Tuesday, February 17, 2009

Social Psychology 101 for Community Managers

These are my speaking notes for my session at the Online Community Unconference East 2009. I believe the wiki is publicly available for viewing.

These are totally messy. I am trying to break my desire for unattainable perfection and allow drafts of my thoughts to be out there. If there seems to be interest, I may turn these into a better formatted series.

- I am not a social scientist, just a Community Manager who has been learning as I go.
- I want to share the moments of, "I know that! It has a name?" with you.
- The format is definitions, examples and some practical applications.
- The main goal is to get you started in learning from other disciplines.
- The secondary goal is to improve ourselves as community managers and participants as community members.

(Additional notes: there is a community manager in Poland named Darek Kłeczek who blogs at Leadership in Social Networks. I came across his post, 10 Social Psychology Tips for Managing Online Communities while searching if anyone else had approached this topic yet. He's worth keeping an eye one as he develops ideas about how we can encourage leaders to develop in our communities.)

Fundamental Attribution Error
People tend to presume the actions of others are indicative of the "kind" of person they are rather than their actions being caused by a situation.

Based on experiment by Edward E. Jones and Victor Harris (1967). Coined by Lee Ross (1977). Sometimes called "correspondence bias", but not by all social scientists.

- Community makes error against their own: First time posters who make a social mistake being considered a troll.

- Community applies attribution error to community host: Considering the host uncaring/out of touch when tech changes occur suddenly or when the host makes a decision about Terms of Service (accused of playing favorites).

Reducing the Effect:
The attribution error occurs usually because there is not enough information about the situation. Studies have shown that when victims of crime learn more about the criminal's circumstances views tend to shift from desiring harsh penalties to compensation for losses.

To prevent yourself from committing the fundamental attribution error, gather situational information:
- Do people tend to behave the same way in the same situation?
- What would I do in the same situation?
- *Ask* the person for help in understanding their situation.

To reduce the effect when others are applying the fundamental attribution error to you, disseminate the situational information.
- publicly realign yourself/your org with community goals/values.
- clarify the way the situation is leading to types of behavior.

- ask potential spammers for personal situation before assuming they are evil. (They may be enthusiastic and not realize the norms of your community)

- Explain situation behind any changes to your community before the changes are implemented. (You may also enlist aid from community to float preferable changes.)

INTERMISSION: A note about conflict management/mediation
I will not be covering conflict management or conflict mediation, though many examples involve these skills. I am willing to help anyone with questions about resources. For the wiki, this might be a good place to compile some resources.

Actor-observer bias
We tend to attribute our own behavior to the situation, but the behavior of others to the "kind" of person they are.

Developed by Edward E. Jones/ Richard E Nesbett (1971) as the flip side of Fundamental Attribution Error.

However, Bertram F. Malle questions actor-observer bias because of a lack of evidence (2006).

- "I am not a bad person."

Reducing the Effect:
- Awareness that this may be happening (on your part, on the part of the person judging you, on the part of two community members towards each other).

- Use the same methods as the Fundamental Attribution Error: learn more about the situation.

- This is as much about self-reflection as it is observation of others.
- I once had an argument with a vendor that seemed to be about their customer service attitude when the truth was I (and our organization) were asking more than the vendor could give. Realizing this allowed us to enter future vendor deals with realistic expectations, fewer conflicts and better results.

Confirmation Bias
We tend to look for, or better remember, information and evidence that supports our preconceptions and avoid/overlook/forget evidence that counters our beliefs.

- Good luck charms, people who believe in psychics/cold readers, belief that a computer problem is because of a virus transmitted by your web site

- We may be wired this way. If a successful survival strategy works, it's better to repeat the strategy than experiment. If a prehistoric tribe succeeds in hunting a deer, they are likely to stick with it rather than risk going hungry. But it's a short jump from attributing the successful hunt to a pretty stone instead and thus the good luck charm is born.
- The person facing evidence to their belief may feel shame, stubbornness or hope.
- Other factors that may filter counter-evidence may include tradition, taboos, religion, ideology.

Reducing the Effect:
- View information impartially. Welcome counterarguments. ("Strong opinions, loosely held" - ******** )
- Combating this in yourself: imagine a deamon, similar to Maxwell's Deamon that acts as a gatekeeper to your senses, allowing agreeing facts in and deflecting facts that counter your beliefs. (aka Morton's Deamon)

Social Facilitation
We tend to do simple tasks that we know well better with an audience than alone. But, we tend to do new or complicated tasks worse in the same situation.

- Performance anxiety

Reducing the Effect:
- Encourage practice.
- Provide graduated experiences such as encouraging poll voting or rating before submitting opinions, reviews or other content. Also, allow people to create content and control ever-widening circles of who sees the content.
- Compliment when a new or difficult task is *attempted*.

- Warm welcomes without correction when someone contributes for the first time.

Social Loafing
When work is pooled and individual performance is not known, people in groups tend to put in less effort.

Reducing the Effect:
Reveal individual performance for simple tasks (to avoid problems in Social Facilitation). For complicated tasks, keep performance private until proficient.

- Badges in online game systems. Simple acknowledgment of simple accomplishments and difficult accomplishments. People sometimes complain when a game does not provide badges showing the power of revealing individual performance.

Bystander Effect
Individuals are less likely to offer help in an emergency situation when other people are present.

Specific to online communities, groups don't handle conflict situations when there is a perception that someone else will handled it faster or with more authority.

Originated around Kitty Genovese who was stabbed to death in 1964 and "no one helped". The story is dramatic, but not entirely true. The media failed to contact the police for information before reporting the story and possibly committed confirmation bias in avoiding information undermining the belief in the bystander effect. In fact, the police were contacted at least once during the attack and "bystanders" heard, but could not see the attack.

The Bystander effect was demonstrated in experiments in 1968 by John Darley and Bibb Latane.

In the late 80's the a gaming service called The ImaginNation Network was having problems with users harassing each other. There was no formal reporting process so members were a bit on their own with few ways to contact the service. The service added a "report abuse" button and, quickly, the number of reports increased. The members stopped trying to resolve the problems themselves and resorted to the abuse button first.

- Social influence. We tend to look at how others are reacting in situations for cues on what to do. If everyone is waiting to see what others do, no one will act.
- Assumption that others will intervene and feel no special responsibility.
- Fear of being evaluated (Social Facilitation), embarrassment, or being superseded by someone more skilled.
- Uncertain help is wanted.
- In online communities, when authority moves fast or decisively.

Reducing the Effect:
- As a Community Manager, drag your feet a little. Give the space for others to step in and encourage them when they do. (Once I had someone who was great at helping in the community in a specific way, he was so fast and so good, others were commenting on him "beating them to it". I praised him for being a great example and ask him to be a little slower off the mark. When he did, others were able to step in. This took the pressure off him and eventually lead to a strong culture of helping in a specific way.)
- As a Community manager, do not use your admin tools first or often. Resolve conflicts publicly as a way to demonstrate to others how they too can handle them. This woudl be a literal empowerment of your community.
- Instill a sense of responsibility and empower the members by following up when they help or get in over their heads. Allow the community to help you spot spam and deal with items they flag. IF someone attempts to resolve a conflict and has trouble, then step in (and later privately help the person improve their skills).

Other social theories
These were not covered during the session. They are presented briefly to provide some background to common terms a community manager will encounter as they read deeper into social psychology and sociolgy.

Small World Experiment
- Stanley Milgram
- Average path length for social networks
- Dropped letter with a note asking to send to person they know who is likely to know the target on letter.
- Multiple factors could have accounted for the average path length he found
- Milgram never used the phrase "six degrees of separation"
- Work in disease transmission indicates that removing the supernodes of a network has little impact on average path length.

Dunbar Number
- Robin Dunbar
- Theoretical cognitive limit to the number of people humans can maintain a *stable* (grooming) relationship.
- Based on primate grooming habits. Max grooming contacts seems limited to volume of neocortex.
- 1992 - Dunbar extrapolated up to human size brain and derived the number 150
- Then compared this to human groups (tribes, basically)
- Dunbar says communities must have high incentive to remain together (stable relationships). Speculated that humans may spend up to 42% of their time in social grooming.
- For a look at group sizes with numbers from online communities, I highly recommend Christopher Allen's six-part series on group sizes starting with: The Dunbar Number as a Limit to Group Sizes and finishing with his two-part Community by the Numbers.

Fixing Broken Windows
- George Kelling and Catherine Coles
- Fix small problems to dissuade larger problems and, eventually, huge problems
- NYC Experiment:
- NYC Transit Authority 1985 - 1993+
- Guliani's zero tolerance 1993 -
- Major crime *did* go down. But it also went down in cities that did not have zero tolerance policies. (See Confirmation Bias)

Robert Cialdini - Influence
(this is a mash up of ideas from Dr. Cialdini's book "Influence" and articles he has written that I blogged previously at: Community Doesn't Sell and Persuasion Revisited.)

Appeal to majority
- "Many guest waste towels. Please don't." - Little effect on reusing towels.
- "Most guests reuse towels. We thank you." - Increase in reuse of towels.

Influence reciprocation
- When a waiter brings mints with the check, there is a slight increase in tips.

- Learned from NPO event days. If you ask people to sign up, but do not charge, attendance will be significantly lower than expected. If you ask people to pay $20 during sign up, attendance will be much closer to expected.

Conformity - social proof
- monkey see, monkey do.
- Be the alpha monkey
- Teach others to be the alpha monkey.

- In the midst of a credit crisis with people being told to stop using their credit cards, what does VISA do? They create a new "exclusive" card. The VISA Black Card. The exclusiveness will generate desire despite rational.

Relationship awareness (a version of commitment)
- When one member of a couple is trying to convince the other to make a change, those who mention the existing relationship before requesting the change had better success. Think to current Obama phrasing that our problems are "American problems". Reminding us of current relationship as Americans before asking Republicans for change.


  1. Thanks for posting this. Happy we got a chance to catch up at the Unconference too!

  2. Great stuff! Been sharing this around quite a bit.

  3. A useful summay. I would note that in addition to what comes from Social Psychology there are relevant findings that Social Psych leverages from Cognitive Psychology and Decision Science.
    For example, experimental results rom Tversky and Kahneman as far back as 75 indicating several types of cogntive biases which produced poor judgements. They attributed these to one of three main heuristics they formulate as: representativeness (judgments influenced by what is typical so it can be a group effect), availability (judgments based on what come easily to mind, which may be because someone just set an idea in your mind), or anchoring and adjustment (judgments relying on what comes first).

    Gary Berg-Cross