1. They justify their conclusions with graphs and numbers
2. They tackle seemingly contradictory results
3. Sometimes they provide more than a single number to justify their conclusions.

You should read the full article as they address several more issues than we’ll get into here (and they also provide sample photos of the types of images they’re referring to), but we’ll call out some highlights that illustrate these points.

Numerical Justification.

Is it better to smile or not? The standard advice is to smile while looking into the camera. But how effective is that really? It apparently depends on whether you’re male or female. Below is their answer for men. I find their chart a bit confusing as the bars below the line at first blush indicate that you get negative hits… how can someone “take back” a dating request? What they’re doing though is graphing against the average. I’ve redrawn it on the right which may make more sense. And of course the results are that no eye contact for men is almost always better.

Results Conflict With Common Sense

They dispel several common pieces of advice regarding dating photos in their article. One example involves those odd-angle self-portrait shots, what they term “MySpace shots”, from their article:

The universally maligned “MySpace Shot,” [is] taken by holding your camera above your head and being just so darn coy.

The graph below shows how effective these are for women. In short, very very effective. They even do a little control testing of this, just to verify that it wasn’t some other feature of these shots that made them so effective (read the article for details) but the “MySpace shot” still came out on top. Again, they’re graphing these results against the average, but with the previous warning this graph should be a little easier to read.

When Yes/No Won’t Do.

Most of the examples they gave provide information to the effect of:

For best results, men should look away from the camera and not smile.

They provide a sense of how much more effective their suggested strategies are by comparing them with other strategies. But this information may be too homogeneous for general use. “Does it really apply to me?” So for a few types of photos OkCupid dug a little deeper. In one case they analyzed how effective “showing skin” was in the photos. For men this tended to be showing off their ripped abs (assuming they had any). But instead of providing a simple comparison against the average or other types of photos, they broke it down by age. The results are below. In short, if you’re young and have ripped abs it’s better to show them off, but it’s not always that much better.

Conclusion.

Every newspaper has an Entertainment section and why not us? But our entertainment may come with ulterior motives. Here’s ours. After seeing this data, how likely are you to abandon your common sense wisdom regarding effective dating photos in favor of a technique that data says works better? Is the ideology of smiling and looking into the camera any different from any other ideology? Shouldn’t any ideology fall under the weight of contradictory evidence? If you change your strategy will your friends call you a “flip-flopper” or are you instead holding fast to a higher principle, one of science?

Alternatively, are you swayed by these results because you believe them to be unbiased? After all, the motivations of OkCupid here are clear, they just want more people to meet each other and they’re reporting on what appears to do that. It’s not like they’re charging more for those “MySpace shots”. Yes, reporting bias does exist and is a real concern, but that cuts both ways, may even cut 3-ways: left-bias, truth and right-bias. Your task is to find the truth and not to blanketly assume your favorite bias is always true. Our task is to help you do that. One way is exemplified by the ripped abs chart above. Rich data, in this case showing where showing the 6-pack helps a lot and where it only helps a little, is believable even if OkCupid had something to gain from pushing 6-pack abs shots. They state themselves that it’s not always tremendously effective. And the results of the “MySpace shot” is another means of verifying unbiased reporting. The results are so stunning that other researchers may very well try to repeat the result, and furthermore the results are so stunning that young women may be able to independently verify its veracity by testing their own dating pics. A marginal result may be hard to independently verify, but a result like this should be fairly easy.

One of the principal objects of theoretical research in any department of knowledge is to find the point of view from which the subject appears in its greatest simplicity. – Willard Gibbs

MANTRACKER

Perhaps the least known of these 2 is “Mantracker” which is a reality TV show where a tracker/ranger/horseman tries to track down two average folks in the woods. The show starts with the “prey” and Mantracker (yes, that’s what they call him on the show) separated by two kilometers when a flare is fired from the prey’s position. This signals both the prey to start heading out toward their destination and Mantracker to start heading toward the last known site of the prey (the flare location). Mantracker is on horseback and has a partner who’s versed in the local area while the prey are given a map and a compass and otherwise can bring anything they want. The prey need to reach their goal usually about 40 kilometers away in 36 hours or less without getting caught by Mantracker.

What makes the show interesting from our point of view is how Mantracker explains himself to the viewing public and how the show illustrates this. He can spot overturned rocks and tell you that this was just recently overturned … the reason? It’s wet on the top, signaling that it’s been flipped over and hasn’t had time to be dried out by the sun. He also points out a trail of bent grass which is sometimes highlighted by the special effects crew to make it stand out more for the home viewer. (The image to the right is the best example I can find of this online, it’s of a footprint in the mud. This particular example doesn’t need highlighting/enhancement, the footprint was obvious on its own, but it illustrates the technique.)

If the average person were to stumble upon some of the trails Mantracker does they might not notice anything amiss. What leads Mantracker to focus on the items that he does is his years of experience at tracking people lost in the woods. Mantracker does not directly teach you how to survey a scene and spot the signs of a trail, but instead shows you why he is choosing a particular trail. You get the sense that at any time you could ask him “why do you think the prey are going this way?” and he’d show you overturned rocks, fresh footprints (and explain why they’re fresh) or a line of broken twigs. Lesser, more egotistical “experts” may simply say “because I’m an expert and you’re not” and leave it at that or go into a jargon laced incomprehensible explanation. Mantracker shows you his reasoning simply and concisely¹.

While I’m a fan I’m not an ardent fan of the show. There are camera crews following both Mantracker and the prey and these undoubtedly help each see each other (it would be nice if the show gave us a behind-the-scenes look at this facet). Also Mantracker frequently comes very close to catching his prey within the first half of the show, and yet somehow they get away. Sometimes the terrain is so rugged that the prey can easily evade the horses, but it would be more believable if the producers would adjust their air schedule so that the occassional episodes where the prey are caught early could be combined into a single hour episode instead of (apparently) letting the prey escape just so they can fit every show into its hour format.

Mantracker is a Canadian series in its 4th season. It airs on the Science Channel in the US.

LIE TO ME

The other show is “Lie to Me” a crime drama staring Tim Roth as the CEO of The Lightman Group, a firm that specializes in detecting lies from people’s facial tics and body language². Each episode usually involves one or more members of the Lightman Group applying their lie detection skills in a crime related way. Sometimes it’s a standard crime, but sometimes not³. The show takes some shortcuts with how these type of experts would work in real life⁴, but two things make it noteworthy here. One is that they point out the facial ticks and body language gestures to the other parties as they happen. Sometimes the plot has the Lightman group reviewing recorded testamony and they’re able to freeze frame and slow motion through the tics and gestures explaining them all the time. Sometimes the plot involves the Lightman group detecting lies in real time where they similarly point out to the would-be liar that he’s just given himself away. But more compellingly, just before a commercial break following one of the above scenes they cut in a rapid fire sequence of shots of famous politicians or celebrities caught in lies betrayed with similar tics or gestures. (Some images below are from the show, others are from unrelated sources on the web but are similar to what is done in the show.)

While not as educational as Mantracker, the visual techniques described above do get you thinking about this in a similar manner. A novice may not be able to spot these tics and gestures on their own, but if an expert shows these tics and gestures on a freeze frame instant replay and explain what’s going on it’s believable.

While researching this article⁵ I found a website run by someone who claims to be a natural. A “natural” is Lightman’s term for someone who was born with the ability to spot these ticks and gestures instinctively, unlike Tim Roth’s character who studied his way to lie detection.

What’s interesting about the blogger is she mentions in her about section how she was unable to explain her ability to friends until she read an article on MSNBC about the work of real-life researchers similar to Lightman. This points out how, even when the expert is motivated to explain themselves in a simple way they may still find it hard to do. Her website discusses cases in the news where there’s enough footage of the people involved that she can render an opinion on how honest the subject is. Armed now with the knowledge of how her skills operate in a subconscious level she explains why she believes or disbelieves various subjects. The difficulty an expert has in presenting his case in a simple manner is not due to any inherent complexity of the issue. It may instead be due to the fact that years of training have pushed the skills of an expert down into the subconscious level and raising them back up to the conscious level is a genuinely hard task⁶.

Lie to Me is entering it’s second season on the Fox network in the US.

WHAT, NO NUMB3RS?

Finally, some may be surprised that I did not include Numb3rs in this list. There are several reasons for this. Firstly, the fancy graphics and explanations in Numb3rs are aimed, not at the specific issue, but instead at the general issue, the mathematical problem that’s similar to the criminal problem at hand. Secondly the proof of the success of the mathematical way isn’t in the inherent veracity of the mathematical technique, but instead in its ability to correctly predict. When the prediction comes true the proof is final. While I love Numb3rs for showing how math may be applied in a variety of unusual circumstances, the techniques they’re using are more akin to the standard “scientific method”. Finally, the FBI lets Charlie (the math whiz on Numb3rs) do his thing largely because it doesn’t restrict what else they are doing. If instead the FBI had to choose between doing things Charlie’s way OR doing things their old standard way I would imagine that Charlie’s explanations would not be sufficiently convincing.

Notes

• 1 I’m sure we’ve all experienced snobbish experts hiding their reasoning from us in similar ways, doctors, coworkers, etc.
• 2 In this sense “The Lightman Group” reminds me a bit of the Second Foundation
• 3 One episode was an accident investigation occurring during the rescue operation where one of the victims knew negligence played a part.
• 4 Like drawing a baseline before jumping to conclusions that a particular tic signals a lie.
• 5 Yea, I research even entertainment articles.
• 6 This is similar to Malcom Gladwell’s theme in Blink.