Social Comparison
Wood, Joanne V (1989). Theory and Research Concerning Social Comparisons of Personal Attributes. Psychological Bulletin Vol. 106, No. 2.
Suls, Jerry, & Martin, Rene, & Wheeler, Ladd (2002). Social Comparison: Why, With Whom, and With What Effect? Current Directions in Psychological Science.
On Suls, et al. and downward/upward comparisons
This is yet another interesting article, seemingly based in pscyhology and having nothing (explicitly) to do with technology or ecommunities. However, the mostly intuitive information gains interesting traction when applied to social comparisons within an ecommunity.
Specifically, the authors spend a significant amount of time discussing the much-accepted downward comparison model, which finds that threatened people are more likely to compare others who are worse off, and that exposure to less fortunate others boosts subjective well being of the onlooker (161). Their emotional example is that of cancer patients who may base their personal outlook on comparing themselves to others. However, as the authors elaborate, this doesn't necessarily mean that upward comparison prodcues negative outlook. The authors even provide that "upward comparison with superior role models can prodive hope and inspiration" (161).
Again, this becomes increasingly interesting when applied to ecommunities. In terms of usage and sticking around, downward comparisons don't seem to work at all. I don't believe I have ever continued using an ecommunity because others weren't using it correctly or weren't as adept at it was I am. On the contrary, I look to superior users (admins, moderators, long-time users) as inspiration on how to take better advantage of the site. What complicates this, though, is that my level of elitism often does dictate my usage of a site. So even if I'm annoyed by novice posters on--for example--Wikipedia, I may be more likely to visit and post because I feel like I'm better in comparison to them. On one hand, I may be very annoyed by their ignorant posts, but realizing my own ability to surpass them and be "better" may encourage me to continue as a member even if I don't strive to compare myself to those "above" me.
Stratification in online communities
I like what you said about looking to role models vs. trying to be more elite than the new users. I do think there's complex stratification in some online communities that provides for a lot of social comparison. In many communities, there are measures of success through participation (number of edits, number of posts in the forum, number of books swapped) and also through competition (the coolest video submission, or the first person to reach x contributions). Both of these provide ways in which one user can be perceived to be better than another.
There are also different user roles that some websites have (I'm thinking of moderators, or volunteers) that form some stratification. I think that new users are often intimidated by the more experienced, and even eperienced users are often intimidated or perhaps feel inferior to moderators. (It all depends on the culture and structure of the site.)
Some potential design claims to investigate:
On Wood and importance of social environment
This is yet another reading that isn't directly correlated to ecommunities, but this one provides more obvious bridges than the Suls piece. In particular, this author's literature review echoes the point that the interplay and importance of social environment and the individual (who may be comparing himself to others) has--as of 1989--been widely been undervalued and underresearched. (244)
This is an interesting finding in relation to ecommunities as it relates to a plethora of possible design claims that aim to value trait X on a site. An ecommunity is a social environment, and so if designers make it so that successful and desirable users earn badges, diggs, titles, access, etc that actually becomes prominently displayed on the site and/or their profiles, it stands to reason that a particular individual may use this as possible upward comparison. The subtlety is in what kind of user a designer is targeting. For instance, a newcomer may have no desire to attain a lofty title within a community that he does not intend to invest much in. However, if values within a community allow ther use more functionality over time, then perhaps they person may begin comparing himself to others and invest more time in the site.
social comparisons - up / down or it depends
John Blair
As with all studies involving people, the authors find that nearly every result they can discover, will be exhibited by someone at some time. It makes sense that some people when comparing themselves to downward targets will feel better about their own problem. This is of course presuming that the personality / emotional / physical / genetic disposition (to name a few) can all be accurately predicted and tested (alluded to on page 160). Without this, it seems we are to expected to forever aim for the middle of the road so as to increase our chances of staying on it (or motivating new members, while not offending existing ones, or vice versa even).
One context in which this data may apply within ecommunties is in the establishment of profiles, settings and options as part of membership within the community. For instance, by completing the profile information and including a picture this allows others to execute more accurate and more expansive searches of the membership because the information would be available to query and help to promote "pyschological closeness". As the authors state, "people typically expect to be similar to those with whom they feel psychologically close", continuing, the sharing of attributes about oneself in comparison to others can be helpful in producing assimilation (162).
Summary of Main Points from Wood
I've attempted a summary of the main points. It is structure as the three main goals of social comparison (evaluation, improvement, enhancement) and then how people select targets for comparison, and how people compare on dimensions related to the dimension under evaluation. Feel free to correct, clarify, expand, etc.
Self-evaluation
Target Selection Along the Dimension under Evaluation
In rank order studies, people will compare themselves to people who ranked similarly. Exception: when score range is unknown, people like to compare themselves to high/low extremes. Performance implications: comparisons to dissimilar others does not seem to affect performance, but comparisons to slightly better others seems to increase performance.
Comparisons on Related Dimensions
People also consider dimensions related to the one under evaluation when chosing who to compare themselves to. People compare themselves with people who are similar in characteristics related to and predictive of performance for the dimension under consideration (related-attributes similarity).
Sometimes people will also compare themselves on dimensions that are not related to the dimension under consideration. e.g. related-ness doesn't always matter.
Self-improvement
Target Selection Along the Dimension under Evaluation
Little evidence on these. But it seems like people tend to look at either slightly better or extremely better others for self-improvement.
Comparisons on Related Dimensions
People are most likely to use someone for self-improvement comparison when they are similar to them on related dimensions. This helps explain why people don't mind looking at people who are better than them (potentially conflicting with goals of self-enhancement). They feel that the comparison person is very similar to them, so if their comparison person is a high performer, then they think they are a high performer as well, or at least have the potential to be a high performer.
Self-enhancement
Target Selection Along the Dimension under Evaluation
People compare themselves to others who are not as good as they are - downward comparison. On negative dimensions, people rate others as being similar to them. On positive dimensions, people rate themselves as being better than others. People are more likely to want to engage in self-enhancement when the target dimension is self-relevant.
Comparisons on Related Dimensions
May compare with others who are similar on related dimensions (if they are similar on related dimensions, then I have potential for success, even if I'm not great now). But also may want to compare based on dissimilarities in related dimensions. Then can attritube superior performance to differences in related dimensions (well, that person is a professional). May even try to create dissimilarity.
Erin, you rock. Thanks for
Erin, you rock. Thanks for this amazingly helpful rundown.
Self, self and self
John Blair
Erin - very nice summary. Reading your post helped me solidify my thoughts on the Wood article. My take away is that throughout the multitude of studies cited, a major theme is that people will do nearly anything to make themselves appear better than others. They intentionally select targets that make them look good, seek dimensions of similarity, select comparative scores higher than their own (thus intentionally ignoring those with lower scores, because they assume they already are better than them), they ignore the comparison entirely, manufacture or manipulate a comparison their ego can tolerate, or avoid it entirely.
Whether this interest in appearing better is driven from self evaluation, improvement or enhancement doesn't seem to really matter as the end result for ones self is often the same, How do I compare to those around me? Surrounding dimensions, emotional status, environment, age, education, mood............ the list of potential influencing dimensions / factors goes on and on.
It can be a cruel world and without some sort of positive mechanism of self, be it the three Wood cites or one of simple self preservation, if you don't feel good about yourself, your odds of the world having you for breakfast are greatly increased.
ethical concerns
I have some ethical concerns about the observation that "people will do nearly anything to make themselves appear better than others". When we design a system, we can please the users by making them "appear better than others". However, is it ethical to do so, considering it might enhance users egoism, vanity, and complacency?
You've described only self-enhancement
You've captured the effect of one of the motives. But self-evaluation
and self-improvement motives do not just lead people to do anything to
make themselves appear better than others. Self-evaluation motive makes
them seek information about where they really stand. And
self-improvement motive makes them seek a comparison that will help
them improve their performance (inspiration, or role model, for
example).
Summary of Suls, et al by terms
Attribution Theory - Individuals with similar attributes are useful for generating base levels of satisfactory competence. Individuals who compare themselves to dissimilar others won't be able to determine if the difference is caused by experience/effort or innate potential.
Proxy Model - An individual can extend similar performance and maximum effort shared by another on a past task that has also completed a new task to accurately predict his own chance of success on the new task. If the effort is known, related attributes are disregarded, but if the effort is unknown, then related attributes are factored in.
Triadic Model - Predicts the chances that one person can influence another based on three variables: current preferences, future preferences, and beliefs. People with similar attributes have the potential to be the most influential. A person must perceive some similarity to an expert to accept their beliefs have personal relevance. Satisfaction with or adherence to future preferences between two people is more likely if they have agreed on preferences in the past.
Self-Enhancement Motive - People at a contextual disadvantage are more likely to identify with people at a further disadvantage to boost their subjective sense of well-being. The authors note potential problems with this hypothesis. 1) No baseline comparison, 2) people may not perceive downward-superiority or upward-inferiority in context, but may rather identify with the target other as upwardly-superior (inflated self-perception of ability or inspiration) or downwardly-inferior (conditions may change for the worse). In other words, conditions can produce affective conditions in both directions:
Inferior - status could decline (negative) or status is advantaged (positive)
Superior - status could improve (positive) or status is disadvantaged (negative)
Self-Improvement Motive - The authors extend the aforementioned argument into a new hypothesis: "upward comparison with superior role models can provide hope and inspiration."
Self-Evaluation Motive - Individuals with similar attributes in a given performance-based task, preference, or belief can make accurate evaluations about themselves by observing the other.
Assimilation/Contrast - If someone wants to tackle this section, I'd appreciate your help. The authors' definition is awkward, poorly written, and divided by subclauses and citations. My takeaway point was that salient similarities or differences between a person and a comparison target will be more likely to result in desired psychological closeness or distance.
Distinctness/Mutability - Comparison target has hard or soft boundaries, respectively. Assimilation is likely when the target's self-views are perceived to be vague and contrast is more likely when the target's views are fixed.
right back at you, jon!
right back at you, jon! compliment returned - nice summary post
Assimilation/Contrast
Assimilative upward comparison-- you feel good about yourself from an upward comparison.
Contrastive upward comparison-- you feel bad about yourself from a downward comparison.
Assimilative downward comparison-- you feel bad about yourself from a downward comparison.
Contrastive downward comparison-- you feel good about yourself from a downward comparison.
Early
research suggested that all comparisons were contrastive. Later,
researchers found that they were sometimes assimiliative. The section
on assimilation and contrast covers what factors are thought to lead to
assimilation vs. contrast, and why.
I think the possibility of
assimilative upward comparison is especially important for online
communities. What presentations of high performers in the community
will make people treat those high performers as role models vs.
presentations that where the high performers just make you discouraged?
The factors, "belief that one could obtain the same status" and "having
related attributes" seem like they might map well to the typical online
community setting.
The ultimate drive: who are you? where are you? and how are you?
People are eager to know all about themselves and the world they are living in. That is the ultimate drive for people to seek those information about who are you? where are you? and how are you? This article enhanced my feeling on people's primary motivation of information uncertainty reduction. And this is what a complex world, an efficient way is to measure ones' position and performance in a relative approach. While similar ones, either a bit superior or a bit inferior, provide good measure to position oneself, as well as with a psychological comfort. Similar ones, on a set of relevant dimensions, have those comparable attributes of the comparator, thus making her reasonable to compare with: because these targets provide the relative criteria within similar constraints of the comparator.
To measure, evaluate, and adjust oneself in such a complex multi-dimensional world, getting a small set of dimensions and a set of comparable targets (this provides particular values on the dimensions) is the only possible way to reducing uncertainty. I feel like this rationale can largely explain all have been found in these studies discussed by this article.
Also, I had a bit comment:
on page 236, the author agues that people some time use unrelated dimensions to compare. I feel the example could not really convince me. It says the workers compare only on the others taking the same job but also tend to compare with those of the same sex. I think sex and wage is strongly related although the job itself may seem to be nothing about gender. Because people perceive the gender difference in wage all over the time. Thus on wage, people would distinguish gender because people of the same sex is comparable.
So I thought people only compare relevant dimensions in ONE comparison. But they would draw from other irrelevant dimensions in sequential comparisons to soften the frustration, which was talked in later part of the article.
you've captured the self-evaluation motive
This was the original formulation of Festinger. And it's real. But
there appear to other drives as well, to self-improvement and
self-enhancement.
other evidence on use of irrelevant dimensions
I agree with your analysis of the wage experiment (since there is
ample data, especially going back to the 80s, women were paid less for
similar work.)
However, the article cites several other studies
where people used irrelevant attributes in comparisons. For example, in
a study evalauting logical reasoning, students chose to compare
themselves with others who were similar in physical
attractiveness.
On Suls et al 2002
This article offers a summary of the research on the subject of social comparison, the act of comparing one's self to others for the purposes of self-evaluation. In this way, we use others similar to ourselves as proxies, and their past and current opinions and behavior to interpret and predict our own inclinations and abilities.
Research has shown that in comparing our abilities with others, both the effort and attributes of our social proxies are important considerations, and in the absence of our knowledge of one, we tend to base our comparisons solely on the other.
The authors suggest that there are three basic types of opinions for which we use social comparison to evaluation:
With beliefs, we tend to choose proxies who are share similar attributes to ourselves, such as background, religion, politics or general world views. In fact, although we often base our beliefs on those of experts, our trust in expert opinion can be overriden when an expert does not have these attributes in common. For this reason, researchers have identified the role of similar expert as a highly important one in trend-setting and opinion leadership.With preferences, it seems that similar past behavior is more important than attributes in proxy selection. This is the case with the collaborative filtering mechanism on Amazon.com: users are given recommendations based the buying patterns of users with similar behavior patterns.Basically, people compare themselves with others for three reasons.
All three of these motives can be seen as regulatory behavior. Self-evaluation tells us when it is time to stop working (e.g. when performance is good enough) self-enhancement helps restore homeostasis in mood and self-improvement may increase the drive to achieve a desired state.
Regardless of the motivation for social comparison, social comparisons necessarily influence both beliefs about performance and affective response towards performance.
Social comparisons can either be upward (against a superior target) or downward (against an inferior target). Although intuitively it seems like upward comparisons should make people feel worse and downward comparisons this is not necessarily the case. The conclusions people reach depend on whether they contrast themselves with the target or assimilate the target. Assimilation effects occur when people conclude that the target is in some way a part of the current or future self. Contrast effects occur when the target is excluded from a representation of the current or future self. Thus, upward assimilation effects and downward contrast effects produce substantially the same result.
excellent summary!
(You reversed the order of self-enhancement and self-improvement from the article's treatment, which
made me think at first that you had the definitions wrong, but then I saw you had just changed the order.)
Comparisons and Environment
My main takeaway point (and somewhat of a concern) from the Wood article wasn't about rationality, bias, or comparisons, but freedom of choice, or lack thereof. How much control do we have over how we reason or who we compare ourselves to? Are we slaves to our emotions - especially those that benefit self-enhancement over self-improvement?
The most hopeful variable would appear to be our environment. To an extent, we can control (or at least limit) the people we spend recreational time with or observe through media. Less control exists for interactions facilitated by institutions like school and work. More unsettlingly, I suppose little to no control exists for when and how individuals make comparisons to themselves. As Freud noted, the child is the father of the man - who we are today isn't who we were yesterday. Though behavioral pattern consistency is more likely across time for individuals than across populations, do comparisons with the self become less valid as time increases between the present and the target past-self?
Using Social Comparison to Establish Norms
When trying to figure out how to apply these readings to online communities, one thought I had was in how social comparison can be useful in socializing newcomers and establishing norms in a community.
This is similar to Chris's earlier point - of how people commonly might look to role models or the 'elite' of the community to understand how to use the site. This is good for establishing norms, in that you want your new users to compare themselves to these people because hopefully it will inspire them to work toward similar behavior and status in the community.
As the article said, people are more likely to compare themselves to
someone who is close but "above" them in some way, than close but
"below". Also, "One's self-evalutative comparisons may ultimately lead to self-improvement." These upward comparisons help people learn from others who are more skilled or accomplished in some way. So in an online community, by doing this comparison, it also helps newcomers understand what these people have done to become successful, how they use the site, etc., so that they can figure out how to do the same.
If I was to make a design claim out of this, I would say that using social comparison to establish norms is more important on an identity-based community, especially an open-source community or one dealing in specialized knowledge. This is because I see people more likely to use comparison when it is dealing with something important to them, like their intelligence, creativity, or innovation. ("Am I a better Wikipedia editor than them?" "Is my SimCity city better than theirs?" "Do I seem to be better at writing Perl than most people on PerlMonks?") Using social comparison on these sites is useful because it can motivate people to want to improve and learn more, to be as good as the work of others they see. (Or, as the article states, there are also drawbacks to this upward comparison - like becoming hopeless that you could ever be as good as the other people on the site.)
On other communities where the focus is more on being social or on less relevant personal characteristics, I see social comparison as less important. "Is my Library Thing list of books I own 'better' than theirs?" "Is my 43 things list of goals better than theirs?". These types of comparisons, in my opinion, are probably less likely to happen, and if they do, not likely to spur users to take some action, like buying more books to be able to list them on their profile.
What do people think? Do you see this as relevant to your community?
yes, but...
I agree that the motive of self-improvement will be stronger on dimensions that are important to one's self-image (in order to promote self-enhancement). However, just because *you* consider creativity, intelligence, and innovation more important to self-image than other characteristics (having the right goals in life; what else?) doesn't mean that everyone shares your sense of which dimensions are important. And what people consider important might change depending on circumstances.
I don't consider resilience or stoicism to be important attributes of myself right now, but I might start to consider them important attributes were I faced with a life-threatening illness.
Two more things...
Two more things I was thinking about with these articles, since I enjoyed them so much :) -
Will you move to Canada?
“My question is, how can these types of group comparisons affect members on a site or the site itself”
I see this effect happening in Newsvine where users will refer to Digg when they want to dissuade members of the community from contributing content that is more infotainment than journalism.
“Are there other examples from online communities where comparison of preferences is used?”
I see this a lot in the political discussions on Newsvine. The obvious example is Obama vs Clinton and the debate among Democrats. I’m not sure it could be quantified but I often feel that when there is a political debate people try to out do each other with their support for a candidate by making claims about how strong their support is. This takes the form that if a few Obama supports are discussing an article one might say “if Obama doesn’t get the nomination I’m going to vote for Nader” and someone will respond “If Obama doesn’t get the nomination I’m going to move to Canada.”
I have always imagined that Paradise will be a kind of library.
-Jorge Luis Borges
CouchSurfer testimony
I'm a recent CouchSurfer, and to be honest, I see Hospitality Club as infereior :) It acutally makes me feel loyal to CouchSurfing.
Despite the fact that CouchSurfers don't provide couches!
This is an interesting display of loyalty from someone who didn't get any couch offers in Boston for DrupalCon!
commitment to group from demonizing outgroup
Denigrating or demonizing an outgroup is a common tactic to build commitment to the ingroup.
This is one of the tactics that was suggested (or should have been if it wasn't) in the bonds and identity discussion a couple sesisons ago.
Homophily and Design choices
Community members are going to have a natural desire to constantly evaluate their performance in the community. Homophily in social networks leads to people clustering with similar others. Inherently we gravitate towards others who are similar to us. After reading Baeur's paper one could argue that homophily exists because of the need to compare ourselves with similar other.
Community managers can make design choices that influence the dimensions of evaluation. As mentioned in the paper, dimension of evaluation is the attribute that used to compare people. In community there are different attributes that could be measured and published. Digg for example can publish a number of attributes that can influence the direction of the community. Digg used to publish the Top Digger list - which had a ranked list of users whose submissions made to the front page. But in Feb 2007 Digg stopped publishing the top users list. The primary intention of this change was to prevent users from gaming the system. But ironically this prevents users from comparing themselves with either their peers or for the leaders to help them improve their submissions. Conversely if Digg decides to publish a different a attribute like number of friends of each user - then some user's focus will be shifted towards forming friends instead of submitting and digging content.
Trouble with social comparison
Mouly brings up an interesting point about the influence of the dimensions of social comparison within a community. The ways community members can distinguish themselves will affect how they participate within the community. So, when using the concept of social comparison to generate design ideas, it's important to think about how the dimension of comparison fits in with the goals of the community and how to balance the welfare of the community with the individual pursuit of prestige that social comparison may inspire.
Efficacy of Leaderboards
I thought that the real reason Digg took down the leaderboard was because it was a constant reminder to Digg detractors that the site is not, in fact, an egalitarian way to determine what is important to people through "digging." At least in the early days, a small group of people were responsible for the majority of "dugg" stories on Digg – earlier submissions from people outside the inner sanctum were buried. This oligarchy/ruling party flew in the face of the purpose of Digg.
Of course, links to many of these criticisms can be found here: http://en.wikipedia.org/wiki/Digg#Criticism
Coming back to social comparison theory, it seems Wood mentions something that might be applicable here. On page 239, the section on Inspiring versus threatening upward comparison addresses how individuals might respond to different types of upward targets. Upward comparisons are "painful when the superior others are close or similar on surrounding dimensions" (p. 239). Perhaps, these detractors didn't see themselves as very different from the leaders (which probably isn't assuming too much since the whole purpose of Digg was to provide a community site where anyone could submit and be an equal), but yet the mechanisms of the Digg algorithm , coupled with the composition of the community prevented them from self-enhancement.
burying
I think this is a good point Mouly. In a real world setting there is no way to attenuate dimensions by which social comparison can occur. In an online setting it is possible to present information far more selectively. The leader board facilitated social comparison along a dimension that doesn’t serve the goals of the site.
Burying in Digg could be viewed similarly as well. Burying serves to move better comments to the top but it also provides a metric that allows for social comparison. If your comment is unpopular why not try and make it the least popular? I recently came across a website that tracks Digg’s most buried articles, so there is some notoriety to be gained by being buried as well.
I have always imagined that Paradise will be a kind of library.
-Jorge Luis Borges
good connection
This is a good connection to make! It's also shown up in some social network literature (e.g. McPherson et al 2001). I also was thinking about it while reading Suls et al, where they talked about people being more likely to accept a response from an expert who had similar beliefs to them than to accept a response from an expert with different beliefs.
The existence of yet another force encouraging homophily is concerning. If people seek out communities of people who are similar along at least some dimensions for purposes of social comparison, then it's questionable how much diverse debate will happen. Perhaps designers interested in attracting a heterogenous user base can help do so by emphasizing user-to-user similarities (e.g., below each user's avatar or username on a post, a logged-in user might see a statement describing a similaritiy, such as "[user x] also likes [a band you like]." -- you can probably think of better ways to do do this, but that's one thought off the top of my head).
Surrounding Dimensions
I found these readings to be pretty interesting, though I think that the discussion of 'surrounding dimensions' has not been fully explored. Unrelated surrounding dimensions are often used during comparison, according to the authors, and while they justify some of this behavior with the explanation of non-causal correlations, they seem to contend that some of these comparisons are simply irrational. In my experience, very little, if anything, in the behavior of human beings is irrational.
I believe the authors miss a couple key ideas when it comes to this topic. (1) Seemingly unrelated dimensions may often be linked in indirect ways. (2) Rational behavior is relative to a particular view of the world, and human beings have imperfect and varying views of the world. (3) We have limited resources, such as time and attention (hello SI500), and so we implement heuristic comparisons that are successful much of the time in order to get the most bang for our buck.
Oh, and was I the only one that started reading "self-enhancement" as "self-denial"?
Information Stratification etc On Online Community
The main claim in these two papers is that people tend to compare themselves with people who are similar with them when they evaluate themselves. Applying this theory to online community design, it would be better to differentiate information provided to users. For instance, instead of information such as”there are 30000 users like this book” given to users, another information with more specific target such as “there are 300 users in your age like this book” might work better.
The other important issue is “Downward V.s Upward” comparison process, where the first one is good for Self-enhancement and the other one is useful in Self-improvement. When we want to implement this kind of comparison process to online community, such as “top10 user board” or “Last 10 user board”, it is necessary for us to know the purpose and the main user types on the specific site. For instance, I would expect that upward comparison can lift users’ participation in a technical-support community as people would like to compete in their techniques between each other.
comparison, attributes and the effect of visibility
My dad once told me that one of the things he loved racing bikes was that he got such a rush from "whizzing past guys 20 years younger than me". He didn't win the bike races, but simply being able to compare himself positively with people who had beneficial attributes made him feel good about his relative position.
I think this ties into Mouly's point that visibility is essential to comparison, and especially if you want people to work harder out of a competitive nature. Bike races could be done asyncronously, with people starting and stoping at different times and having them officially recorded, but the syncronized nature of it means you can track your relative progress against other agents.
PerlMonks has a very simple, very poignant means of comparison and that is the experience system. And this metric is made visible by the "Other Users" nodelet which is on every page in the site -- whenever a user logs on, their name appears in that block. The order of the users is determined by their experience in the system -- so users have a very basic, concrete way of comparing themselves with others.
--------------------------------------------------------------
oostendo@umich.edu
--------------------------------------------------------------
Social Comparison in Last.fm
Something that both of these readings allude to in a way is the idea of recommender systems. If people are to compare themselves to similar others, it helps to have a way of finding others who are similar.
Last.fm has a feature that shows you "neighbors." These are people who have listened to the same artists and songs that you have in the last week or so. So they're basically similar others in the dimension of musical taste.
I think Last.fm has done a good job of taking advantage of the desire for people to compare themselves to similar others by providing this list of neighbors. I like to check it now and then to see people who have strangely similar musical taste. I guess I don't really judge myself against them, but it is novel to explore the list, and one could argue it keeps people coming back to the community due to its novelty.
is it a comparison process?
When you look at a "strangely similar" neighbor, on what dimension would you compare yourself? Are you enhancing your self-assessment by confirming that others like your music?
If there's no comparison going on, what else are you looking at? Perhaps you're using it as a recommender system to see what else they like that you don't know about yet?
Self-improvement motive
I was interested in the self-improvement motive idea which had really short sections in both articles saying that there is not a lot of research here. (Still?) As designers it would seem quite helpful to think about motivating concepts and what compels people to action, and what action under what circumstances. Marketers have, of course, already developed a science around this.
Mouly mentioned Digg. One of the interesting takeaways from the Futurtech crowdsourcing panel (thanks awesome organizer, Jon Cohen) was hearing the panelists get into some of the details of what is effective or not for motivating. The Reddit founder said, for example, that they had needed to redesign their vote incentives because it was too demotivating for new people when they saw the most elite members scores/ranks because they seemed so high, and thus unattainable. This is the caveat from the Suls article that "exposure to upward targets increased self-evaluations of competence and motivation when individuals believed in the possibility of a change in their status." So, as designers, it seems that we'd want to try to understand these thresholds (when is a status change incentive credible) in order to design effective systems.
Helper + Similar New Comer = Turn Off ? (Wood)
I found the following quote from the Wood piece most relevent to the process of formulating and / or evaluating design claims:
From Brickman & Bulman, 1977:
"helping situations prompt social comparisons that often imply to the recipient that he or she is inferior to the helper.(p.236)"
Wood goes on to mention a couple of studies that indicate that this concept has an even-worse impact on self-esteem if the subject sees themselves as similar to the helper.
Some related questions to design claims:
1. Does establishing an adoptee / or mentor system have the potential to turn off new users who may feel condascended to by being offered a "helper"?
2. Are there ways to orient newcomers while avoiding a hazing process?
Motives
In terms of three different motivations of social comparisons, I am still a bit confused about when and how different motivations may occur. For example, Liz provides one design claim: “Communities with little means for social comparison are better at welcoming newcomers,” which makes sense to me.
However, can I say that communities with a means for social comparisons are better at welcoming (socializing) newcomers because newcomers have self-enhancement motivation? Also, in the same kind of context I wonder if one person may change the motivation from one to the other.
Assimilation in E-Communities (Thoughts on Sul et al)
Factors Impacting Assimilation according to Sul et al (applied to e-communities):
1. degree to which new members believe they can achieve the same status as veteran users
2. degree to which new members feel a sense of closeness to the community
3. degree to which new members feel a sense of identification between their own attributes and those of the existing community members
4. degree to which the new members feel part of a "we"
These factors are all useful to keep in mind when designing features into an e-community's architecture that have the potential to either embrace or marginalize new comers.
Lisa McLaughlin
see comment above on assimilative/contrastive
I think you may have in mind "assimilation" as in "blending into the group". I guess there's some connection, but the article is claiming that the factors above affect whether an upward comparison will make the person feel better about themselves, not necessarily whether they will want to blend into the group.
Not your usual self-improvement read
My takeaway from here is that, motivating members in the context of an online community can be done via social comparison, besides using the goal setting, individuality and self-benefit approach.
But to compare, members first need to evaluate themselves. If evaluating ability, the proxy model (comparing task similarity, degree of effort, and relatedness to proxy) is used to predict performance in a future task. In evaluating current preferences, the “similar other” model is used. In evaluating belief, the “similar expert” model is used. In evaluating future preferences, the collaborative filtering approach is used (like in a recommender system).
Once members know where they stand, they can start to compare with others – either within the group or externally. For positive effect, the downward comparison ("I am less inferior"-approach) is used to get a positive effect, as opposed to the upward comparison which the theory says can produce a negative effect on self. Nevertheless, further studies are inconclusive. Now they say, regardless of how better or worse the person you are comparing with, a more important factor is the similarity of that person to you. People tend to focus on similarities (in relatedness or connection) with people more familiar (which leads to mutability), and on differences with unfamiliar (incongruent) people (which leads to distinctness). What a mess.
Can changing prototype for Social Comparison create more loyalty
Social Comparison and Goal formation can also be an
important part of creating regular “events” which can bring users back
regularly to the site.
For example in PagalGuy there are regular tests which take
place and people post their scores and their experiences with the test (for example
which questions were difficult, which sections were easy, what was the optimal
strategy). It is very common for people
to identify others who are like themselves (similar in scores, ability to do
math, logic etc.) and track their progress as well as their own and evaluate
themselves based on those people’s progress. However, since the person they are
comparing may change quite a bit (they may have improved their score quite a
bit from before, or may have improved their score way slower than the comparer)
the prototype person people compare themselves changes completely and further
the change is hard to predict people have to come back after each test to see
who to compare themselves against. Hence a changing prototype to compare them
against also makes an “event” to which people will regularly return and may
create loyalty amongst the members.
Release of scores for comparison as an event
Interesting design idea
My community (cgsociety.org)
My community (cgsociety.org) has an example of the social comparison
mechanism bound by self evaluation and self improvement: Paying a yearly
subscription fee.
In my community, advanced digital artists (both pro & non pro) go
directly to the paying membership route given the many more amenities that they
are offered. However, the level of critiquing is at a much more advanced
level. This inherently separates novices from advanced artists. A
novice will not become a paying member until he/she is a highly coveted novice
non paying member no longer being challenged and truly wants to improve their
artistic skills and be peer reviewed. Otherwise, that person (novice)
will just remain at the highest level within the non paying sector.
Similarly, advanced artists are consistently comparing their work with each
other and competing against one another (competitions are constantly being held
and sponsored by companies within the digital artistry industry). I believe these
mechanisms, arguably, allow the cgsociety.org community to continue growing
while staying cutting edge.
teasing out the argument
I think you're claiming that one of the things people buy with the paid subscription is the option of perform social comparisons against a different set of people. For an advanced novice, this would be the option of switching from downward comparisons to upward comparisons.
What would the readings suggest about the conditions under which people would want to make that switch? Which motive (self-evaluation, improvement enhancement)? And which dimensions for comparison are involved (other than quality of one's artwork)?
Social Comparison
Suls, et al, write about social comparison.
What is interesting is that they observe people "comparing one’s own performance with the performance of someone possessing similar related attributes." For political philosophy this is significant in that it highlights how we cannot begin to discuss "difference" on any axis without also simultaneously expressing "identity" on another axis.
Again, they say "Comparing one’s performance with the performance of someone with superior or inferior related attributes is not as informative because any performance difference could be attributed to differential standing on the attributes rather than to differences in inherent ability." Interestingly, this is the intuition behind ruling out confounding factors in statistics.
This is also present within "expert" knowledge relations: "The model assigns considerable importance to the role of the similar expert (someone who is similar in some ways but not others)" which could be used to thinking about things like wikipedia.
Of particular interest is the section on "distinctness and mutability" which become increasingly important under current conditions of pluralization and interpenetration.
We should note however that comparisons depend on values. Imagine 2 people: one with high education and low income, and the other with high income and low education. These individuals could perceive a comparison as "downward" from both directions, thus both would feel superior to the other "in the thing that matters" to each, i.e. on different axes. In other words, salience of the comparison is a crucial factor, i.e.does one CARE if the Other is "above" or "below" the Self on any given axis?
Humorous and possibly worth looking into that Rousseau blames all of the subsequent evils of civilization on the first time that human beings began to compare themselves to each other.
Wood:
No sooner had I written the above, than I read the Wood paper, which is in keeping with my assessment and even delves much further into the complexities of social comparisons (cf the investigation of "similarity" on p236).
(p238) The notion of closeness very neatly maps onto the way I have been using closeness in relation to network topology and distance.
Regarding the now, as we have more and more access to others with whom to compare ourselves (via teh interwebs), it makes sense that social comparison theory would get a boost.
--------------------------------------------------------
PHartzog@umich.edu
--------------------------------------------------------
The Universe is made up of stories, not atoms.
--Muriel Rukeyser
Does it achieve the goal? :[Reposting from the blog]
Social comparison seems to be a concept where (in time) good becomes
better but bad becomes worse. Digg is a classic example for that (for
diggers as well as for stories). My question is: Is it a good thing for
a social ecommunity? Imagine how a real-life community (or a group of
friends) would be if it starts behaving like digg? It would lead to the
formation of several layers of "classes" within the community: some
better than the others. Does it attain the objectives of the community?
I have my doubts.