Event Registration Now Open

Our next event will be held at a bar in Tanjong Pagar on a weeknight. Interestingly, the bar is on the second floor of a shophouse and offers outdoor al fresco seating on the roof. I've viewed the space and it looks like a great venue for events. Registration closes in just 10 days, so we recommend registering early.

As always, registration is confirmed only after payment has been received and acknowledged.

Data Analytics Wrap-up

We've had five posts on data analytics related to speed dating, and if you haven't had time to read our previous posts, this final post is a summary of all our findings. Based on our data and analysis, our conclusions are:

1. It's really hard to find someone you like who likes you back. So, dating is invariably going to be frustrating and a lot of hard work. Be persistent and you might be rewarded someday.


2. Even if you aren't the most attractive person, being sociable and approachable helps improve your chances. It really does.


3. Age matters. How it matters, however, is a separate question altogether. Keep this in mind as you search for a romantic connection.


4. If you're going to attend dating events, go to facilitated events. Facilitated events work better than unstructured events.

A Tableau Storyboard on Interests of Participants

Ever been curious about what activities and interests gay men are into? Wonder no more.

In my day job in data analytics, I use and also develop a number of software tools. One of the tools I use is a data visualization software called Tableau.

Today's post will be a little different as it will be much less wordy, and rely on visuals from a Tableau storyboard that I have constructed. So, let's begin.

These are the top 30 interests among participants sorted from most to least popular. By a large margin, the top 3 interests are movies, travelling and listening to music.

Some interests are universal; they have broad appeal across age groups. The top 3 universal interests are listening to music, movies and dining out. If you're on a date, and you find yourself at a loss for conversation topics, talk about any or all of these three things to keep the conversation flowing. Go into specifics to deepen the conversation: artistes, genres, restaurants etc.

Interestingly, across all age groups, the the top 7 interests tend to be fairly similar, but their relative ranking varies significantly by age group. For example, while travelling was the number 2 interest overall and popular with most age groups, it was not a universal interest because it did not figure in the top 7 for the under-25 age group. The under-25 age group was the only age group that did not find travelling important.

Here are the top 7 interests for each age group.

There are some discernible patterns in how interests vary with age. Here are 5 examples. 

Interestingly, team sports are much less popular than individual sports. Maybe gay men just didn't get into the whole male bonding thing back in school. Perhaps some of us were less athletic in our youth, and ended up being the proverbial last kid that's picked for a team; ergo, the lifelong aversion to team sports.

Do Age and Common Interests Matter?

When I first devised my seating algorithm, the prime consideration was coming up with an algorithm that could avoid seating arrangements where participants would meet someone they already knew or had speed dated previously. The secondary consideration was optimizing the seating arrangement so that people whom I thought were more likely to connect would meet each other on a speed date. I designed the seating algorithm to minimize the age difference while maximizing the number of mutual interests between participants who speed dated each other.

Naturally, I wanted to verify whether this hypothesis of age difference and mutual interests actually mattered when it came to improving the chances of a match. I analyzed the data in the following way:

1. I looked at the pairs of participants who speed dated each other. For each pair, I calculated the absolute age difference and the number of mutual interests between them. 


2. Across all speed dating pairs, I calculated the average absolute age difference and the average number of mutual interests. 


3. I repeated the procedure in (1) and (2) separately for two subsets of the pairs of participants who speed dated each other:
   1. Those pairs where only one participant was interested in the other (I call this the “expressions of interest” group).
   2. Those pairs where both participants were interested in each other, resulting in a match (the “matches” group).


4. Finally, I compared the average absolute age difference and the average number of mutual interests across the three groups: the speed dating group, the “expressions of interest” group, and the matches group.

It should be noted that I carried out the analysis only for guys who speed dated each other. Carrying out the analysis for guys who mingled at the event but did not speed date each other would not have been useful because not everyone meets everyone else at the mingling. Some participants leave right after the speed dating portion of the event.

More importantly, and this may come as a surprise to some of you, the incidence of matches that occur between guys who mingle but do not speed date each other is actually very, very low. Across all 3 events I have held, there are only been 5 such matches. The low number of such matches precludes any credible analysis. In contrast, there have been 18 matches between speed dating pairs.

Now, let’s look at what I found from the analysis.

Age Difference

As it turns out, there is a difference in average age difference across the groups. The chart below shows that the age difference is significantly smaller in the matches group as compared to the speed dating or expressions of interest groups.

A less thorough analysis would conclude here that the smaller the age difference between a dating pair, the more likely they are to match. That would be a premature conclusion however.

One thing that stood out while I was organizing events was the age distribution of participants who come to 7for7 events. The majority of participants who attend my events are in the 20-40 age group. Only about 20% of participants are in the above 40 age group. This distribution of ages, however, is not preserved when we look at the matches group. The chart below shows the data.

When we move from the speed dating group to the matches group, the proportion of participants above the age of 40 halves. Only about 10% of participants with a match are above the age of 40. Obviously, since the above-40 group accounts for a lower proportion of the matches group, the 20-30 and 30-40 age groups must take up a correspondingly larger proportion in the matches group.

The increase is greatest for the 20-30 age group: an increase from 36% to 42% represents an increase of 17% (we’re taking ratios here, not absolute differences). The increase is more muted for the 30-40 age group (44% to 47% represents a 7% increase).

So, what does all this mean?

It means that it is age itself, rather than the age difference, that is the important factor in explaining the smaller absolute age difference in the matches group as compared to the speed dating or expressions of interest groups. Older guys receive disproportionately fewer matches compared to the younger guys, particularly those in their 20s, and that means that guys in the matches group are generally younger guys matching up with each other. This is confirmed when I plot the same age difference data not as a simple column chart, but as a box plot, shown below.

The distribution of age differences gets tighter as we move from the speed dating group, to the expressions of interest group, to the matches group. For those who need a primer on what box plots are, here’s the Wikipedia entry.

I hate to sound ageist when I say that age matters, but the data is the data. [And as for those who like to claim that age is "just a number"...well, my numbers - the data I have collected, say otherwise.] While age matters, how and why it matters is a completely different question altogether. One that I cannot answer definitively with the data I have.

There could be a plethora of reasons why age matters. Maybe older guys are clearer in their minds what they’re looking for, so they tend to be more selective and receive fewer matches as a result.

Maybe the older guys who come to my events are less physically attractive; age certainly affects how a person looks. In which case, the difference that I am observing in the data has less to do with age than with youth and physical attractiveness. Age is simply a proxy indicator.

Or maybe things work pretty much the same in the gay world as they do in the straight world. Guys, no matter their age, prefer to date youthful partners. Straight men well into their 50s still prefer the company of younger women in their 20s rather than someone closer to their own age. Gay men might well be the same.

If you are above the age of 40 and reading this, I am not discriminating against you. Frankly, I'm not that far off from the big 4-O myself. I am interpreting what I see in the data.

Does this mean that if you're above 40, you shouldn't attend speed dating events? No. There are lots of reasons why you might choose not to attend speed dating events, but this shouldn't be one of them. That would be like blaming the mirror for why you look the way you do. Also, the data above shows that guys in their 40s can still get matches, albeit at a lower frequency.

Mutual Interests

It may seem counterintuitive, but my data showed that the number of mutual interests didn’t seem to matter. We would expect that the number of mutual interests would increase going from the speed dating group to the expressions of interest group to the matches group. In fact, there was little discernible difference in the average number of mutual interests between the three groups.

I can think of a number of reasons to explain this counterintuitive result:

1. People don’t take the information they provide on their interests and activities seriously when they fill up the registration form. For example, something like a third of all guys put down board games as an activity, but in my experience, very few people actually regularly play board games. This is essentially a problem with bad data.


2. Just because people have activities in common doesn’t mean they will hit it off on the basis of those activities. It depends on what kinds of activities we're talking about.
   
   For example, something like 50% of the guys who register on my website indicate “gym” as an activity. But if you’re a gay man, going to the gym regularly is practically a necessity of the gay condition. Many guys genuinely enjoy going to the gym, but an even larger number go to the gym because they feel they have to build a body that will garner some attention. It’s probably not likely that guys will bond over an activity that is regarded as a chore by some.


3. Some interests are so banal as to render them useless in predicting attraction between dates. For example, listening to music, movies and dining out are the top 3 activities. More than three quarters of participants indicated interest in these activities. Come to think of it, they're probably the top 3 activities of Singaporeans in general.


4. I collect data on activities and interests that participants claim they engage in (whether or not they actually engage in these activities is a separate question altogether). I do not, however, collect data on the intensity of these interests and activities.
   
   Based on my analysis of the data, most guys list too many activities rather than their most important activities and interests. In retrospect, I should have limited the number of activities and interests that registrants could select to perhaps their top 5.
   
   It seems more likely for two guys who have 1 deep mutual interest (particularly if it is an unusual or uncommon interest)  to hit it off than if they had 10 shallow mutual interests. This reason is probably the likeliest one for the number of mutual interests to apparently not have an impact on the probability of a match.

A more detailed post on the interests of participants, and by extension, single gay men, will follow next week.