female identification (again, I will go with the point made that many women will avoid gender identity).
If we are even running at the 70-30 mentioned, that’s significant in my corner of the world.
Just some thoughts.
[…] Gender Bias: A Twitter Folly? There’s a claim made by researchers at Harvard Business School that men are followed disproportionately on Twitter. That may be true on a straightline basis. But there may be more – or less — here than the authors make out. The fact is, we can’t tell yet. […]
Thanks for highlighting this report. Here are my 2 cents –
1. My initial reaction to all of this is similar to Shuying’s first thought – what about brands on twitter? Are they considered male/female or neuter? Wouldn’t it be interesting to do a study on how/if people perceive a gender of brands?
2. As a woman, I agree with the idea that women may “have more stringent thresholds for reciprocating relationships.” Since the early days of the internet, we have been warned to be smart about who we connect with online. This may also account for the report’s suggestion that “men [may] find the content produced by women less compelling (because of a lack of photo sharing, detailed biographies, etc.).”
Again, thanks for sharing your thoughts on this study!
Shuying – Your points are solid. The fact is that folks follow others for many reasons; gender is one of many, perhaps – but the authors’ analysis doesn’t rinse out, test, and then repack the factors to see which are most active, and with who. They have no analysis or hypothesis about following behavior, and so can’t test it with any reasonable set of tools.
Kelcy, good questions. I find the original study suspect because it oversimplifies the demographic of twitter users.
Some thoughts on this:
1. How does one account for people following group/company twitter accounts?
2. Apart from gender and age, would there be any other factors for following? And how strongly correlated to follow rate would these factors be versus gender or age?
3. Assuming we remove markers like photos and names aside (which some people have), how do we quickly discern gender and age from content?
Making the big assumption that people are somewhat like me, I would agree with Drew that Twitter relationships are mainly content driven. When I decide to follow someone, it’s a function of 1. knowing them in real life aka friends and associates OR 2. having met new people offline and using twitter and other online networks as a means of keeping in touch; OR 3. that person or entity having a common interest or profession which comes across because of the kind of recent content they post. In this case, this person or entity would have come to my attention either because they replied to a tweet or they followed me first or they used words that came up in my saved searches that pertains to topics of interest to me.
Kelcy,
I think your first point is a very strong one, and something that deserves a lot of attention for those that want to leverage Twitter for the dissemination of information. In a very non-scientific way, I can say that name and avatar are very important. Nearly all of the “What to/not to do on Twitter” post you read by the so-called social media gurus has some mention of the importance of both of these things. It would be a difficult but very interesting analysis to examine this in a scientific way.
On the second, from my own personal experience I do not think there is a bias in any direction based on age. I think Twitter relationships are mainly content driven, but of course this is my assumption and something that should be tested!
Good evaluation of both the study and follow-on research areas. I think this clearly illustrates the need for being careful to report any research instead of rushing to report on the “sexy” technologies and associated cultural issues.
A couple of questions come to mind if you do any follow-up evaluations/study. 1. If gender is kept ambiguous through avatar and name, can content help provide a gender clue that allows the user to decide who to follow. 2. Is there an age factor in choosing who to follow – are younger males more likely to follow females or does this matter. Or is it perhaps tied with profession where women may be more prevalent in certain professions or equally balanced depending on age.