Picture your ideal partner - what do they look like? What do they talk about? What are their interests? What do they do for a living?
It turns out that, when you actually come face-to-face with someone, all of that stuff goes right out of the window. What you use to check out a real-life, right-in-front-of-you-now, potential partner has pretty much nothing to do with what you use to construct your ideal partner.
Take one easy example of this. The cliché that men go for looks and women go for earning potential - that pretty much does hold up when you ask people about their ideal partner, or even ask them to rate written-descriptions of potential partners. And yet, put real people in front of them and this cliché falls apart. Men and women are no different in the way looks or money inspires romantic interest.
How do we know this? By testing the match between peoples' ideal partner traits and their romantic interest in imaginary or real people in different contexts.
Studies using potential partners you're never going to meet
In this set-up you present people with photographs and a written description of potential romantic partners. In this kind of situation, if you know what their ideal partner is like you can predict which of the written descriptions of potential partners they will show the most interest in. People will choose the partners that best fit their ideal.
However, put people in a speed-dating situation, where they get to meet real-life potential romantic partners for a few minutes at a time, and things start to get less clear. In fact, in this situation, there's hardly any match between someone's ideal partner characteristics and the characteristics of the people they show an interest in. This even holds up when you ask them about their ideal partner characteristics in the context of a speed-date: How much do you think earning potential will matter in your decision to ‘yes’ or ‘no’ someone after your 4-min date?
Why does meeting someone change your mind?
You can start answer this question by giving someone a photo and a written description of a potential romantic partner and then having them meet the same person face-to-face.
Imagine you're in a lab and you've been given a range of photos with written descriptions of potential romantic partners. You rate them for how much you'd be interested in dating them. At this point, the way you rate them fits really well with your ideal partner template. If you've said you are generally attracted to people who are outspoken - you go for the descriptions which convey a person who seem to be outspoken.
But then we take just one of those potential partners - one which you rated as very desirable - and you have a 4 minute face-to-face meeting with them. Unknown to you, this person is an actor, sticking to a script so you get no new information about them - nothing more than you already found out from their written description.
At this point, everything changes.
All the characteristics of your ideal partner completely fail to predict whether you will find this person attractive or not by the end of the face-to-face encounter. The ideal partner template worked on paper, but not in real life.
And there are some interesting things which happen here. For example, imagine you said that your ideal partner is someone who is outspoken. On paper, that was straight-forward - there were statements like, I like to speak my mind. But face-to-face, you don't find this person outspoken ...you find them tactless. In other words, identical traits, scripted in identical ways, can be interpreted completely differently by different people during a live interaction.
The thing is, with real people, much of the attraction process is happening at an implicit (kind of unconscious) level. That’s why, when you give people an implicit test about the attributes of their ideal partner, you get a much more realistic pattern. This implicit test is done by flashing words on a screen and asking participants to hit the space bar as quickly as they can when they think the word is synonymous with attractiveness.
For example, 'confident' is flashed up and you hit the space bar (or not). This will all happen within 750–1,000 milliseconds, so you don't get a chance to think about it consciously - that's why it's tapping into your unconscious.
By looking at these split-second decisions, and tiny changes in reaction speed, you can assess the preferences people have before they get a chance to use their conscious 'filter'. It turns out that these implicit, spontaneous reactions tend to paint a totally different picture of your ‘ideal partner’ than the one you talk about consciously. What’s more, results from this kind of test are much more accurate at predicting who you will find attractive in a real-life interaction.
Internal Working Model
So that leads us all the way back to the internal working model. We don't consciously know what we want in a romantic partner because there are powerful unconscious forces at work. These forces are forged from our first relationship - with our parents.
These interactions became our understanding of what relationships are and how they work. This, in turn, informs our understanding of our own self-worth. And it includes what we expect - what we find familiar and comfortable - in a romantic partner. Being aware of this can help to break cycles of dysfunction - help to stop you from repeatedly seeking refuge in the familiarity of bad relationships.
The Predictive Validity of Ideal Partner Preferences: A Review and Meta-Analysis. Eastwick, Paul, et al. s.l. : Psychological Bulletin, 2013, Vol. April.