Science Of Love - Online Personals Watch: News on the Online Dating Industry and Business

« Want To Marry A Millionaire? | Main | Niche Dating Sites Grow Steadily As Mainstream Ones Flail »

Comments

Fernando Ardenghi

"it's not surprising to see that even more scientific methods are being leveraged"
Good News Mr. Brooks!!!

Until now, year 2008, it seems:

* Homophily dominates human attraction.
* what is important in attracting people to one another may not be important in making couples happy.
* "Opposites attract but then they attack", Dr Buckwalter's dixit.
* Persons with opposites characteristics feel attraction between them for short_term mating.
* Persons with similar characteristics are more suitable/qualified for long_term mating.
* Many online daters opt for sameness (when they search) more often than chance would predict.
* Big-5/Big-7 as a main core of compatibility matching is not enough for online dating, because Dating Sites have very big databases (like 10,000,000 of different profiles or even more!!!) and if Big-5/Big-7 is used, the whole precision is less than any person could achieve searching by his/her own!!!
* Conclusions from scientific papers obtained using small samples (small scale of researching) are not valid with real world (large scale of researching).
* The 16PF5 normative personality test available in different languages and any other personality tests (like 15FQ+), are not tests of abilities, so they do not need to be hierarchically structured using Item Response Theory / Rasch analysis. To my best knowledge there is NO Scientific Paper published YET proving/showing that personality scales developed via Item Response Theory measure better or more precise than those developed via Classical Test Theory.
* For DNA matching: Randy Thornhill discovered that women, in an unconscious bid for better genes, will have affairs with men who are more attractive (though perhaps less likely to commit) than their long-term mates (sweaty t-shirt experiment).
* Many Compatibility Tests offered by actual online dating sites are only *fueled* by Marketing Policies/Strategies and not by serious scientific evidence.


My effort is directed to prove that temporal patterns of relationship variables may indeed play a significant role between prospective mates with last stage of temporal patterns: high level on personality* similarity* between mates is the core of relationship stability and satisfaction == Dyadic Success for 26_and_more_years_old_persons interested in serious dating.
personality*: measured with the 16PF5 normative test in different languages (no other actual online dating site is using it!).
similarity*: calculated using quantum math equations with the quantitative method I had invented, named LIFEPROJECT METHOD.

Regards,

Fernando Ardenghi.
Buenos Aires.
Argentina.
[email protected]

James Houran

As someone who has actually conducted and published several academic studies on compatibility and compatibility testing in the real world (as well as helped build commercial and customized matching systems), I can say that Fernando misses the mark on most of his points and arguments here related to psychometric theory and the variables that foster compatibility.

That said, Fernando is correct about one thing: temporal patterns do matter in relationship compatibility. Of course, this is not a new idea -- Levinger, for one, talked about such things nearly 25 years ago! See:

Levinger, G. (1986). Compatibility in relationships. Social Science, 71, 173-177.

Today's dating sites need to partner with assessment vendors who apply modern psychometric theory to modern studies of love and attachment. Unfortunately, the trends speak for themselves:

1. Many, if not most, vendors use Classical Test Theory, which has been outdated since 1960!

2. Many, if not most, vendors continue to focus on a couple's similarity across personality variables, which research has shown to be misguided!

3. Many, if not most, vendors use Classical Test Theory approaches to build, validate and match people on compatibility "tests." Classical Test Theory has been demonstrably shown to be inferior to Item Response Theory since 1960!

4. Many, if not most, vendors neither have the psychometric expertise or theoretical expertise to produce quality and defensible compatibility systems. In fact, many popular matching systems today seem to have little to no scientific rigor behind them!

See:

Houran, J., Lange, R., Rentfrow, P. J., & Bruckner, K. H. (2004). Do online matchmaking tests work? An assessment of preliminary evidence for a publicized ‘predictive model of marital success.’ North American Journal of Psychology, 6, 507-526.

I suspect the proliferation of these other matching methods that claim to use physiological psychology (Chemistry.com, DNA matching, etc.) stems from the need to establish a clear differentiator in this space. After all, to lay people most compatibility tests look alike or at least are assumed to be nearly alike. Thus, sites with newly launched compatibility tests that look like eHarmony's approach come across as passe.

Plus, these "new" methods sound more scientific than old-fashioned questionnaires asking about lifestyle preferences and personality traits. Finally, these "new" approaches might well speak to users at a visceral level -- they play into marketing catch phrases such as "chemistry" and "passion." Thanks to unrealistic Holloywood movies, love songs and kindred media, lay people seem to instinctively equate "compatibility" and "soul mates" with these more emotional, ephemeral notions.

Thanks,

James Houran, Ph.D.
OnlineDatingMagazine.com

Rense Lange, Ph.D.

Fernando's attitude is classic: Most anything goes in psychological testing, and personality testing therefore does not need sound modeling. Given what most dating sites are doing with daters' test data, the casual observer might indeed be forgiven for thinking that anything goes in online compatibility testing.

However, personality testing can in fact be done right and in this case the results are treated with the same respect as ETS and the College Board treats your test scores. As is increasingly witnessed by the literature care is taken to make personality testing internally valid and unbiased as well. I expect that daters' attitudes will soon follow suit, and that amateur tests no longer suffice.

Better testing is not inexpensive because extensive pilot testing are needed, and this may well begin to separate the big dogs from the little dogs. And, yes, doing all this requires considerable expertise in psychometrics and test design. I am professionally involved in educational assessment in several countries, and I have found that people genuinely appreciate high quality unbiased measurement. Tests for online dating are no exception here.

However, online dating also has an entertainment aspect, and psychometrics can provide this as well. For instance, it allows one to identify and exploit idiosyncratic answer patterns that provide information beyond the scores proper. We have improved the process for identifying "exceptional" answers - both in a qualitative and in a quantitative way. As a result, some of our customers now exploit IRT to identify "misfitting" answers to fine-tune the feedback they give to individuals. Also, when applied to "matching" such answers are valuable and truly surprising source of new insights.

So, Fernando's comments are largely supersided as many new things have been developed over the past years. As a result, the old debates have shifted away from quantitative matching to qualitative issues that people find far more interesting and important. At the same time, the additional issues that you mention concerning DNA, sociobiology, etc. can be integrated for more easily within an IRT framework. From my perspective, the old compatibility testing approach is dead. But, fortunately long live the newer approaches that can easily (and in a more meaningful way) be connected to modern developments in a wide variety of areas.

Fernando Ardenghi

Hi Drs. Houran and Lange!!!

To my best knowledge the only three Online Dating Sites using IRT are True, PlentyOfFish and Cybersuitors.

I do not see True, PlentyOfFish and Cybersuitors providing better "compatible prospective mates" than Chemistry, PerfectMatch, eHarmony, OKCupid and others.
All of those sites have the problem that the whole precision is less than any person could achieve searching by his/her own!!!


And as for
Houran, J., Lange, R., Rentfrow, P. J., & Bruckner, K. H. (2004). Do online matchmaking tests work? An assessment of preliminary evidence for a publicized "predictive model of marital success." North American Journal of Psychology, 6, 507-526.
That paper has a confusing title, it was the answer to contradictory results provided by eHarmony with "Helping Singles Enter Better Marriages Using Predictive Models of Marital Success"
Presented at the 16th Annual Convention of the American Psychological Society, May 2004

It is 2008 now, not 2004.
Where are the new papers?
Any new one using PlentyOfFish Chemistry Predictor?

Regards,

Fernando Ardenghi.
Buenos Aires.
Argentina.
[email protected]

James Houran

Hi Fernando!

Always fun to talk to you -- I hope all is well.

1. You wrote,

[Your North American Journal of Psychology]... "paper has a confusing title, it was the answer to contradictory results provided by eHarmony with "Helping Singles Enter Better Marriages Using Predictive Models of Marital Success"
Presented at the 16th Annual Convention of the American Psychological Society, May 2004."

Response: Please read the paper, then you will not be confused. While we specifically examined eHarmony as a case study of poor psychometrics and theoretical rationale, other dating site systems were examined for their scientific backgrounds as well. Moreover, that paper did not present results from the APS conference. That was a separate study using data from Glenn Wilson. It substantiated the pattern of results we observed previously and subsequently published in Dynamical Psychology -- namely that similarity in compatibility is an oversimplification and that complementarity is a better predictor of relationship quality. Further, IRT is needed to take into account idiosyncratic issues in compatibility.


2. You wrote,

"I do not see True, PlentyOfFish and Cybersuitors providing better "compatible prospective mates" than Chemistry, PerfectMatch, eHarmony, OKCupid and others.
All of those sites have the problem that the whole precision is less than any person could achieve searching by his/her own!!!"

Response: How do you possibly know this? Where is your evidence for your claims? Precision in matching has never been an issue; we've been over and over this before. Rather, good measurement underlies everything, so if you have poor measurement, you have poor outcomes. IRT is the statistical gold standard in msurvey/questionnaire/assessment measurement. See our Dynamical Psychology paper (it's online) for an introduction to these issues.

3. You wrote,

"It is 2008 now, not 2004.
Where are the new papers?
Any new one using PlentyOfFish Chemistry Predictor?"

Response: We are working on new studies all the time, as our track record demonstrates. Plus, we work on proprietary applications based on our previous work, whose results have not needed revision yet because our models actually work. You can also expect additional publications soon. However, what I have not seen at any point are any published studies from you proving the internal validity of your method or its predictive validity in predicting relationship quality. I have also not seen any papers proving your arguments in point #2 above.

You're right... it's 2008...where are your studies and papers?

Thanks,

James Houran, Ph.D.
OnlineDatingMagazine.com

Philippine girls

It's always good to know that there are people concern on other people's finding their lifetime partner activity. I'm sure with the help of this great minds it will be more easier, faster and assurance of meeting the best lifetime partner they will have.

The comments to this entry are closed.

Research

  • Dig Deeper - Research Categories

We're Social

  • Facebook  X   Youtube Linkedin