This paper bothers me for a number of reasons. I will briefly summarize McKenna and colleagues' arguments and then address my concerns. Briefly though (so you can determine if this is worth reading), I intend to address the following concerns:

• They only articulate a process model for one of their three hypothesized causes (absence of gating), the other two are speculation.
• They have no evidence that the gatekeeper filters used in real life are bad (or that they don't exist in a different form on IRC)
• The statistics used to analyze their data do not alow them to make the conclusions that they do and in some cases are grossly wrong. 
• There is no evidence that the lonely and socially anxious people they talk about have a shortage of real life friends.
• Many of the relationships they find can be explained without appealing to a "true self online"by reversing the causal direction or thinking about what the questions they ask really mean.

Briefly, the authors argue that the internet fosters more intimacy for three reasons. First,  people can disclose more to others without concern that they will violate the "dyadic boundary" (by disclosing secrets to all their friends). Second, people can "get past the gates" and be judged based on what they have to say rather than other visible features such as appearance, stuttering or anxiety. Finally, it is easier to find others who share similar interests on the internet.

MkKenna et al. argue that these conditions are ideal for people who have difficulty getting past the gates in face to face interactions. Note that at this point, reasons one and three kind of fall out of their model.

To support this claim, the authors conduct three studies.

Study One found that people who are lonely or socially anxious are more likely to locate their "true self" on the Internet and that people who locate their "true self" on the net are more likely to have contacted other people they met through the Internet in some way other than email.

Criticism:

***Stats Nerd Diatribe**** 

the measures of whether the true self is located online seems confounded with whether people end up contacting others they meet on the net. Essentially, the claim is "people who feel like they can disclose to others that they meet on the Internet are likely to  do more than just exchange email with them." SEM cannot test causal assumptions about this link meaning that it is also plausible that you need to engage in some form of conversation other than email in order to self-disclose with someone. Second, is the "location of the true self" is in fact a mediator, than their should be a relationship between loneliness/social anxiety and behaviors such as calling people met on the Internet only until the "true self" index is plugged into the model, at which point these relationships should fall away. This would be the more convincing causal argument but it is not what they report on pg. 17.

****End Diatribe****

Study Two

Study Two a follow up to Study One. They find that many people were as close or closer to their Internet friends two years after the original survey. They also find evidence to support the claim that their "real me" index is basically stable. They also find that people reported being less lonely and socially anxious than two years prior.

Criticism:  It is not clear to me that you can conclude from the change in
loneliness that "using the Internet makes people less lonely". It could
simply be that people reach out to Internet friends at particular
points of life in which they happen to be lonely (say, when they move
to a new city) and that loneliness returns to baseline over time.the statistical test they ues to demonstrate that their "location of true self" index is stable is wrong. 

****Stats Nerd Diatribe****

A t-test will not tell you this, it will only tell you if the population moved clearly in one direction or another.

****End Diatribe**** 

Study Three 

Study Three is a lab anologue to Studies One and Two, and is the only evidence the authors can present of the process involved in the observed relationships. Participants were randomly assigned to three conditions (two meetings face to face; two meetings IRC; and two interactions w/ the belief that they are interacting with two seperate people "trading places". Really these are two versions of the same study - one between subects (condition 2 vs. condition 1) the other within (condition 3; IRC vs. face to face). They found that repeated Internet interactions increased liking more than  face to face interactions. Additionally, quality of the conversation predicted liking when people communicated over IRC but did not in the face to face condition.

Criticism: Of course people rely on the quality of the conversation when they interact over IRC, they are deprived of all other cues. MkKenna et al. assume that these other cues are wildly uninformative and should not be used to make judgments about people. However, maybe they are wrong about this, people are also really good about picking up many non-verbal cues that do matter, such as signals of trustworthiness and interest. Maybe this has nothing to do with the socially anxious guys who get a raw deal. Maybe people like others that they cannot see because they cannot filter out creepy free-riders. Why should we assume that "gating" features are superficial?

**** Warning: Stats Nerd Diatribe ****

First, all of the data in this study are non-independant. Analysis should have been done at the level of the dyad, not the individual. This alone would be enough to drive most of these effects down to the level of insignificance. There are a couple of other weird  results reported in here (e.g. an ANOVA with 190 degrees of freedom).

****End Diatribe****

As a final thought McKenna et al. seem to equate loneliness and social anxiety with the stereotypical shy computer geek who has difficulty making friends in the real world". This feels kind of wrong to me. First, I am not sure what the correlation is between loneliness, social anxiety and difficuly making friends. I could just as easily tell a story that invoked the stereotype of the sorority girl or frat boy who is constantly worried about rejection by others and surrounds themselves with many acquaintences in an attempt to hide their inability to form a substantial relationship with others. Maybe these are the people who make friends online with the purpose of transcending the internet/real-life barrier.
 

Notes 

Internet has special qualities that make it a different medium for relationships such as greater anonymity, lack of gating features, and ease of finding similar others.  The implication for relationship formation is that they will be intimate more quickly and be stable over time.

Critique

The authors claim that such relationships should be able to better
survive a f2f meeting with the lack of gating features. I think that
might be the case for the first meeting, but in time those
relationships should converge to regular ones.

I had a concern about the scale used in Study 1. I am not sure if it is right to measure friendship and romantic relations using the same scale (3=very close friend, 4=romantic partner). Not every pair of people who hook up on the Internet would be romantically involved. But they would still have an intimate relationship. The problem I had with Study 2 is the time interval. I think people would forget the intimate details of their intimate relationships after two years. How do they define intimacy anyway? Plus, it is hardly convincing that 71% of the romantic partnerships remained relatively stable and durable over the 2-year period. For the last study, I found the amount of time (20 minutes) limiting to make meaningful inferences. But maybe that is enough, I’m not sure.

Personally, I don't like to share too much information about myself because I always think that information is traceable back to me. In the past, my email account was compromised. It wasn't really a nice memory. So, I don't trust people easily. I feel like I am sharing a subset of information about myself (mini me), not the real me. Moreover, my mini me is not necessarily the same in every community I am a member of; it changes depending on the context. And that's perfectly fine. In short, people do not necessarily want to disclose more information because it might not be safe for them. For instance, some students might not want to post crazy pictures on Facebook, thinking employers might check on their profile.

Connections

There is a connection to Sunstein’s work on social fragmentation where the authors talk about finding similar others.

Summary
The hypothesis is that close relationships are more likely to form online between people who reveal their "true self" to others, and these online relationships are more likely to result in face-to-face relationships. After performing three studies, the authors claim that 1) meaningful relationships form online; 2) these relationships are stable over time, and; 3) people like one another more when they meet face-to-face after starting their relationship online. They posit that gating feaetures (barriers to relationship formation such as ugliness, shyness, etc.) that are absent online help explain why online relations hit it off so well when they meet face-to-face.

Critique
Jesse pointed out some concerns he had about the statistics used in this study, so I won't go there. He also has a lot more insight into the experimental design, and there appear to be several serious concerns there, too.

I'm not as qualified to address those things, but I was generally just troubled by the premise of the study. Key to their analysis is how they determine what qualities or information defines one's "true self." Unfortunately they don't define those qualities. I'm pretty suspicious of using this label, as it implies that 1) I know what my "true self" is, and 2) I can communicate my "true self" in a CMC environment. Something about that doesn't sit right with me at all. Does the "Real Me" consist only of whatever behaviors and messages I present to the community that give me the best chance of forming a relationship? If I am online actively seeking a meaningful personal relationship, I'll probably tinker with my identity in order to give me the best chance of finding someone. Is that my true self? Or just self-marketing?

Critique 

Overall, McKenna study sounds quite convincing to me (intuitively), and especially the darkened room experiment by George et al. (p24) seems to nicely support McKenna’s study assumptions about the relationships between “Real Me” and “Gates” and the ease of forming relationship online.

However, besides Jesse's  and Matt’s criticisms, I have questions regarding an assumption and reference in the study.

From Study 3, I would like to see whether receiving any kind of a message from someone itself evokes the feeling of liking.  I once heard that if you meet someone while playing sports, there is more chance you fall in love with her/him because your heart rate is much higher than normal, and your brain takes it as the signal of love.   Likewise, one might automatically link getting any kind of a message to the signal of liking or attraction by the sender.  My understanding is when you ask the phone number of someone, you express your attraction to him or her if it is not a work place.  I guess communicating with someone might have broken down the mental guards of the subjects regardless whether they revealed their Real Mes or not.   

The other question I have is about the Kraut et al. study (p29).   Their study shows that people who were more active online had also healthier relationships in their real life.  However, I believe that it will be much more meaningful if a study can show how one person’s social life has changed after she or he uses Internet a lot.   In other words, it might be the case that only 10 people’s social lives in the real world might have improved after extensive use of the Internet, and the remaining 90 people’s social lives might have become worse.  However, overall, all the subjects might have still had time for their families.   I thought about this possibility because as mentioned many times by many people, time is very limited for the most people. I am not quite convinced that the extensive use of Internet does not badly affect an individual's time with his/her family members.

Application

It seems if real names are used as user IDs in an online community, the community cannot have the "darkened room" effect anymore; therefore some people will not be able to reveal their Real Mes.  The use of real names is a powerful measure for the protection of a community; however it seems that designers should be extra cautious to use real names.   

In this interesting article, the authors attempt to show that those who better rexpress their true self over the Internet are more likely than those who dont to form close on-line relationships quickly, and move these relationships to r/l.

Because anonymity reduces risk, it leads to greater personal disclosure, which leads to greater intimacy and closeness. The authors refer to this self disclosure as the "Real Me", based around identity-important yet usually unexpressed aspects of oneself. I guess examples might be an individual who is gay but still in the closet, or someone who never expresses political beliefs out of politeness, even though they are deeply important to her .

They also present the idea that people get to pick the groups and relations they use to define themself. Somehow, these relations are incorporated into our identity (p12), and let us chose where to locate our "real me".

But, i am still a little hazy on what it means to "locate the real you online". It seems like it shoudl mean more than just more revealing more things about your self online. What about time spent online v. offline attempting to make relationships? much of our offline time is spent in other pursuits, and relationships emerge, but online time may be spent explicitly attempting to form relationships.

I agree with Matt that communicating true self seems like a moving target. The authors never directly address the idea of people lying about their identity to foster relationship formation online. The same properties of the net that make it easier for your real self to shine make it easy to present yourself differently to each person you meet.

Holy crap-- gold stars to Jesse for that insightful post!

This paper presents a few main ideas about why the Internet can be a good place for people to form strong relationships. If we take these ideas for face value, ignoring the authors' dubious association of them with people who are lonely and depressed, they do provide some valuable insights to consider:

• The Internet can produce a greater degree of intimacy and closeness because of the anonymity it affords; people do not need to worry about sharing personal information that will be attached to their physical identity
• The Internet enables people to socialize regardless of social stigmas that may limit socialization in face to face interaction
• The Internet brings together groups of people with similar interests who may otherwise never get to meet

The challenge in this all working out is in one's ability to locate and define one's "actual self" either online or offline. The author attempts to show that people who have a well defined "actual self" in the online world do quite well at forming relationships on the Internet. The problem here is clearly that there isn't a good way to gauge and locate this concept of "actual self"-- I may know that the way I present myself online and offline are different, and I may have a sense of which one is more "me" but that does not mean that this one is my "actual self." It could still be lacking quite a bit, even if it's more "me" than my other ways of presenting myself. I'm just having a hard time wrapping my mind around this.

Regardless, I think that what the authors are trying to say is similar to what Kim pointed out in the discussion of profiles-- people feel better able to form connections when they can learn some relevant information about one another, such as shared interests and involvement in a social circle. For some people, sharing this information may simply be easier done on the Internet than in person. Further, the self-realization of what one likes, what one is looking for in connections with other people, and where one is comfortable making these connections makes it easier to actually go out and form these connections and likely form strong relationships. Or, at least, this is my interpretation of the point that McKenna is making on p 13 under "Being the Real Me".

 Application to Geo-Aware Communities
As I mentioned in my discussion of Cummings' article, the communities I am studying rather weakly support relationship formation currently. There's ways for members to leave comments to one another about each other's contributions to the community, but these comments tend to be more like one-way interactions than two-way conversations.

However, these communities do embody the idea that I think McKenna (weakly) tried to point out-- people who are interested in these very unique hobbies may not find many people in real life with which to share this interest. Having communities on the Internet in which they can display their personality and find others with similar personality traits provides a method of relationship formation that may not have occurred in the offline world. The three communities I am studying put the members at the center of interactions with the community-- when viewing a contributed arrow, cache, or map there is a clearly-available link to the contributor's profile, where one can learn about the contributor's history in the community, demographic information, interests that he or she wants to make avaiable, and so on. This enables visitors to build trust in the community by seeing that real people contribute, and enables them to gain a better idea of what the community is about and who comprises the community. This level of clarity opens the possibility of increased trust in a community, which Kim also mentioned in her article as an important factor in relationship formation. Without the centrality of the user profile, these communities would just seem like dynamic databases of interesting locations that may not particularly motivate the GPS lover to look for others like him/her.

Anecdotally, this article makes a lot of sense. Growing up with AOL, I can remember having a lot of conversations with people from random places in the country about issues with which we would never discuss face-to-face. Yet I can't help but wonder about the "Real Me" idea. Have you ever spoken to people in a chat room with an assumed identity, or a falsified your age, background or gender? In such a situation, a person can create entire relationships that are based around "disclosure" of events and feelings that have never happened. Indeed, on a smaller scale it is possible that Internet relationships contain much more of this than real world relationships because of the lack of gating—and therefore lack of validating their truthfulness—as evidenced in Ayaça's cartoon.

I'm struggling to relate this article to DragonRealms, although I believe on some level it must exist. Players, from my observations, usually play for years, and establish relationships in-game which at times become in game husbands or wives and even children. Yet because these communities feature assumed personas, the "Real Me" connection seems weak as well. A person could feel that their "Real Me" is an axe-weilding Dwarf barbarian (which although humorous, said façade might be a comment on themselves, a la the essay I posted to the del.icio.us aggregator), but then the connections that draw people to real-world connections seems more tenuous. How does one feel when the Elven warrior mage with mead-colored hair that you've been mooning over turns out to be a forty-something man?

Richard, if you read this, I believe our communities are similar in this regard, and would be interested to hear your ideas on the subject.

From personal experience, I think that McKenna's true self argument works.  I have witnessed those who (mostly bloggers I know) have expressed their "true self" more openly on the Internet seem to form on-line relationships better than those who don't.  In fact, one my friends who stayed with me in Ann Arbor met up with one such friend here through a progressive rock messaging board.  

 I also think users who believe they are reading the "true self" of another user tend to treat that this person with more credibility and authenticity.  It seems that people are willing to believe someone who is more open because it is seems like they have nothing to hide.  Although I should point out there have been many cases of people making up fraudulent Interent personas to attempt to gain sympathy and presumably forge some sort of relationships with others.  One prominent example is the Kaycee Swenson case.

This issue applies to the Fighting 44s.  When I contacted them about my study, they replied back that several changes were being made to the site.  One of which is that the founders were going to reveal their true identities.  Currently the site founders go by their screennames (ie. Dialectic) and generally refrain from disclosing much, if any personal information.  Perhaps, by revealing more of their true selves the founders will develop better on-line and off-line relationships with its membership.  It remains to be seen.