Skip to main content
SAGE
Search form
  • 00:22

    [MUSIC PLAYING]

  • 00:26

    PRIEST: What is true love?Love is patient.Love is kind.Love is not a mathematical equation.

  • 00:34

    NARRATOR: Wait.[SCREECH] Says here that it is.

  • 00:37

    MAN: Science-based algorithm matching.

  • 00:40

    NARRATOR: Algorithms.Maths.It's what dating websites use to find your perfect match.Look!Worldwide each month almost 150 million peoplego online to find dates.Some of them, looking for the one.Even Australia's help websites boastover a million members each.But is there any evidence that arming Cupid's bow

  • 01:02

    NARRATOR [continued]: with algorithms, not arrows, will hit true love?[MUSIC PLAYING]Well, let's go back a step to find out.[BIRDS CHIRPING]If Lady Luck's been a bit slow to step in and find youa soulmate, why wouldn't you trust online mathematiciansto get the job done?

  • 01:23

    NARRATOR [continued]: All they ask of you is data.

  • 01:26

    MAN: Do you like rock music?How often do you brush your teeth?

  • 01:29

    WOMAN: Do you get angry easily?

  • 01:30

    BOYFRIEND: Would you just shut up!

  • 01:33

    NARRATOR: Your answers are comparedwith those of potential partners, often on the basisthat similarity is the key.

  • 01:40

    ELI FINKEL: The idea isn't crazy, right?We all have this intuition that we end uphappier in relationships where we have personalityand values in common with the person that we're marrying.

  • 01:51

    NARRATOR: So maybe matching like with like could work.Do you want to be the center of attention?

  • 01:58

    BOTH: Yes.

  • 02:02

    NARRATOR: Or maybe not.

  • 02:03

    ELI FINKEL: The problem is that the evidence over the last fiveyears is really clear that similarity on those thingsjust isn't a very strong predictor of whether peopleend up happy in their relationships.Large-scale studies recently haveshown that we do tend to match up and marry people whoare more similar to us on variableslike race and age and other sociodemographic variables.

  • 02:24

    ELI FINKEL [continued]: But what you really want to know is are peoplehappier in relationships with somebodywho is more similar to them rather than less similar?And the answer is they're about equally happy in both cases.

  • 02:36

    NARRATOR: Right.But similar or not, can't you just let the algorithmknow what you want?

  • 02:41

    GIRLFRIEND: I'd like him to not wantto be the center of attention.That's what I'm for.

  • 02:51

    ELI FINKEL: Most of us go through life assumingthat we know what it is that we're lookingfor in a romantic partner.But it assumes that we're accurate.The problem is we keep finding in study after studythat you don't really have that much insight into those things.

  • 03:06

    NARRATOR: Advanced algorithms alsofactor in that not all the data they collectis equal, at least to you.Each question is given a weighting by youfor its importance to you.That way his belief in God can still trump an annoying habit.

  • 03:23

    BOYFRIEND: Annoying?

  • 03:24

    NARRATOR: Some matching algorithms claim to mine datafrom the way you answer the question, not justthe answer itself.So loud and outgoing people--

  • 03:34

    GIRLFRIEND: Am I the life of the party?Whoo!

  • 03:37

    NARRATOR: --will answer the questionnairedifferently to those who are shy and reserved.That creates individual patterns of datawhich may reveal hidden personality traits that evenyou didn't know yet.Matching algorithms can also compare potential couplesagainst thousands of real couplesthey've already studied, some still happily together

  • 03:59

    NARRATOR [continued]: and others have separated.It all sounds relatively scientific.Potential compatibility based on real compatibility.But just because they are using scientific methods doesn'tmean their algorithms are scientific.

  • 04:13

    HARRY REIS: Well, there's lots of researchthat looks at what are the characteristics of the coupleswho've done well and the couple who don't do so well.The problem is that that kind of workis not the same as predicting which couples will do wellin advance.Nearly all of the information that youwould need to predict well is simply not availablebefore two people have met.

  • 04:35

    HARRY REIS [continued]: Things like life stresses, having children,coping with illness.And therefore, your access to the thingsthat really matter in predicting long-term successin a relationship is simply not there.

  • 04:48

    ALL: Ahh.

  • 04:49

    NARRATOR: So if a couple are truly perfect for each other,is there any algorithm they can trust to rate them as the mostmutually compatible couple?Well, the scientists say no.

  • 05:02

    BOTH: Really?

  • 05:03

    NARRATOR: Yep.That's despite some of the industry's claims.

  • 05:07

    SCIENTIST: We've conducted years of extensive researchand have an algorithm that delivers lasting relationships,not just great dates.

  • 05:15

    NARRATOR: But how?

  • 05:16

    SCIENTIST: Wouldn't you like to know?

  • 05:17

    NARRATOR: Yes.

  • 05:18

    SCIENTIST: Oh.

  • 05:19

    HARRY REIS: None of the algorithmsthat is currently available has been testedby valid scientific means.That means that on some fundamental level,we cannot ascertain whether these methods really work.

  • 05:30

    ELI FINKEL: There's a second issuetoo, which is in principle, how likely is itthat an algorithm that focuses on information about people whohave never met can actually predict,not only short-term but also long-term compatibility?And we delved deeply into this question in a recent articleand concluded that the odds of them

  • 05:50

    ELI FINKEL [continued]: being successful with these methods are close to zero.Hey!Where's everyone gone?

  • 05:58

    HARRY REIS: I don't know.[CHURCH BELLS RINGING]

  • 06:01

    NARRATOR: Though there's no compelling evidencethat mathematical algorithms did as they claimed.

  • 06:06

    PRIEST: Have you finished?[INAUDIBLE]

  • 06:08

    NARRATOR: But it is a numbers game.And the authors did say that the websites are still a great wayto meet new people.

  • 06:14

    MAN: Finally.[SWOOSH]

Video Info

Publisher: Australian Broadcasting Corporation

Publication Year: 2013

Video Type:Documentary

Methods: Algorithms, Probability

Keywords: dating, courtship, and marriage; online dating; personality assessment; romantic love

Segment Info

Segment Num.: 1

Persons Discussed:

Events Discussed:

Keywords:

Abstract

A look inside how dating websites use complex math to find a user's perfect match.

Looks like you do not have access to this content.

Algorithm of Love

A look inside how dating websites use complex math to find a user's perfect match.

Copy and paste the following HTML into your website