Beste online dating site belgie

Extracting multistage screening rules from online dating activity data

Extracting multistage screening rules from online dating activity data.,Associated Data

This paper presents a statistical framework for harnessing online activity data to better understand how people make decisions. Building on insights from cognitive science and decision theory, we develop a discrete choice model that allows for exploratory behavior and multiple stages of decision making, with different rules enacted at each stage 14 rows ·  · Significance. Online activity data—for example, from dating, housing search, or social This paper presents a statistical framework for harnessing online activity data to better understand how people make decisions. Building on insights from cognitive science and  · This paper presents a statistical framework for harnessing online activity data to better understand how people make decisions. Building on insights from cognitive science and  · Extracting Multistage Screening Rules From Online Dating Activity Data by Elizabeth Bruch, Fred Feinberg, Kee Yeun Lee published in Proceedings of Extracting ... read more

Behav Processes , 90 2 , 05 Mar Cited by: 10 articles PMID: Krupp DB. Arch Sex Behav , 37 1 , 01 Feb Cited by: 9 articles PMID: Contact us. Europe PMC requires Javascript to function effectively. Recent Activity. Search life-sciences literature 41,, articles, preprints and more Search Advanced search. This website requires cookies, and the limited processing of your personal data in order to function.

By using the site you are agreeing to this as outlined in our privacy notice and cookie policy. Elizabeth Bruch Department of Sociology, University of Michigan, Ann Arbor, MI ; Center for the Study of Complex Systems, University of Michigan, Ann Arbor, MI ; ebruch umich.

edu Search articles by 'Elizabeth Bruch'. Bruch E 1 ,. Fred Feinberg Ross School of Business, University of Michigan, Ann Arbor, MI ; Department of Statistics, University of Michigan, Ann Arbor, MI ; Search articles by 'Fred Feinberg'. Feinberg F 2 ,. Kee Yeun Lee Department of Management and Marketing, Hong Kong Polytechnic University, Kowloon, Hong Kong. Author profile Search articles by ORCID Lee KY 3. Affiliations 1 author 1. Department of Sociology, University of Michigan, Ann Arbor, MI ; Center for the Study of Complex Systems, University of Michigan, Ann Arbor, MI ;.

Ross School of Business, University of Michigan, Ann Arbor, MI ; Department of Statistics, University of Michigan, Ann Arbor, MI ;. Department of Management and Marketing, Hong Kong Polytechnic University, Kowloon, Hong Kong.

Share this article Share with email Share with twitter Share with linkedin Share with facebook. Abstract This paper presents a statistical framework for harnessing online activity data to better understand how people make decisions. We find that mate seekers enact screeners "deal breakers" that encode acceptability cutoffs. Our statistical framework can be widely applied in analyzing large-scale data on multistage choices, which typify searches for "big ticket" items.

Free full text. Proc Natl Acad Sci U S A. Published online Aug doi: PMCID: PMC PMID: Elizabeth Bruch , a, b, 1 Fred Feinberg , c, d and Kee Yeun Lee e. Elizabeth Bruch a Department of Sociology, University of Michigan, Ann Arbor, MI, ; b Center for the Study of Complex Systems, University of Michigan, Ann Arbor, MI, ; Find articles by Elizabeth Bruch.

Fred Feinberg c Ross School of Business, University of Michigan, Ann Arbor, MI, ; d Department of Statistics, University of Michigan, Ann Arbor, MI, ; Find articles by Fred Feinberg. Kee Yeun Lee e Department of Management and Marketing, Hong Kong Polytechnic University, Kowloon, Hong Kong Find articles by Kee Yeun Lee.

Author information Copyright and License information Disclaimer. a Department of Sociology, University of Michigan, Ann Arbor, MI, ;. c Ross School of Business, University of Michigan, Ann Arbor, MI, ;. d Department of Statistics, University of Michigan, Ann Arbor, MI, ;. Email: ude. hcimu hcurbe. Edited by Susan T. Fiske, Princeton University, Princeton, NJ, and approved July 13, received for review November 14, Copyright notice. This article has been cited by other articles in PMC.

Associated Data Supplementary Materials Supplementary File. pdf 3. Significance Online activity data—for example, from dating, housing search, or social networking websites—make it possible to study human behavior with unparalleled richness and granularity. Keywords: choice modeling, noncompensatory behavior, mate selection, computational social science. Modeling Noncompensatory, Heterogeneous, Multistage Choice Processes: An Application to Online Mate Choice Fig. Open in a separate window.

Data and Results Our data consist of over 1. Table 1. Fit statistics for proposed, nested, and alternate models. Men Women L 2 df BIC L 2 holdout L 2 df BIC L 2 holdout One-class model Linear , , 55, , 1, , , Quadratic , , 48, , 1, , 92, Cubic , , 47, , 1, , 91, Splines , , 47, , 1, , 90, Five-class model Linear , , 48, , , 96, Quadratic , , 44, , , 88, Cubic , , 44, , , 87, Splines , , 43, , , 86, No.

of users 1, No. of observations , 56, , , Different Rules at Different Stages. Sharp Cutoffs. Deal Breakers. Heterogeneous Behavior. Discussion Online activity data throw open a new window on human behavior. Materials and Methods We describe two key features of our modeling strategy: first, how we allow for multiple decision stages; and second, our strategy for estimating the model coefficients.

Modeling Multiple Decision Stages. Model Estimation. Human Subjects. SI Appendix. Supplementary Material Supplementary File Click here to view.

Acknowledgments We thank Dan Ariely for helping us obtain the data used in this project. Footnotes The authors declare no conflict of interest. This article is a PNAS Direct Submission.

References 1. McFadden D. Conditional logit analysis of qualitative choice behavior. In: Zarembka P, editor. Frontiers of Econometrics. Academic; New York: Newell A, Simon H. Human Problem Solving. Prentice Hall; Englewood Cliffs, NJ: Payne J, Bettman J, Johnson E. The Adaptive Decision Maker. Cambridge Univ Press; Cambridge, United Kingdom: Cowan N. The magical mystery four: How is working memory capacity limited, and why? Curr Dir Psychol Sci. McCelland G, et al.

Effects of choice task on attribute memory. Organ Behav Hum Decis Process. Payne J. Task complexity and contingent processing in decision making: An information search and protocol analysis.

Organ Behav Hum Perform. Hastie T, Tibshirani R. Generalized Additive Models. Hastie T, Tibshirani R, Friedman J. The Elements of Statistical Learning. Springer; New York: De Boor C. A Practical Guide to Splines.

Harrell FE. Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis. Kim J, Menzefricke U, Feinberg FM. Capturing flexible heterogeneous utility curves: A Bayesian spline approach. Manage Sci. Hitsch G, Hortaçsu A, Ariely D. Matching and sorting in online dating. Am Econ Rev. Lewis K. The limits of racial prejudice. Proc Natl Acad Sci USA. Lin K, Lundquist J.

Mate selection in cyberspace. Yancey G, Emerson M. Does height matter? An examination of height preferences in romantic coupling. J Fam Issues. Kendrick D, Keefe R. Age preferences in mates reflect sex differences in human reproductive strategies. Behav Brain Sci. Fernandez R, Weinberg N. Sifting and sorting: Personal contacts and hiring in a retail bank. Am Sociol Rev. Bills D. Sociol Q. Hoekstra M. The effect of attending the flagship state university on earnings: A discontinuity-based approach.

Rev Econ Stat. Zimmerman S. The returns to college admission for academically marginal students. J Labor Econ. Talarchek M. Sequential aspects of residential search and selection. Urban Geogr. Clark W, Smith T. Housing market search behavior and expected utility theory: 2. The process of search.

Environ Plan A. Shafir E. The Behavioral Foundations of Public Policy. Princeton Univ Press; Princeton: Hauser J. Consideration set heuristics.

J Bus Res. Young DS. Mixtures of regressions with changepoints. Stat Comput. Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences. Full text links Read article at publisher's site DOI : Smart citations by scite. ai include citation statements extracted from the full text of the citing article.

The number of the statements may be higher than the number of citations provided by EuropePMC if one paper cites another multiple times or lower if scite has not yet processed some of the citing articles. Explore citation contexts and check if this article has been supported or disputed. Human social sensing is an untapped resource for computational social science.

Galesic M , Bruine de Bruin W , Dalege J , Feld SL , Kreuter F , Olsson H , Prelec D , Stein DL , van der Does T Nature , , 30 Jun Cited by: 3 articles PMID: Inference of a universal social scale and segregation measures using social connectivity kernels. Hoffmann T , Jones NS J R Soc Interface , 17 , 28 Oct Cited by: 2 articles PMID: PMCID: PMC Free to read. Computational Social Science and Sociology.

Edelmann A , Wolff T , Montagne D , Bail CA Annu Rev Sociol , 46 1 , 01 Jul Cited by: 1 article PMID: PMCID: PMC Free to read. Impressions of HIV risk online: Brain potentials while viewing online dating profiles.

Schmälzle R , Imhof MA , Kenter A , Renner B , Schupp HT Cogn Affect Behav Neurosci , 19 5 , 01 Oct Cited by: 0 articles PMID: Choice Set Formation in Residential Mobility and Its Implications for Segregation Dynamics. Bruch E , Swait J Demography , 56 5 , 01 Oct Cited by: 2 articles PMID: PMCID: PMC Free to read. Data Data behind the article This data has been text mined from the article, or deposited into data resources. BioStudies: supplemental material and supporting data. Similar Articles To arrive at the top five similar articles we use a word-weighted algorithm to compare words from the Title and Abstract of each citation.

Mate preferences do predict attraction and choices in the early stages of mate selection. Li NP , Yong JC , Tov W , Sng O , Fletcher GJ , Valentine KA , Jiang YF , Balliet D J Pers Soc Psychol , 5 , 05 Aug Cited by: 32 articles PMID: The science of online dating. Can the application of science to unravel the biological basis of love complement the traditional, romantic ideal of finding a soul mate?

Frazzetto G EMBO Rep , 11 1 , 01 Jan Cited by: 1 article PMID: PMCID: PMC Free to read. Aspirational pursuit of mates in online dating markets. Collective decision making by rational individuals. Mann RP Proc Natl Acad Sci U S A , 44 :EE, 15 Oct Cited by: 6 articles PMID: PMCID: PMC Free to read. Computational mate choice: theory and empirical evidence. Castellano S , Cadeddu G , Cermelli P Behav Processes , 90 2 , 05 Mar Cited by: 10 articles PMID: Review.

Through evolution's eyes: extracting mate preferences by linking visual attention to adaptive design. Krupp DB Arch Sex Behav , 37 1 , 01 Feb Cited by: 9 articles PMID: Review. Funding Funders who supported this work. HHS NIH National Institute of Child Health and Human Development 2  Grant ID: KHD 1 publication Grant ID: RHD 1 publication.

NICHD NIH HHS 3  Grant ID: P2C HD publication s Grant ID: R24 HD publication s Grant ID: K01 HD 5 publication s. Claim to ORCID Open PDF Get citation.

Follow us News blog Technical blog Twitter YouTube. About About Europe PMC. Joining Europe PMC. Tools Tools overview. ORCID article claiming. Journal list.

Grant finder. External links service. Annotations submission service. Developers Developer resources. Articles RESTful API.

Grants RESTful API. API case studies. SOAP web service. Annotations API. OAI service. Bulk downloads. Developers Forum. Help Help using Europe PMC. Search syntax reference. Contact us Helpdesk. Tech blog. Laptops market is a highly competitive market in the present scenario. Due to rapid changes in the field of technology day by day , new notebook brands comes in to existence, also, the consumer desires and expectations are moving on.

The research paper further uses the analytic hierarchy process AHP approach to find out which features are more important for the students decision-making process when purchasing a laptop, this paper will help student to clarify the important criteria which affect their decision making process for their next laptop as well as helping the producing companies to know how university students evaluate their laptops brand in terms of the selected criteria.

Mahmoud Maqableh. Computer-Aided Civil and Infrastructure Engineering. Rick Andrews. Fred Feinberg. John Payne. Steve Gaskin. Nadya Khoirina. James Agarwal. Amitav Chakravarti. Yoel Mansfeld. Víctor Flores. victor flores. Chris Janiszewski. JoLaine Draugalis. Maria Oskarson. Joffre Swait. James Helgeson. Monica Martinussen , David Hunter. Olivier Toubia. Ling Rothrock. José Kerstholt. Jose Luis Bermudez. Patrick Brockett. Joseph K Goodman.

Christian Seidl. Journal of Experimental Psychology: Learning, Memory, and Cognition. Gerd Gigerenzer. Dan Zakay. Benedict Dellaert. International journal of geographical information systems. Ahmet Sayar. Robert Paul Jones. Holger Dietrich , Anja Dieckmann. Guo Hu. Eduard Brandstätter. Divakaran Liginlal. Scott Klemmer. Randall Westgren. Guy Hochman. International Journal of Culture, Tourism and Hospitality Research.

Joan Buckley. Log in with Facebook Log in with Google. Remember me on this computer. Enter the email address you signed up with and we'll email you a reset link. Need an account? Click here to sign up. Download Free PDF.

Extracting multistage screening rules from online dating activity data. Abstract This paper presents a statistical framework for harnessing online activity data to better understand how people make decisions. Related Papers. Decision Support Systems Predicting and explaining patronage behavior toward web and traditional stores using neural networks: a comparative analysis with logistic regression.

Download Free PDF View PDF. Geographical Analysis Optimal Stopping and Residential Search Behavior. Broken Rationality: The Ecological Rationality of Simple Inference Heuristics. J Consum Aff Enhancing Consumer Choice: Are We Making Appropriate Recommendations? What Makes You Click? Saudi Online Courtship Script: Qualitative Interviews of Using Matrimonial Websites among Saudi Users. AHP Study on laptops selection Criteria. Modern Applied Science Qualitative Interviews of Using Matrimonial Websites among Saudi Users.

Extracting multistage screening rules from online dating activity data Elizabeth Brucha,b,1, Fred Feinbergc,d, and Kee Yeun Leee a Department of Sociology, University of Michigan, Ann Arbor, MI ; bCenter for the Study of Complex Systems, University of Michigan, Ann Arbor, MI ; cRoss School of Business, University of Michigan, Ann Arbor, MI ; dDepartment of Statistics, University of Michigan, Ann Arbor, MI ; and e Department of Management and Marketing, Hong Kong Polytechnic University, Kowloon, Hong Kong Edited by Susan T.

Fiske, Princeton University, Princeton, NJ, and approved July 13, received for review November 14, This paper presents a statistical framework for harnessing online adapt these behaviorally nuanced choice models to a variety of activity data to better understand how people make decisions.

To we develop a discrete choice model that allows for exploratory that end, here, we present a statistical framework—rooted in de- behavior and multiple stages of decision making, with different cision theory and heterogeneous discrete choice modeling—that rules enacted at each stage. Critically, the approach can identify if harnesses the power of big data to describe online mate selection and when people invoke noncompensatory screeners that eliminate processes. Specifically, we leverage and extend recent advances in large swaths of alternatives from detailed consideration.

The model change point mixture modeling to allow a flexible, data-driven is estimated using deidentified activity data on 1. A nonparametric account of heterogeneity re- tentially different rules at each. Our statistical framework can ciency of a particular attribute, regardless of their merits on be widely applied in analyzing large-scale data on multistage others.

We apply our computational social science modeling framework to mate-seeking behavior as observed on an online dating site. In doing so, we empirically establish whether V ast amounts of activity data streaming from the web, smartphones, and other connected devices make it possible to study human behavior with an unparalleled richness of detail.

behavioral data: strings of choices made by individuals. Taking full advantage of the scope and granularity of such data requires Significance a suite of quantitative methods that capture decision-making processes and other features of human activity i.

Discrete associations among variables rather than behavior of human choice models, by contrast, can provide an explicit statistical repre- actors. Harnessing the full informatory power of activity data sentation of choice processes.

However, these models, as applied, requires models that capture decision-making processes and often retain their roots in rational choice theory, presuming a fully other features of human behavior. Our model aims to describe informed, computationally efficient, utility-maximizing individual 1.

mate choice as it unfolds online. It allows for exploratory be- Over the past several decades, psychologists and decision havior and multiple decision stages, with the possibility of dis- theorists have shown that decision makers have limited time for tinct evaluation rules at each stage. This framework is flexible learning about choice alternatives, limited working memory, and and extendable, and it can be applied in other substantive do- limited computational capabilities.

As a result, a great deal of mains where decision makers identify viable options from a behavior is habitual, automatic, or governed by simple rules or larger set of possibilities. For example, when faced with more than a small handful of options, people engage in a multistage choice process, Author contributions: E. designed research; E. in which the first stage involves enacting one or more screeners performed research; E. and F. analyzed data; and E. wrote the paper. to arrive at a manageable subset amenable to detailed processing and comparison 2—4.

These screeners eliminate large swaths of The authors declare no conflict of interest. options based on a relatively narrow set of criteria. This article is a PNAS Direct Submission. history is available, such as for frequently purchased supermarket html. However, these models are not directly applicable to major 1 To whom correspondence should be addressed.

Email: ebruch umich. problems of sociological interest, like choices about where to live, This article contains supporting information online at www.

what colleges to apply to, and whom to date or marry. Such splines consist on Site Partners Viewed of linear functions joined at specific points called knots. The key im- Fig. The multistage mate choice process. pediment to efficient estimation is that the space of all possible knots is typically very large for our final model, on the order of in fact , and therefore, brute force exhaustive search is out of the question. Thus, one needs a powerfully efficient way to Modeling Noncompensatory, Heterogeneous, Multistage explore potential knot configurations Materials and Methods.

Choice Processes: An Application to Online Mate Choice Experimentation with our data and prior empirical work 11 Fig. Utility functions then, among those browsed, to whom to write. One may, for example, browse a narrow age , and dummy variables for intrinsically categorical attributes band of ages and then be relatively indifferent to age thereafter e. Specifically, in standard notation 9, 11 , when writing.

Empirical studies suggest that the choice process utility VijB for user i browsing B superscript; an analogous commences using cognitively undemanding, cutoff-based criteria formulation holds for writing W potential mate j is operating on a small number of attributes e. Decision theorists iii sum of utilities of L discrete attributes coefficients γBil and distinguish screeners that are conjunctive deal breakers from covariates xBijl. Large positive or negative slope dicates a set of qualities where any one suffices.

With V on although offering great flexibility to fit data well, typically encode a logit scale, a difference of three represents a difference in odds two procedures at odds with how actual humans seem to process and thereby, probability on the order of being 20 times less large amounts of information.

From the standpoint of capturing noncompensatory decision rules, there are three problems with this approach. recalling it at will, and weighting it judiciously that is, computa- First, polynomial functions conflate nonlinearity with nonmonotonicity.

However, as in tionally —for the decision maker are easier to model and estimate Fig. linear and monotonic. Higher-order polynomials allow for a wider range of functional forms but at a cost of greater imprecision and intrinsic multicollinearity.

Second, non- For example, the compensatory model can be readily estimated compensatory decision rules impose a screener denoting the acceptability cutoff for a using standard regression-based techniques; even allowing for the given attribute. Third, polynomials are notoriously sen- straightforward with standard software.

However, noncompensa- sitive to outliers, so that the resulting shape of the function in any given region may be driven by observations with values far from that region. Our aim is to allow the func- tory decision rules that allow for i abrupt changes in the relative tional form to be driven primarily by local information and not by asymptotics. We show desirability of potential partners as an attribute passes outside an that our model both fits better and tells a different substantive story compared with acceptability threshold and ii an attribute to have a disproportionate more conventional specifications.

A B Data and Results Our data consist of over 1. For categorical attributes, dummies capture potential interactions. Dif- ferences likely matter more at low vs. Illustration of how choice model captures alternative decision rules.

Both BMI and age are, therefore, accommodated as differences A depicts a linear compensatory rule; B depicts a nonlinear but compensa- on a log scale [e.

tory one. C is a conjunctive rule where being outside of the range δ1ik and Table 1 reports the fits of two-stage models with and without δ2ik acts as a deal breaker, and D is a disjunctive rule where being greater than δ2ik acts as a deal maker. heterogeneous decision rules latent classes as well as models that allow for conventional representation of continuous covariates i.

Based on standard fit metrics [Bayesian Information which may be large enough that no other attribute combination Criterion BIC and L2], the proposed model with five latent classes SOCIAL SCIENCES can overcome it: a deal breaker.

for both men and women fits the data better than all nested models Fig. These out of sample estimates reaffirm that a model allowing for nonsmooth response and heterogeneity out- utility slope is large enough to render all other attributes and performs other more traditional specifications. In addition to su- their differences irrelevant, a nonlinear but ostensibly compen- perior fit, our model captures features of decision processes that are satory rule can function as deal breaker or deal maker.

Similar distorted by traditional approaches. Additional details are in SI logic applies to the L categorical attributes: the dummy slope Appendix, Section S4. coefficient γBil determines whether the attribute l functions as deal Although our models produce many results, we focus here on breaker or deal maker. All results reported i nonlinear, even noncompensatory, evaluative processes; in the main text are significant at the 0. For our specific application to online dating, it allows for distinct but statistically intertwined accounts of both the brows- Different Rules at Different Stages.

Distinct subsets of attributes ing and writing stages and explicit quantification of the relative are implicated at the browsing and writing stages. For example, importance placed on observable attributes included in online when men select among women, age plays a greater role in the profiles.

Importantly, decision rules need not be prespecified: the browsing stage. Consider Fig. The model also accommodates exploratory younger women. Among women, age matters in both browsing and stochastic behavior, thus guarding against a deal breaker on, and writing, but its effects can vary across stages. For example, as say, age being tautologically inferred as the oldest or youngest we see in Fig. BMI also figures differently into browsing erences and deal breakers can stand out.

and writing decisions. Thus, it observed data. For example, if a particular site user wrote only to people above a certain age, we might declare that being below that age is a deal breaker. It would also ignore important statistical information: if The site skews toward a specific demographic subgroup with distributions, discussed that respondent wrote to other users, 99 who were over 50 y old and 1 who was 25 y below, that closely match the general online mate-seeking population.

The greater old, the model should not merely spit out that a deal-breaker age was anything below number of women in our sample reflects site base rates. A nondisclosure agreement the much lower figure. Bruch et al. Fit statistics for proposed, nested, and alternate models Men Women 2 2 2 L df BIC L holdout L df BIC L2 holdout One-class model Linear , , 55, , 1, , , Quadratic , , 48, , 1, , 92, Cubic , , 47, , 1, , 91, Splines , , 47, , 1, , 90, Five-class model Linear , , 48, , , 96, Quadratic , , 44, , , 88, Cubic , , 44, , , 87, Splines , , 43, , , 86, No.

of users 1, No. of observations , 56, , , seems that women can never be too thin to write to; conditional who are substantially older. The median woman in this class is on browsing. around 40 y old; she is 2.

Our model Sharp Cutoffs. The results for height, as shown women very different from themselves. These men are, on av- in Fig. smooth changes. Overall, women seem to prefer men who are 3—4 In our final set of results, we show that analogous analyses can in taller across the board, with substantial drop offs for men below be distorted by traditional statistical modeling approaches. Be- this cutoff.

This finding is consistent with prior research showing cause unobserved heterogeneity is standard in most statistical that women prefer a partner who is not taller than she is in heels software packages, an appropriate comparison is between our With regard to age Fig. Given that these men are, on Fig. Any such available in SI Appendix, Section S4. First, we see that, although crisp criteria would be smoothed over in a model that captured nonlinearities via polynomial specifications.

Deal Breakers. Age differences are the biggest deal breaker. Even A B within the bulk of observations i. The model can also locate deal breakers in categorical covariates, although this is not unique to its framework. In online dating, one that stands out is not demographic but an act of omission: failing to provide a photo. Both men and women are roughly 20 times less likely to browse someone without a photo, even after controlling for all other attributes in the model age, education, children, etc.

C D Nearly as strong is smoking behavior: among those who do, non- smokers are nearly 10 times less likely to be browsed and, there- fore, smoking is evidently a decisive screen. In short, we find clear evidence of deal-breaking behavior, although the strength of ef- fects varies across the revealed classes.

Note that, although none of these may be truly inviolable, they are practically insurmount- able within the observed range of available covariates. Heterogeneous Behavior. By allowing for unobserved heteroge- neity, we can both assess what behaviors hold across the board and identify subclasses of users pursuing unique mate selection Fig.

The probability of browsing and writing someone of a given value of age relative to the probability of browsing or writing someone of equal strategies.

potential matches. The y axis shows the associated probability ratio for both For example, although most women pursue partners who are browsing and writing. The probability of browsing and writing someone of a given value Fig.

The probability of browsing and writing someone of a given value of body mass relative to the probability of browsing or writing someone of of height relative to the probability of browsing or writing someone of equal body mass.

A and B show results for men, and C and D show results for equal height. The x axis is height difference in inches between the user that for potential matches.

Either your web browser doesn't support Javascript or it is currently turned off. In the latter case, please turn on Javascript support in your web browser and reload this page. Proceedings of the National Academy of Sciences of the United States of America , 30 Aug , 38 : DOI: Free to read. b Center for the Study of Complex Systems, University of Michigan, Ann Arbor, MI, ;. e Department of Management and Marketing, Hong Kong Polytechnic University, Kowloon, Hong Kong.

Author contributions: E. designed research; E. performed research; E. and F. analyzed data; and E. wrote the paper. Online activity data—for example, from dating, housing search, or social networking websites—make it possible to study human behavior with unparalleled richness and granularity. However, researchers typically rely on statistical models that emphasize associations among variables rather than behavior of human actors.

Harnessing the full informatory power of activity data requires models that capture decision-making processes and other features of human behavior. Our model aims to describe mate choice as it unfolds online. It allows for exploratory behavior and multiple decision stages, with the possibility of distinct evaluation rules at each stage. This framework is flexible and extendable, and it can be applied in other substantive domains where decision makers identify viable options from a larger set of possibilities.

This paper presents a statistical framework for harnessing online activity data to better understand how people make decisions. Building on insights from cognitive science and decision theory, we develop a discrete choice model that allows for exploratory behavior and multiple stages of decision making, with different rules enacted at each stage. Critically, the approach can identify if and when people invoke noncompensatory screeners that eliminate large swaths of alternatives from detailed consideration.

The model is estimated using deidentified activity data on 1. A nonparametric account of heterogeneity reveals that, even after controlling for a host of observable attributes, mate evaluation differs across decision stages as well as across identified groupings of men and women. Vast amounts of activity data streaming from the web, smartphones, and other connected devices make it possible to study human behavior with an unparalleled richness of detail.

Taking full advantage of the scope and granularity of such data requires a suite of quantitative methods that capture decision-making processes and other features of human activity i. Discrete choice models, by contrast, can provide an explicit statistical representation of choice processes. However, these models, as applied, often retain their roots in rational choice theory, presuming a fully informed, computationally efficient, utility-maximizing individual 1. Over the past several decades, psychologists and decision theorists have shown that decision makers have limited time for learning about choice alternatives, limited working memory, and limited computational capabilities.

As a result, a great deal of behavior is habitual, automatic, or governed by simple rules or heuristics. For example, when faced with more than a small handful of options, people engage in a multistage choice process, in which the first stage involves enacting one or more screeners to arrive at a manageable subset amenable to detailed processing and comparison 2 — 4.

These screeners eliminate large swaths of options based on a relatively narrow set of criteria. Researchers in the fields of quantitative marketing and transportation research have built on these insights to develop sophisticated models of individual-level behavior for which a choice history is available, such as for frequently purchased supermarket goods.

However, these models are not directly applicable to major problems of sociological interest, like choices about where to live, what colleges to apply to, and whom to date or marry. To that end, here, we present a statistical framework—rooted in decision theory and heterogeneous discrete choice modeling—that harnesses the power of big data to describe online mate selection processes. Our approach allows for multiple decision stages, with potentially different rules at each.

We apply our modeling framework to mate-seeking behavior as observed on an online dating site. In doing so, we empirically establish whether substantial groups of both men and women impose acceptability cutoffs based on age, height, body mass, and a variety of other characteristics prominent on dating sites that describe potential mates. The pool of potential partners includes all relevant users active on the site. Informative features of mate choice behavior are revealed at each stage, and choices made at the browsing stage restrict which alternatives are subsequently available.

One may, for example, browse a narrow band of ages and then be relatively indifferent to age thereafter when writing. Empirical studies suggest that the choice process commences using cognitively undemanding, cutoff-based criteria operating on a small number of attributes e. Our proposed framework can accommodate an arbitrary number of sequentially enacted winnowing stages.

Here, we focus on two intrinsic to the medium: browsing and writing. At each stage, choice is governed by one or more possible decision rules, which are uncovered by the model. Alternately, they may impose noncompensatory screening rules, in which they browse only those profiles meeting some threshold of acceptability on one or more attributes.

Decision theorists distinguish screeners that are conjunctive deal breakers from those that are disjunctive deal makers ; the former indicates a set of qualities where all must be possessed, and the latter indicates a set of qualities where any one suffices. Even sophisticated modeling approaches in social research 7 , 8 , although offering great flexibility to fit data well, typically encode two procedures at odds with how actual humans seem to process large amounts of information.

However, noncompensatory decision rules that allow for i abrupt changes in the relative desirability of potential partners as an attribute passes outside an acceptability threshold and ii an attribute to have a disproportionate effect on choice outcomes over some region of values lack anything approaching a turnkey solution.

Such splines consist of linear functions joined at specific points called knots. If knot positions are known in advance—for example, a downturn in utility for men under a given height—estimating the slopes of each of the component linear functions is straightforward and quick; however, here, we seek to identify both the slopes and the knots themselves, which are highly nontrivial The key impediment to efficient estimation is that the space of all possible knots is typically very large for our final model, on the order of 10 62 in fact , and therefore, brute force exhaustive search is out of the question.

Thus, one needs a powerfully efficient way to explore potential knot configurations Materials and Methods. Specifically, in standard notation 9 , 11 , utility V i j B for user i browsing B superscript; an analogous formulation holds for writing W potential mate j is. With V on a logit scale, a difference of three represents a difference in odds and thereby, probability on the order of being 20 times less likely that the potential match will be browsed or written to, which may be large enough that no other attribute combination can overcome it: a deal breaker.

Similar logic applies to the L categorical attributes: the dummy slope coefficient γ γ i l B determines whether the attribute l functions as deal breaker or deal maker. Illustration of how choice model captures alternative decision rules. A depicts a linear compensatory rule; B depicts a nonlinear but compensatory one.

C is a conjunctive rule where being outside of the range δ 1 i k and δ 2 i k acts as a deal breaker, and D is a disjunctive rule where being greater than δ 2 i k acts as a deal maker. In summary, the model accommodates three key constructs: i nonlinear, even noncompensatory, evaluative processes; ii heterogeneity across individuals; and iii multistage choice behavior. For our specific application to online dating, it allows for distinct but statistically intertwined accounts of both the browsing and writing stages and explicit quantification of the relative importance placed on observable attributes included in online profiles.

The model also accommodates exploratory and stochastic behavior, thus guarding against a deal breaker on, say, age being tautologically inferred as the oldest or youngest value observed for each individual. Our data consist of over 1. For categorical attributes, dummies capture potential interactions.

Differences likely matter more at low vs. Both BMI and age are, therefore, accommodated as differences on a log scale [e. Table 1 reports the fits of two-stage models with and without heterogeneous decision rules latent classes as well as models that allow for conventional representation of continuous covariates i.

Based on standard fit metrics [Bayesian Information Criterion BIC and L 2 ], the proposed model with five latent classes for both men and women fits the data better than all nested models e.

To safeguard against overfitting, we also assess goodness of fit using a holdout sample consisting of men and women who joined the site immediately after the estimation period. These out of sample estimates reaffirm that a model allowing for nonsmooth response and heterogeneity outperforms other more traditional specifications.

In addition to superior fit, our model captures features of decision processes that are distorted by traditional approaches. Additional details are in SI Appendix , Section S4. Although our models produce many results, we focus here on key features of mate choice behavior that would be, as a whole, inaccessible with alternative modeling approaches: i different rules at different decision stages, ii sharp cutoffs in what attribute values are desired or acceptable, iii invocation of deal breakers, and iv heterogeneity in behavior.

All results reported in the main text are significant at the 0. Distinct subsets of attributes are implicated at the browsing and writing stages. For example, when men select among women, age plays a greater role in the browsing stage. Consider Fig. Among women, age matters in both browsing and writing, but its effects can vary across stages. For example, as we see in Fig.

BMI also figures differently into browsing and writing decisions. Thus, it seems that women can never be too thin to write to; conditional on browsing. The probability of browsing and writing someone of a given value of age relative to the probability of browsing or writing someone of equal age. The y axis shows the associated probability ratio for both browsing and writing. The probability of browsing and writing someone of a given value of body mass relative to the probability of browsing or writing someone of equal body mass.

The y axis shows the associated probability ratio. By identifying sharp cutoffs in acceptability criteria, the model can identify norms or rules that would be difficult to extract using traditional methods. The results for height, as shown in Fig. Overall, women seem to prefer men who are 3—4 in taller across the board, with substantial drop offs for men below this cutoff. This finding is consistent with prior research showing that women prefer a partner who is not taller than she is in heels With regard to age Fig.

Any such crisp criteria would be smoothed over in a model that captured nonlinearities via polynomial specifications. The probability of browsing and writing someone of a given value of height relative to the probability of browsing or writing someone of equal height.

The x axis is height difference in inches between the user and potential match. Age differences are the biggest deal breaker. Even within the bulk of observations i.

The model can also locate deal breakers in categorical covariates, although this is not unique to its framework. In online dating, one that stands out is not demographic but an act of omission: failing to provide a photo.

Both men and women are roughly 20 times less likely to browse someone without a photo, even after controlling for all other attributes in the model age, education, children, etc. Nearly as strong is smoking behavior: among those who do, nonsmokers are nearly 10 times less likely to be browsed and, therefore, smoking is evidently a decisive screen. In short, we find clear evidence of deal-breaking behavior, although the strength of effects varies across the revealed classes.

Extracting multistage screening rules from online dating activity data,Significance

 · This paper presents a statistical framework for harnessing online activity data to better understand how people make decisions. Building on insights from cognitive science and This paper presents a statistical framework for harnessing online activity data to better understand how people make decisions. Building on insights from cognitive science and decision theory, we develop a discrete choice model that allows for exploratory behavior and multiple stages of decision making, with different rules enacted at each stage  · Extracting Multistage Screening Rules From Online Dating Activity Data by Elizabeth Bruch, Fred Feinberg, Kee Yeun Lee published in Proceedings of Extracting It allows for exploratory behavior and multiple decision stages, with the possibility of distinct evaluation rules at each stage. This framework is flexible and extendable, extracting Building on insights from cognitive science and decision theory, we develop a discrete choice model that allows for exploratory behavior and multiple stages of decision making, with 14 rows ·  · Significance. Online activity data—for example, from dating, housing search, or social ... read more

Guy Hochman. Keywords: choice modeling, noncompensatory behavior, mate selection, computational social science. Open in a separate window. Decision theorists distinguish screeners that are conjunctive deal breakers from those that are disjunctive deal makers ; the former indicates a set of qualities where all must be possessed, and the latter indicates a set of qualities where any one suffices. In the first stage, the probability that the i th mate seeker will consider browse the j th option at a particular time, which for simplicity, we leave unsubscripted can be written as a binary choice model, which we operationalize as softmax i.

All results reported in the main text are significant at the 0. Help Help using Europe PMC. Journal of Behavioral Decision Making Ignorance or integration: the cognitive processes underlying choice behavior. Funding Funders who supported this work. Abstract This paper presents a statistical framework for harnessing online activity data to better understand how people make decisions.

Categories: