The Geography Of Online Dating Fraud

2y ago
17 Views
2 Downloads
371.11 KB
7 Pages
Last View : 17d ago
Last Download : 2m ago
Upload by : Baylee Stein
Transcription

1The Geography of Online Dating FraudMatthew Edwards , Guillermo Suarez-Tangil† , Claudia Peersman Gianluca Stringhini† , Awais Rashid , Monica Whitty‡ Cyber Security Group, Department of Computer Science, University of Bristol, UK† Information Security Group, Department of Computer Science, University College London, UK‡ Cyber Security Centre, WMG, University of Warwick, UKAbstract—This paper presents an analysis of online datingfraud’s geography. Working with real romance scammer datingprofiles collected from both proxied and direct connections,we analyse geographic patterns in the targeting and distinctcharacteristics of dating fraud from different countries, revealingseveral strong markers indicative of particular national originshaving distinctive approaches to romance scamming. We augmentIP geolocation information with other evidence about the datingprofiles. By analysing the resource overlap between scam profiles,we discover that up to 11% of profiles created from proxiedconnections could be assigned a different national origin onthe basis of text or images shared with profiles from directconnections. Our methods allow for improved understandingof the origins of dating fraud, beyond only direct geolocationof IP addresses, with patterns and resource sharing revealingapproximate location information which could be used to targetprevention campaigns.I. I NTRODUCTIONThe online romance scam is one the most prevalent forms ofmass-marketing fraud in many Western countries. False datingprofiles are created by scammers as a prelude to a sustainedfalse romance, during which the target is repeatedly defraudedof large sums of money. The impact on victims in terms ofboth monetary loss and emotional harm can be substantial.However, technical analysis of the methods used by thesescammers remains sparse, with few quantitative analyses ofattacks and attackers.Previous work has explored victim understanding of thescam process in interview settings [1], text reuse in romancescammer approaches via Craigslist [2] and strategies deployedin an anonymous Chinese dating site [3]. A major unaddressedhurdle for combatting this fraud is understanding its trueglobal origins, as misrepresentation of location is common.Uncertainty about location and international legal obstaclescan hinder investigation and prosecution.The locations scammers give in their profile are typicallyregarded as being as false as the profile picture, calculated toattract the interest of their targets [1]. Dating sites record theIP addresses used by scammers in creating and accessing theirprofiles, and may compare those addresses to blacklists or usethe IP geolocation (especially when compared to the profile’sdeclared location) to inform a judgement about the likelihoodthat a profile is genuine. In response, most scam profileauthors make use of web proxies to disguise their IP addressconnection information, and so they appear to be using aconnection from the location given in their profile information.Dating sites are predictably countering by banning access totheir site via known web proxies and similarly allocated IPblocks. There are however limitations to the effectiveness ofthese countermeasures, with privately hosted or intentionallydisguised proxies escaping the checks of proxy listing services.The real location, even at a national level, of the creators ofthe scam profiles is of interest both to law enforcement andfor other preventative efforts – not only for the purpose ofidentifying that a given profile is a scam, but for following upwith appropriate countermeasures once a significant origin ofscams has been identified (e.g., contacting local law enforcement, funding targeted preventative campaigns). This paper isthe first study we know of to address this topic.In this paper, we use a dataset of real online dating scamprofiles which includes profiles created via both proxied anddirect connections. We set out to answer the following researchquestions: Where does dating fraud come from? What does IPgeolocation evidence tell us about the origins of profilescreated via direct connections, and how does this connectto the locations given in the profiles? Do profile elements get reused internationally? Doesreuse suggest different origins for dating profiles? Canwe complement IP geolocation by examining profileelements being reused between unproxied and proxiedconnections? Does dating fraud from different regions presentdifferent characteristics? Do countries tend towardscertain forms of romance scam in a distinctive manner?In Section II below, we describe the available data, and noteits limitations. In Section III below, we outline the significantorigin countries within the SOURCE dataset, and the nationallocations those profiles present. In Section IV we look attext and images being shared between romance scam profiles,and what these patterns suggest about the PROXY dataset.In Section V, we examine the major scam origin nationsto identify patterns in other elements of the profiles, beforeconcluding in Section VI with a discussion of the policyimplications of this analysis.II. DATA S OURCEThe data used in this paper comes from a public onlinedating scamlist maintained at scamdigger.com, which offersup romance scammer profile data for public awareness. Anexhaustive collection of the 5,402 scam profile instances, ascollected during March 2017, was examined with respect totwo sources of geographic information:

21) The location given in the scammer dating profile information.2) The IP address used to create the profile, as reported bythe dating site.Other profile elements of note include the age, gender,occupation, marital status and self-description, which areanalysed in detail in related work. Of the two sources ofgeographic information, the former was recorded as a string,often specifying location to a city level. This was geocodedto lat/lon coordinates and a standard format through queriesto the Open Street Map’s Nominatim service1 . For the sakeof brevity, the locations given in profiles are referred to as thepresented locations.The IP address information was mapped to a locationthrough the use of a geolocation service 2 , providing both coordinates and structured address information. Some 368 recordscontained no IP address information and were excluded,leaving 5,194 profile instances. Of the IP addresses used,many (67.9%) have been identified as known web proxies orVPN end-points by the dating site, raising doubts about thereliability of the inferred geographic location. For this purpose,we separate the data into the SOURCE (i.e., un-proxied users)and PROXY (i.e., proxied users) subsets, of 1,666 and 3,528profiles respectively. It is possible that IP addresses fromthe SOURCE dataset are in fact unknown proxies, perhapsshared secretly amongst criminals, and similarly, it is possiblethat PROXY users are only masking their specific connectioninformation rather than their national origins. We address thesepossibilities below as they touch upon our results.Some important limitations of the data source mustbe acknowledged as context for our analysis. Firstly, thescamdigger.com site is primarily a scam-list for profiles submitted to a particular dating site, datingnmore.com, whichreviews submitted profiles with particular focus on onlinedating fraud, and lists those identified as scammers either atregistration or after interaction with members. The profilespresented are thus those of scammers that attempt to targetthis particular dating site, which may be a source of unknownbias. As with almost all criminal data analysis, these are alsothose dating fraud profiles from scammers who have beenidentified or caught, and it is possible that they are not representative of a more skillful subpopulation, which could alsobe geographically biased. The former issue could be exploredfurther through comparison with statistics from other datingsites, where they can be persuaded to release this information.The latter is an inherent limitation of criminological data.III. G EOGRAPHIC O RIGINS OF DATING F RAUDTable I lists the significant origin countries for the SOURCEdataset. The largest single origin by far was Nigeria, atover 30% of the dataset. West Africa in general accountsfor over 50% of the SOURCE locations. These proportionsclosely match previous observations of the national originsof advance-fee fraud, as determined by email header IPaddresses [4], [5], suggesting potential commonality between1 https://wiki.openstreetmap.org/wiki/Nominatim2 http://freegeoip.net(September 2017)these types of fraud. The next largest origins, Malaysia andSouth Africa, are also well-known for producing other formsof internet fraud. All of the listed nations score below 50on the 2016 Corruption Perception Index [6], except for theUnited States and the United Kingdom, suggesting these maybe unusual siaSouth AfricaUnited KingdomUnited iaIvory .0250.0250.0180.0170.0150.0140.014TABLE I: The SOURCE countries for 20 scam profilesFigure 1 plots the major scam origins against their profile’spresented location, as directional arrows weighted by volumeof scams. The United States is the location most commonlypresented in dating profiles, at 63% of the SOURCE dataset,followed by the UK (11%), Germany (3%) and Canada (2%).As presented locations are usually indicative of the victims’nationality, we can understand the data as reporting thatresidents of the US are the major target of romance scams,followed by those of other western nations.Africa: Most African sources focus their attention on themajor western targets reported above. A notable exceptionis a cluster of profiles from Ghana which appear to reporttheir location accurately. This may be a simple reaction to ascam-detection methodology which uses mismatches betweenpresented and IP-geolocated locations3 ; or could represent amore ‘honest’ scam format aimed at extracting funds throughstraight seduction. A similar but smaller group appears inSouth Africa. Other exceptions include a small cluster of profiles from South Africa and Ghana which present their locationas Iraq and Afghanistan. These are classic “military scam”profiles, purporting to be members of the US military stationedoverseas. A small number of Nigerian profiles present theirlocation as Malaysia, for unclear reasons.Europe: Almost all SOURCE profiles from the United Kingdom presented themselves as from the United States, withonly 9% targeting the United Kingdom itself, despite this alsobeing an internationally targeted location. Profiles originatingin Turkey targeted the United States and Germany, in keepingwith the international norm. Most interestingly, profiles fromthe Ukraine and Russia almost always presented their nationallocation as consistent with their IP address. This marked deviation from the pattern of romance scams originating elsewherehighlights the distinctive nature of Russian and Ukrainiandating fraud.(March 2017)3 Sucha method is in use by the dating site operators

3Fig. 1: The major paths from SOURCE IP addresses to the locations given in profilesAsia: India follows the international norm in presenting profiles as from the United States and United Kingdom, althoughthe ratio allocated to each is weighted more in favour of theUnited Kingdom (2:1 vs the 10:1 in West Africa), perhapsdue to closer national ties. There are some small groupsof Indian source IPs which present profiles in Singapore orMalaysia. Malaysian scammers also present profiles in the USand UK at the Indian 2:1 ratio, with small secondary clusterspresenting from Malaysia and nearby Australia. Scammers inthe Philippines split their presentation between the Philippinesitself and the US, an unusual pattern that likely reflects theclose links between the US and the Philippines.United States: Almost all SOURCE profiles from the UnitedStates gave their location as within the United States. However,the most common presented state locations were New Yorkand Texas, while the source addresses were mostly located inArizona, California and Virginia, suggesting a degree of location misrepresentation within the nation or else imprecisionof unknown proxying attempts.IV. AUGMENTING G EOLOCATION E VIDENCEAs previously highlighted, SOURCE IP addresses are notnecessarily accurate origins – they could be unknown proxieswhich escaped detection. While this is inherently an unknownfactor, we can make use of certain additional evidence as anaugmentation. For SOURCE IP information we can assess thelikelihood of impersonations, and for the unknown PROXYsubset’s true locations we can examine the reuse of text andimages with direct connections.A. Probabilistic AssessmentWe can first estimate the likelihood of this possibility bycomparing the ratio of SOURCE and PROXY IPs for nationallocations. It is known that proxy lists will have a certain degreeof error or incompleteness, which, under a base assumptionthat knowledge of proxies is affected similarly despite theirlocation around the globe, means we are searching for anunknown threshold at which to discard the idea that certainorigins are genuine – the rate of false negative error in theseproxy lists. As we cannot be certain of this rate, no hardconclusions can be drawn from proxy ratios alone, but wecan say that a large SOURCE:PROXY ratio is a signal carryingsome information about the credibility of location information.Where the number of profiles with an unknown IP addressis a small fraction of the number of known proxies for thislocation, we will regard these locations as suspect. Where thisis not the case, we can be more confident that the IP addressaccurately reflects the origin of the scam profile.NationUnited StatesUnited ry CoastMalaysiaSouth AfricaNigeriaSenegalTogoGhanaP ROXY S OURCE :P LE II: Ratio of suspected source IPs to known proxies bycountryTable II presents this ratio for the major SOURCE countries.From this, we can say that we have the most reason to besuspicious of the validity of IP addresses situated in the UnitedStates, with the observed count of scam IP addresses notknown to be proxies being a very small fraction of those fromknown proxies. We also know that the majority of the SOURCE

4dataset from outside the US have presented their locationas being in the US, attesting international effort at exactlythis form of misinformation. Looking at temporal reportinginformation, we find that the proportion of SOURCE profiles inthe US has been decreasing since 2013, suggestive of graduallyimproving proxy detection.The UK is the next most suspect IP location, also attractinga large volume of SOURCE profiles as a falsely presentedlocation, and with more PROXY than SOURCE IP addresses.However, scammers would have to be an order of magnitudemore effective at masking their IP addresses as UK locationsthan as US locations, in order to explain the ratios of scamprofiles generated by these IP addresses. It is notable thatboth SOURCE and PROXY profiles from UK IP addresses mostoften present themselves as located in the US. This suggestseither that the UK supports a population of relatively securityconscious romance scammers targeting the US, or is acting as asignificant staging ground for fraud from elsewhere directed atthe US. Temporal information here also suggests a downwardtrend since a spike in 2014.Russia and the Ukraine are also locations with a significantnumber of PROXY profiles, but here there is less reason tosuspect the SOURCE IP addresses do not reflect the nationalorigin of the scam. Unlike the US and UK, we do not see anysignificant number of other SOURCE profiles presenting Russiaand the Ukraine as their location, and unlike the SOURCEprofiles, most PROXY profiles from these locations presenttheir location as the US. The reporting figures appear stableover the observed period. The few presented Russian andUkrainian PROXY profiles may simply be scammers protectingtheir individual location and connection information, withoutinterest in masking their national origins. Similarly, knownproxies account for just over a quarter of the IP addressesfrom the Philippines, but there are few profiles traced fromoutside the country which purport to be located there, so thereis little reason to suspect large-scale misrepresentation.The remaining locations are only lightly populated by IPaddresses from known proxies, and we may have confidencethat these are genuine national origins of online dating fraud.Some locations show up neither as significant SOURCEorigins nor as presented locations in profiles, but only astransit points in the PROXY dataset. These are locations withsignificant proxy populations, but apparently of low appealas targets for international dating fraud. All such profilespredominantly presented as located in the United States, withthe proxy country being at best a distant second. Notabletransit locations include the Netherlands, Switzerland, Sweden,France, Australia, Romania and Finland.B. Profile Description ReusePrevious work has shown that romance scammers engagein substantial reuse of certain profile elements to save onlabour, using certain cached images and making use of textual “scripts” which can be copied and pasted with minimalediting [2]. We here seek to explore how these sharingpatterns appear geographically. Understanding which sourcesare sharing resources can help identify cooperating criminalsand similar scam types. Geographic clusters of resources canalso be useful in identifying the true origins of profiles usingproxies to hide their location.Text reuse is common in scam profiles, with key chunks oftext and expressions being observed across different uniqueprofiles. To identify these overlaps, we first preprocessed thetextual descriptions to standardise case and remove punctuation, and then used a longest common substring method tocluster texts. Any two texts which shared more than a thresholdof 10 tokens (words) were considered to be part of the samecluster. By this method, 899 unique profiles could be assignedto a cluster, sharing text with at least one other profile4LocationNigeriaGhanaMalaysiaItalySouth AfricaIndiaUnited 5544415TABLE III: Inferred true locations of PROXY profilesLooking first of all at reuse within the SOURCE subset,the greatest text reuse occurred within nations, with multipleunique profiles originating in Nigeria and South Africa sharingdescription text. The greatest international text reuse wasbetween Nigeria and South Africa, with multiple profiles ineach country sharing elements, and, interestingly, betweenNigeria and the United States. Given the previous evidencethat the SOURCE profiles in the United States may have beencreated through undetected proxies, we can take these Nigerianand South African scripts appearing in the US as furtherevidence of this under-detection. Similarly, scripts appearing inthe United Kingdom suggest that there are undetected proxiesamongst the SOURCE IP addresses from the UK. Text reusewithin Africa and between Nigeria and to a lesser extentwith all of Malaysia, India and Turkey, suggest a commonapproach to romance scamming in these nations. Notably, wesee little to no direct text reuse from Russia, the Ukraineor the Philippines, either internally or externally, though it isworth noting that we have relatively few examples from thesecountries in comparison to the numbers from West Africa.Turning to the PROXY dataset, we find that 241 (11%) sharetext with SOURCE profiles, meaning that their true location canbe indirectly inferred. Table III reveals the results of assigningthe majority national label for shared clusters. As well asadding significantly to the totals for the already-dominantWest African and Malaysian scam origins, this inference alsoreveals a number of Italian scam profiles. Combining thesediscovered origins with the smaller number of Italian SOURCEprofiles which enabled this inference, Italy would place 11thin Table I, with more profiles originating here than in Russiaor the Ukraine.4 This number does not count variants of the same profile identified as suchfrom the dataset, so these 899 reflect 28% of the dataset

5(a) 2d.a8.(b) d4.9e.(c) 6c.89.(d) 15.bf.Fig. 2: Images reused by scammers in different profiles. Eachsub-caption shows an excerpt of the hash of the image. Notethat although certain images are perceptually equal, theirhashes are different.nations in West Africa. These images might be copied fromother scammers, or profiles in our dataset could have beencreated by the same scammer under an unresolved alias.Turning to the PROXY subset, 19 image clusters in this datawere connected to the SOURCE subset, allowing a total of 48proxied profiles (1% of the subset) to be connected to profilesfrom unproxied connections. The major connected locationswere Nigeria (22), Ghana (13), Togo (5) and the UK (4),with the majority of the PROXY profiles affected presentinga US location. Again, this is congruent with other evidenceof a largely West African scammer population making use ofproxied connections to present themselves as US citizens, withsome hints of scammers also acting from within the UK.C. Profile Image ReuseThe use of images plays an important role in onlinedating sites. Scammers often reuse profile images thathave been shown to attract vulnerable users in other locations. The military, the academic and the medical context are recurrently exploited [1]. Figure 2 shows four examples of image reuse. These images appear in differentscammers’ profiles. While some images are perceptuallythe same picture, their hashes are totally different. Thisis the case, for instance, of Figure 2a and 2b, wheretheir hashes are 2da1883450f2b74357465d3031cfd2a8 andd43c4519edc110c6a53dd10e40414e9e respectively.In our work, we use perceptual hashing to fingerprintimages. This type of hashing extracts features from the imagesso that two images will have the same perceptual hash whenfeatures are similar. These hash functions can distinguishbetween dissimilar images, while being robust against differenttransformations and “attacks” [7]. For the purpose of thispaper, we leveraged different perceptual hashing algorithms including the classic perceptual hash function—computed fromthe Discrete Cosine Transform (DCT) between the differentfrequency domains of the image—and wavelet hashing—using the Discrete Wavelet Transformation (DWT) [8]. Perceptual hashes within the dataset are compared in a pairwisemanner using their Hamming distance, and then tested forequivalence based on a distance threshold, and manuallyverified to exclude false positives. We observe a total of 187images which are perceptually equivalent, with some beingreused across up to five different scam profiles.Image clusters were then aggregated from perceptuallyequivalent images which were connected by being presentedon the same profile page (our assumption being that theseare attempts to portray the same subject, even if perceptuallydissimilar in setting). There were a total of 183 profilesconnected by 57 image clusters. Within the SOURCE subset,there were 45 profiles connected by 27 clusters of images.Examining reuse within the SOURCE subset, images werepredominantly shared between profiles created within Nigeria(14 internal connections to 4 external), Ghana (12 to 2), theUK (5 internal) and South Africa (5 internal). The externalconnections from Nigeria and Ghana were to Ghana, Nigeria,Benin, Kenya and Turkey. Though the numbers here aresmall, they fit with the more substantial body of text evidenceshowing resource sharing largely appearing to happen withinV. C HARACTERISING G EOGRAPHICAL D IFFERENCES INS CAM P ROFILESA previous section has explored how the presented location in a scam profile can differ according to the actuallocation of its creator. Other profile elements may also varygeographically, according to the particular flavours of romancescam being employed in each location. In the section below,we examine how demographic characteristics are distributedaccording to the origin of scams from the SOURCE dataset.We survey the demographic information—age, gender, occupation, ethnicity and marital status—for each major scamorigin country5 in the SOURCE dataset. Z-tests were performedfor the age, gender, and topmost category of occupation, ethnicity and marital status, compared to the SOURCE populationaverages. Table IV presents the results, with statistically significant differences (α 0.05) highlighted in bold. Bonferronicorrection was applied to adjust for multiple comparisons.Gender is presented as the proportion of males.An immediate division can be drawn between countrieswhich predominantly present male profiles (e.g., Nigeria,Malaysia, South Africa) and those which present mostlyfemale profiles (e.g., the Philippines, Ukraine, Senegal). Theage of scam profiles corresponds with their gender, withfemale scam profiles typically averaging around the age of30, and male profiles averaging towards 50. The rates bywhich profiles declare themselves single also appear to begender-biased, with female profiles being far less likely touse alternative statuses such as divorced or widowed. Thesewould seem to correspond to very different top-level strategiesof online dating fraud being pursued in different countries,with, presumably, different targets in mind.Within nations presenting mostly male profiles, the strategies appear to be fairly similar. They all mostly report whiteethnicities, and most frequently use military or engineeringoccupations. Two exceptions are India, where the all-malescam profiles mostly present themselves as ‘businessmen’, andItaly, where the profiles most commonly report professionsin the real estate sector. Marital status provides the mostinteresting distinctions. It is clear that a heavy use of the‘widow’ backstory is especially favoured by South Africanand Turkish scammers, also most evident in the profiles with5 Those in Table I, plus Italy, which is promoted to importance whenconsidering text reuse evidence

6Nation (S OURCE IP)NigeriaGhanaMalaysiaSouth eRussiaIvory 07-6.11-5.72-2.44-2.72-1.09-Genderx̄z 0.340.260.320.370.570.500.220.430.300.320.630.17z .390.380.481.000.960.480.550.890.63z 4.243.518.02-0.822.55-Marital nglesingle0.47 -1.730.63 3.700.46 -1.170.57 7.470.33 5.510.30 0.180.58 4.660.53 0.390.68 2.340.88 4.810.97 5.140.89 4.260.79 2.940.65 1.480.64 1.300.89 3.590.50-TABLE IV: Dominant demographic characteristics by origin country. Significant differences highlighted.the (suspect) location in the USA. The UK features profilesunusually willing to make use of a ‘divorced’ status. Theseexceptions mark certain nations as following variant patternsfrom the “Nigerian” approach. The commonalities betweencountries could be attributed to larger-scale campaigns oflocation disguise of the same individuals, or an internationalcriminal group or community following similar patterns ofactivity, perhaps actively sharing tactics.Nations presenting mostly female profiles have moremarkedly distinct strategies. Profiles from Senegal and theIvory Coast are most likely to state a black ethnicity, and toreport being students. Profiles from Russia and the Ukraine arealmost universally white, but may be distinguishable by theirdeclared occupation, with accounting professions being particularly distinctive of Russian profiles. Profiles from Togo arenotable for including a number of females reporting militaryoccupation, whilst the Philippines often present distinctivelyas mixed-race, working in sales positions.Ghana presents an unusual picture, with a distinctive biastowards a more mixed-gender approach to scamming. Furtherexamination reveals that Ghanaian profiles represent a mixof two competing local approaches: male profiles followingthe preferred pattern from nearby Nigeria, with dominantlymilitary occupations, whilst female profiles borrow from atradition more akin to that in Senegal and the Ivory Coast,presenting mostly as students – though, interestingly, Ghanaianfemale profiles are still more likely to be white than black.Kenya also appears to represent a balanced gender mix ofscam profiles, but the comparatively small number of profilesfrom there make it difficult to be confident about this pattern.The language in profile descriptions also shows some regional characteristics. To analyse the variety of topic categoriesfound in the profile descriptions, we used dictionary terms thatare mapped to categories from the LIWC 2015 dictionary [9].Normalised category frequencies were recorded for each profile description and grouped by country of origin.As can be seen in Figure 3 our results showed that scammersbased in Ru

The data used in this paper comes from a public online dating scamlist maintained at scamdigger.com, which offers up romance scammer profile data for public awareness. An exhaustive collection of the 5,402 scam profile instances, as

Related Documents:

May 02, 2018 · D. Program Evaluation ͟The organization has provided a description of the framework for how each program will be evaluated. The framework should include all the elements below: ͟The evaluation methods are cost-effective for the organization ͟Quantitative and qualitative data is being collected (at Basics tier, data collection must have begun)

Silat is a combative art of self-defense and survival rooted from Matay archipelago. It was traced at thé early of Langkasuka Kingdom (2nd century CE) till thé reign of Melaka (Malaysia) Sultanate era (13th century). Silat has now evolved to become part of social culture and tradition with thé appearance of a fine physical and spiritual .

On an exceptional basis, Member States may request UNESCO to provide thé candidates with access to thé platform so they can complète thé form by themselves. Thèse requests must be addressed to esd rize unesco. or by 15 A ril 2021 UNESCO will provide thé nomineewith accessto thé platform via their émail address.

̶The leading indicator of employee engagement is based on the quality of the relationship between employee and supervisor Empower your managers! ̶Help them understand the impact on the organization ̶Share important changes, plan options, tasks, and deadlines ̶Provide key messages and talking points ̶Prepare them to answer employee questions

Dr. Sunita Bharatwal** Dr. Pawan Garga*** Abstract Customer satisfaction is derived from thè functionalities and values, a product or Service can provide. The current study aims to segregate thè dimensions of ordine Service quality and gather insights on its impact on web shopping. The trends of purchases have

To understand how online dating fundamentally differs from conventional offline dating and the circumstances under which online dating promotes better romantic outcomes than conventional offline dating, we consider the three major services online dating sites offer: access, communication, and matching. Access refers to users' exposure to and

Chính Văn.- Còn đức Thế tôn thì tuệ giác cực kỳ trong sạch 8: hiện hành bất nhị 9, đạt đến vô tướng 10, đứng vào chỗ đứng của các đức Thế tôn 11, thể hiện tính bình đẳng của các Ngài, đến chỗ không còn chướng ngại 12, giáo pháp không thể khuynh đảo, tâm thức không bị cản trở, cái được

online dating (and the design of online dating services), online communities and social networks more generally. Self-Presentation in Online Dating As noted in the introduction, one of the advantages of online dating sites is that they allow members to construct a caref