Last week we announced a prize for the best example of an evaluation. The winner of the evaluations prize is David Manheim, for his detailed suggestions on quantitative measures in psychology. I selected this answer because, although IAT was already on my list, David provided novel information about multiple tests that saved me a lot of work in evaluating them. David has had involvement with QURI (which funded this work) in the past and may again in the future, so this feels a little awkward, but ultimately it was the best suggestion so it didn’t feel right to take the prize away from him.
Honorable mentions to Orborde on financial stress tests, which was a very relevant suggestion that I was unfortunately already familiar with, and alexrjl on rock climbing route grades, which I would never have thought of in a million years but has less transferability to the kinds of things we want to evaluate.
How useful was this prize? I think running the contest was more useful than $50 of my time, however it was not as useful as it could have been because the target moved after we announced the contest. I went from writing about evaluations as a whole to specifically evaluations that worked, and I’m sure if I’d asked for examples of that they would have been provided. So possibly I should have waited to refine my question before asking for examples. On the other hand, the project was refined in part by looking at a wide array of examples (generated here and elsewhere), and it might have taken longer to hone in on a specific facet without the contest.
B) Has at least implied that he wants to use EA thinking in the role
My default belief is that a politician implying something he knows the listener wants to hear is not evidence he's believes or will act on that implication. Do you disagree with that, in general or for Hsuing in particular?
I find "Wayne has told me he wants to use evidence-based reasoning for deciding city policy and has identified as EA for years" to be extraordinarily weak evidence. Anyone can say either of those things.
Not a lawyer, repeating the words of someone else who's not a lawyer.
I talked to my bank (a 3 state credit union) today, specifically asking whether I counted as my own employee. The representative implied the SBA itself didn't know, and if they did they hadn't shared the information yet. She also very strongly implied that as long as I filled out the forms in good faith and my supporting documents were accurate, I would not be punished for fraud.
One kind of reason to do both: It's not a true tradeoff.
This argument comes up a lot in the EA/veganism debate, and I think it's a "minds very different from our own" situation. Some people don't find eating vegan to be costly, or find it cheap enough to not notice. Some people find it prohibitively costly, or so costly that it's not worth considering. What I would ask is that people who find veganism cheap acknowledge that their experience is not universal, and for some people it really is that hard.
This isn't a moral argument. Sometimes the morally correct thing to do is costly. But it doesn't help anything to pretend it's cheap.
Thanks for the plug Milan. For those who don't want to click through: via a grant from LTFF, I've been working on a method for bootstrapping a deep grounding in a subject, starting from knowing nothing. I don't want to take over the thread, but I'm happy to talk about it with anyone who's interested.
Slate Star Codex just published on this. His argument is basically "lots of things look very promising and then fail, and LSD is especially prone to this because it stimulates the insight part of your brain.", although I encourage everyone to read the full post because obviously there's more to it.
If this comment wins a prize I'll pass it on to Scott.
Understanding Conflict in Geographically Distributed Teams: The Moderating Effects of Shared Identity, Shared Context, and Spontaneous Communication
Pamela J. Hinds, Mark Mortensen 2005
“t shared identity moderated the effect of distribution on interpersonal conflict and that shared context moderated the effect of distribution on task conflict”
“spontaneous communication was associated with a stronger shared identity and more shared context, our moderating variables. Second, spontaneous communication had a direct moderating effect on the distribution-conflict relationship, mitigating the effect of distribution on both types of conflict. W”
Insights for Culture and Psychology from the Study of Distributed Work Teams
Screw you google books and your disabling of copy/paste and hiding of pages
The mutual knowledge problem: people act on disparate knowledge without knowing the knowledge is disparate
Note: I skipped descriptions of problems for which I had already described the paper they were based on.
“Language asymmetries” are a big deal.
“Strategies included avoiding conversations and meetings at which the lingua franca would be required, deleting correspondence unread, leaving meetings early, trying to control attendance at meetings to exclude speakers of the less familiar language, and switching languages in the middle of meetings.”
“For example, on the Indian side, team leaders were active work monitors. The importance of tasks was judged by how often the team leader asked about progress. Lack of active leader monitoring indicated to Indian developers that a task was unimportant. This permitted developers to agree to all requests, while tacitly judging the true importance of a task by the level of team lead monitoring. Thus the timing of work was linked to leader monitoring. Work also was often carried out in bursts of activity after a stated deadline had passed. This was because the deadline was understood within culture to be nominal in deference to immediate instructions of superiors.” meanwhile, their German superiors expected to be proactively notified of delays.
In the Flow, Being Heard, and Having Opportunities: Sources of Power and Power Dynamics in Global Teams
“Over the last 10-15 years, research on global teams has grown dramatically, including investigations of team dynamics [6, 8, 23, 33], communication structure [10, 19], conflict [20, 21], coordination [16, 18] and leadership [11, 40]. Scholars have also studied the use of technology to support [13, 17, 33, 38] and alleviate many of the challenges encountered by globally distributed teams [3, 27, 28]”
“ Few studies, however, have examined the power dynamics in global teams, particularly the sources of power and how workers respond to power imbalances. This is despite the fact that the distribution of power can have profound and far-reaching effects on individual behavior  and team outcomes [2, 41]”
“Social power has been defined as “asymmetric control over valued resources in social relations””
“power is more objective whereas status is in the eye of the beholder.”
“work by French and Raven  reported six sources (bases) of power – legitimate (positional), reward, coercive, informational, expert, and referent”
“ O’Leary and Mortensen  who examined the configurational imbalance (relative numbers of team members) across two sites. They found that locations with (numerical) minority subgroups were at a disadvantage compared to locations with a relatively larger number of team members.”
“They report that separation in time and space accentuated differences in sources of status (social capital), but that the onshore leaders of these teams sometimes shared their resources as a way to renegotiate status differences”
9 teams interviewed over 18 months
“Our analysis suggests that team members at some locations felt that they had less influence and power than those at other locations. Interestingly, it was not only those at headquarters or those collocated with the largest number of team members who felt powerful, although both of these were factors. We found that the sources of power among these globally distributed professionals resided at different locations and fell into three categories; access to information (being in the flow), access to decision makers (being heard), and opportunities for growth. When team members perceived that they had access to these resources, they felt that they could work effectively, influence decisions of importance, and have career opportunities, but if few were present, they expressed frustration and dissatisfaction in their jobs.”
“, having the expertise located elsewhere created a dependency on team members at another location, which reduced local developers’ sense of power”
“Having direct access to customers was also a critical resource for developers and varied significantly by location”
Factors that matter in likelihood of power struggle:
Being near executives
Access to customers
Access to decisionmakers (e.g. prod people)
Opportunities for growth (terminal goal)
Team with shared power had equal access to execs and other sources of power
“Power Contested In 4 of the 9 teams in our study, power was contested, sometimes fiercely. Surprisingly, in all of these teams, the sources of power were relatively evenly split across locations.”
“in the absence of a powerful organizational boundary (vendor-client), power contests may be less easily and quickly decided”
“In our analysis, power contests seemed to be dominated by concerns over opportunities for capturing work and for growth.”
“ perceiving opportunities for growth diluted power contests.”
“individuals in power are less likely to share knowledge [26, 30]”
“ two-way channel between leaders and distant workers that would provide more access to strategic information, more opportunity to contribute to decisions, and more visibility between leaders and distant workers. Visibility, in particular, would need to be bidirectional, “
Group projects consisting of two students each from three different schools, sometimes international. Note that there was no control and students are just bad at things, and didn’t have as much time to gel.
“Five types of problems constituting failures of mutual knowledge are identified: failure to communicate and retain contextual information, unevenly distributed information, difficulty communicating and understanding the salience of information, differences in speed of access to information, and difficulty interpreting the meaning of silence”
“unrecognized differences inthe situations, contexts, and constraints of dispersed collaborators constitute "hidden profiles" that can increase the likelihood of dispositional rather than situational attribution, with consequences for cohesion and learning”
“ Mutual knowledge is knowledge that the communicating parties share in common and know they share”
“Establishing mutual knowledge is important because it increases the likelihood that communication will be understood”
“ Proceeding without mutual knowledge, people may speak and understand what is said on the basis of their own information and interpretation of the situation, falsely assuming that the other speaks and understands on the basis of that same information and interpretation”
“ Krauss and Fussell (1990) describe three mechanisms by which mutual knowledge is established: direct knowledge, interactional dynamics, and category membership”
“It is well established that groups that meet face-to-face tend to dwell on commonly held information in their discussions and overlook uniquely held information...When a group's discussion is mediated by technology, the problem seems to be worse”
“ The computer-mediated groups exchanged less information overall and took more time doing it. One of the most robust findings concerning the effect of computer mediation on communication is that it proceeds at a slower rate than does face-to-fac”
“"A delay of 1.6 seconds is sufficient to disrupt the ability of the sender to refer efficiently to the . . . stimuli, despite the fact that the back-channel response is eventually transmitted”
Computer mediated interactions are less rich, leading people to read too much in to what non-verbals they do get.
Computer mediated interaction with people you don’t know well creates feelings of “ isolation, anonymity, and deindividuatio”
It is important that when a problem arises, peolpe attribute it to the correct cause. In particular, don’t blame the person when the situation is the cause
“ people using computer-mediated communication with remote others they do not know well rely heavily on social categorizations to guide their relationships. The social categorizations provide a basis for affiliation if participants share a significant social identity. However, they also can provide fodder for in-group/out-group dynamics if remote others are seen as belonging to social categories different and less attractive than oneself”
“The data were contained in an archival dataset that was created in the course of a collaborative project involving graduate business faculty and students located at nine universities on three continents”
No co-located controls.
“Despite efforts to make project requirements consistent across all nine universities, differences were discovered as the project unfolded”
“s. In seven of the 13 teams, conflict escalated to the point that hostile coalitions formed. In five of these teams, members at two sites began to complain about partners at the third site, refusing in some cases to send them pieces of the team's work or put their names on finished work. Two teams evidenced shifting coalitions among subgroups at the three sites”
“Eventually, Icharacterized five types of problems: (1) failure to communicate and retain contextual information, (2) unevenly distributed information, (3) differences in the salience of information to individuals, (4) relative differences in speed of access to information, and (5) interpretation of the meaning of silence.”
(1) failure to communicate and retain contextual information
E.g. Team members refused to schedule group chat and insisted on phone call, but chat was a required part of the project for other team members
Not informing each other of spring break (which occurred at different times for different schools).
Not informing team members they were disappearing for other projects or tests
Assume co-located team members shared more info than they did
Team members on e-mail missed that other people weren’t included or that their emails were misspelled. “ Sometimes people knew they were exchanging mail with only part of the team, but failed to understand how this affected the perspectives of team members who did not receive the mail, or how it affected the dynamics of the team as a whole”
People retained resentments that others were “unresponsive”, even when they saw proof the person had attempted to communicate and it was not their fault (e.g. misspelled an e-mail address).
“In relationships conducted face-to-face, it is a challenging cognitive exercise to interpret a set of facts from the perspective of another person. It is far more difficult to determine how the information before the other party differs from one's own, and then see things from the other's perspective. Geographic dispersion makes these two activities more difficult because of undetected "leaks in the bucket," because partnerseem to have difficulty retaining information about remote locations, and because feedback processes are laborious”
Private conversations give an inaccurate perception of the pace of work
Differences in the salience of information
Confusion to do indirect wording
“there was a tendency to request feedback from the team indirectly, yet to expect quick responses from everyone”
Incorrect assumption that everyone had 24/7 access to email. Some could only check email at school
Physical distance meant that US <-> US communication was faster than US <-> AUS communication (I expect this to be less of an issue now). This meant everyone’s messages were followed by unrelated messages rather than responses. This was “... attributed to remote partners' lack of conscientiousness”
Got time difference wrong
People failed to convey which topics in an email were most important. In person you can check comprehension as you go and look for reactions.
“t silence had meant all of the following at one time or another: Iagree. I strongly disagree. I am indifferent. I am out of town. I am having technical problems. Idon't know how to address this sensitive issue. I am busy with other things. Idid not notice your question. I did not realize that you wanted a respons”
This turned out to be a very big deal.
“ The data suggest four constellations: (1) good performance, task focus, moderate relationship demands, relatively low volume of communication, and low coalition activity; (2) good performance, high task and relationship demands, relatively high volume of communication, and high coalition activity; (3) weaker performance, relatively high volume of communication, many and diverse information problems, and high coalition activity; and (4) weaker performance, relationship focus, task secondary, relatively high volume of communication, and low coalition activity”
“Information problems seemed to be more damaging to relationships than to task performance”
Highest-harmony team did not grade that great, perhaps because their unwillingness to disagree meant they didn’t hone each others’ ideas.
“ there may be a tendency to generalize such social perceptions, particularly negative ones, to the locational subgroup of which a person is a member, which sets in motion in-group/out-group dynamics (Point ) that are destructive to group cohesion”
“ Members of the teams I studied often failed to guess which of the many features of their context and situation differed from the contexts and situations of remote partner”
“The authors propose that trusting action and demonstrated reliability increase trust in dispersed teams. However, my work suggests that human and technical errors in information distribution may be common in dispersed collaboration, particularly during the early phases of activity. If these are interpreted as failures of personal reliability, they are likely to inhibit the development and maintenance of trust. “
“ when people work under heavy cognitive load, they become more likely to make personal rather than situational attributio”
Situated Coworker Familiarity: How Site Visits Transform
Relationships Among Distributed Workers
Pamela J. Hinds
Stanford University, Stanford, California 94305, email@example.com
Catherine Durnell Cramton
Site visits engender closer co-worker relationships that change behavior upon return to distant sites, continuing the closeness.
“qualitative study of 164 workers on globally distributed teams”, over 1.5 years
“Wilson et al. (2006) demonstrate that trust starts lower in mediated groups but develops to levels comparable to face-to-face groups (although pure electronic groups never achieved the same levels of cooperation as those who met face-to-face)”
“Walther (1992, 1996), for example, leverages social information processing theory to argue that, although the process may be slower, rapport among members of distributed dyads that never meet can eventually exceed that of collocated dyads. Similarly, Wilson et al. (2006) demonstrate that trust starts lower in mediated groups but develops to levels comparable to face-to-face groups (although pure electronic groups never achieved the same levels of cooperation as those who met face-to-face). Field research on virtual teams has also claimed that knowledge repositories are an adequate substitute for face-to-face interaction (Malhotra et al. 2001)”
“Grinter et al. (1999), for example, report that coworkers who did not meet face-to-face had more difficulty creating rapport and developing a longterm working relationship. Alge et al. (2003) also find that coworkers who met face-to-face and got to know one another before embarking on a mediated collaborative task reported as much openness and trust as those who worked face-to-face the entire time. In an early review, Olson and Olson (2000) conclude that face-toface interaction is important to establishing the conditions necessary for distributed work. In their longitudinal study of three globally distributed teams, Maznevski and Chudoba (2000) further describe how regular faceto-face meetings create the rhythms that enable higherlevel coordination. Touching (e.g., handshakes, pats on the back) and “breaking bread” together also have been found to contribute to a state of communicative readiness among distant workers (e.g., Nardi 2005). With few exceptions, research on distributed work consistently points to the importance of face-to-face contact as a means of building trust and rapport (e.g., Orlikowski 2002), translating locally situated knowledge (Sole and Edmondson 2002), interacting more rapidly on tasks (Crowston et al. 2007), and building social networks (Orlikowski 2002). This research generally suggests that periodic face-to-face interaction plays an important role for distributed workers, although what happens during these encounters remains largely a mystery.”
“Mortensen and Neeley (2012) that shows that workers who travel to the location of their coworkers have more knowledge, both direct (about their distant coworkers) and reflected (about their own location), and that this knowledge is associated with feelings of closeness and trust”
“more time together leads to more familiarity, although a plateau seems to be reached fairly quickly in laboratory settings “
Workers felt more comfortable with distant co-workers after meeting in person.
Socializing outside work hours was important
Seeing how co-workers interact with other people is useful
“n seeing workers acting on their knowledge and working directly with them as they did so that coworkers truly began to understand the capabilities of their coworkers and how they needed to work together to leverage those skills most effectively.”
“. Informants told us that after site visits, they and their distant coworkers responded in a much more timely manner to emails and requests”
Not integrating a visitor has a terrible effect on their morale and doesn’t generate any of the benefits of a good visit.
“We speculate that travel needs to occur on a regular basis with intervals of about six months, plus or minus depending on the ambiguity of the project”
“We studied three global virtual teams, collecting data over a period of 21 months: 9 months of intensive observation and collection, preceded by 3 months of informal discussions with the teams and their managers, and followed by 9 months of more discussions.”
“Effective outcomes were associated with a fit among an interaction incident’s form, decision process, and complexity”
“effective global virtual teams sequence these incidents to generate a deep rhythm of regular face-toface incidents interspersed with less intensive, shorter incidents using various media.”
“ In some studies face-to-face groups performed better than technology-mediated groups (e.g., Hightower and Sayeed 1996, Smith and Vanecek 1990); in others they performed worse (e.g., Ocker et al. 1995–1996, Straus 1996); in others there was no difference on quality-related outcomes (e.g., Farmer and Hyatt 1994, Valacich et al. 1993). Furthermore, these relationships changed and evolved over time (e.g., Hollingshead et al. 1993). Although task type was often proposed to moderate the relationship between a medium and its effect on performance (e.g., O’Connor et al. 1993), there did not seem to be a consistent pattern of task types for which communications technology was better or worse. Some studies concluded that a combination of media including face-to-face outperformed one without face-to-face (e.g., Ocker et al. 1998).”
“trust, which was critical to the team’s ability to manage decision processes, could be built swiftly; however, this trust was very fragile.”
“This basic pattern is defined by regular faceto-face meetings in which the intensity of interaction is extremely high, followed by a period of some weeks in which interaction incidents are less intense. Moreover, the decision process is organized to match this temporal pattern, rather than the other way around”
MakeTech and SellTech chose the kind of interaction (phone, fax, etc) to match the task, NewTech did not.
Reflected Knowledge and Trust in Global Collaboration
INSEAD, 77305 Fontainebleau, France, firstname.lastname@example.org
Tsedal B. Neeley
Harvard Business School, Harvard University, Boston, Massachusetts 02163,
“an equally important trust mechanism is “reflected knowledge,” knowledge that workers gain about the personal characteristics, relationships, and behavioral norms of their own site through the lens of their distant collaborators. Based on surveys gathered from 140 employees in a division of a global chemical company, we found that direct knowledge and reflected knowledge enhanced trust in distinct ways”
“some researchers have argued that trust is more easily generated in collocated settings, because collocated workers are better placed than their distributed counterparts to use behavioral cues to read intentions and foster a collective identity (Frank 1993, Wilson et al. 2006)”
“t firsthand experience would be positively related to individuals’ reflected knowledge”
“neither the path linking reflected knowledge to understanding distant collaborators nor the path linking direct knowledge to feeling understood by distant collaborators were significant in our original model”
“We found significant positive paths linking direct knowledge and understanding collaborators ( = 00531 p < 00001) and linking reflected knowledge to feeling understood by distant others ( = 00281 p < 0005)”
An Empirical Analysis of Communication on Trust Building in Virtual Teams
Makoto Shinnishi1orcid, Kunihiko Higa2
Compared 5 teams with text-only communication, “ five teams could use a non-text communication tool through which one can see other member’s situation with a web camera image and a short text message. “
“use of non-text communication tool did not affect trust building; however, amount of awareness communication affected trust building. Log-in to the communication system at the same time also affected trust building. The findings of this study showed the tendency of awareness communication helping team building trust in the remote environment.”
Awareness = “knowledge about the work and worker of current and predicted future status and situation.”
Out of sight, Out of sync: Understanding conflict in distributed teams
Pamela J. Hinds •Diane E. Bailey
“empirical studies suggest that distributed teams experience high levels of conflict”
“In this paper, we develop a theory-based explanation of how geographical distribution provokes team-level conflict.”
“Geographically distributed teams face a number of unique challenges, including being coached from a distance, coping with the cost and stress of frequent travel, and dealing with repeated delays (Armstrong and Cole 2002).”
“Field studies further indicate that geographically distributed teams may experience conflict as a result of two factors: The distance that separates team members and their reliance on technology to communicate and work with one another.”
“Armstrong and Cole (2002) reported that conflicts in geographically distributed teams went unidentified and unaddressed longer than conflicts in collocated teams.”
“t geographical distribution will have a significant impact on each type of group conflict proposed in recent organizational studies: task, affective, and process. Task conflict refers to disagreements focused on work content. Affective conflict (sometimes referred to as relationship or emotional conflict) refers to team disagreements that are characterized by anger or hostility among group members. Process conflict refers to disagreements over the team’s approach to the task, its methods, and its group processes.”
“task conflict has been found to be beneficial for performance on many traditional teams, but we contend that it will not be so for their distributed counterparts.”
“Distributed teams enable firms to take advantage of expertise around the globe, to continue work around the clock, and to create closer relationships with far-flung customers. “
“We contend that all other traits that may be associated with geographical distribution derive from distance or technology mediation, and we consider them in our analysis of these two factors”
“members of distributed teams may have difficulty establishing a shared context”
“in a study of the use of new machines in a factory, Tyre and von Hippel (1997) observed that engineers and operators had trouble resolving equipment problems over the phone because the engineers needed to “see for themselves” the technology in context.”
“When team members have different understandings of the task, task conflict is likely to result (Jehn et al. 1997). Moreover, when team members’ understanding of the issues differs, conflict is difficult to resolve (Brehmer 1976)”
“Team members who lack a sense of a shared context as a result of distance also are likely to adhere to different norms”
“site-specific cultures and expectations acted as significant sources of misunderstandings and conflict between distant sites. “
“Grinter et al. (1999) found that members of distributed software development teams, regardless of the way they structured their work, were “constantly surprised” and confused about the activities of their distant colleagues.”
“research on friendship suggests that distributed teams will experience less friendship and, thus, less affective conflict.”
“studies that suggest that time can remedy the relational problems that ensue from technology mediation”
“Feelings of not “being there” with one’s communication partners stand to prevent distributed team members from sharing relational information that help teams to develop trust”
“technology mediation engenders negative relational effects that we contend will precipitate affective conflict.”
“Several problems related to information sharing and seeking emerge from the literature, including uneven distribution of information, unevenly weighted information, and information that resists transmission.”
“Uneven distribution can occur in at least two ways: Team members may be purposely or accidentally excluded from communications, or members may not reveal information that they uniquely hold.”
“Purdy et al. (2000) reported that student groups working face to face collaborated more than distributed groups working over video, telephone, or chat”
“Longitudinal studies report that groups adapt communication technologies to good effect”
“Over time, effective teams generate a shared team identity.”
“Distributed teams appear to gain more if they meet early in the development of the group (Kraut et al. 1992), enabling members to form relationships that can be supported over technologies (Armstrong and Cole 2002)”
“. If distributed teams are able to meet face to face at the points with the most potential for conflict, then conflict may be reduced or diminished.”
“Establishing collaborative norms, however, may be significantly more difficult in distributed teams.”
Distributed Work (The MIT Press) 1st Edition, Kindle Edition
by Pamela J Hinds (Author, Editor), Sara Kiesler (Editor)
Chapter 4 The Place of Face-to-Face Communication in Distributed Work
Bonnie A. Nardi and Steve Whittaker
Chapter 5: The (Currently) Unique Advantages of Collocated Work
Judith S. Olson, Stephanie Teasley, Lisa Covi, and Gary Olson
“ communication frequency among individuals drops considerably with distance and that after about thirty meters, it reaches asymptote. This means that if two people reside more than 30 meters apart, they may as well be across the continent”
This paper is talking about radical co-location: 6-8 peeps in the same room. Any further than that doesn’t count.
“Teams who experienced radical collocation—pioneer teams—were much more productive than standard teams at both this company and in the industry as a whole.”
“These teams produced twice as much as other teams did in their multitasked work, in standard office cubicles, in projects with more variable scoping. The collocated teams got the job done in about one-third the amount of time compared to the company baseline—and even faster than the industry standard. Both of these differences are significant using a z-score against company baseline (p < .001).”
I think they had other advantages over the control group, such as having only one task. And of course there’s Hawthorne effect.
“Follow-on teams were even more productive than the pioneer teams; function points per staff month doubled again while cycle time stayed about the same. We believe this second increase has to do in part with the fact that some of the team members now had experience (some pilot team members served on follow-on teams), and there was some organizational learning about how to run and manage such groups.”
I’m extremely curious how often team members used the available solo offices.
Productivity measured in function points, so it’s possible they were just better at estimating.
Morale effects (positive and negative) were more contagious. Even the team with shit morale was more productive than standard teams.
White boards and sticky notes were very popular
Tivoli vs. IBM: quite close
Chapter 6: Understanding Effects of Proximity on Collaboration: Implications for Technologies to Support Remote Collaborative Work
Robert E. Kraut, Susan R. Fussell, Susan E. Brennan, and Jane Siegel
“ For example, in collaboration at a distance, communication is typically less frequent, characterized by longer lags between messages, and more effortful.”
Is less frequent a downside?
“Results showed that even in this environment, pairs of researchers were unlikely to complete a technical report together unless their offices were physically near each other, even if they had previously published on similar topics or worked in the same department in the company. “
But international collaborations happen
What tools can replace hallways?
Elizabeth’s observation: this works better for extroverts and multitaskers
“ the most important problem is that when conversation is initiated in person, the people must be simultaneously present”
“ Ancona and Caldwell (1992),demonstrated that problems can arise when people concentrate communication within a supervisory group and fail to exchange enough information with others outside the group”
Chatrooms are a partial substitute for hallways
“Because spoken utterances are ephemeral, unlike messages on an answering machine or in a written document, the listener cannot pause or reread the message when some portion is difficult. Again, however, the ability to ask for clarification partially compensates for the ephemeral nature of speech. When there are many listeners, however, it is far more costly for a single one whose attention has wandered to stop the speaker for clarification”
“ease of local communication and information acquisition may bias the information tracked by a work group, causing them to overattend to local information at the expense of more remote, contextual information.”
The Good, the Bad, and the Unknown About Telecommuting: MetaAnalysis of Psychological Mediators and Individual Consequences
Ravi S. Gajendran and David A. Harrison
“Managers who are unwilling to or who lack the training to
change their management and control styles would likely see
deterioration in the depth and vitality of their connection with
Telecommuting led to: increased feelings of autonomy, lower work-life conflict, better supervisor relationships (not statistically significant), no change on co-worker relationships (low intensity only), worse co-worker relationships (high intensity), increased job satisfaction, increase in objective or external (but not self) reported performance, lower turnover intent, less stress, no impact on career prospect feelings.
This is true even for very high intensity remote work (> ½ time)
Could be supervisor relationship improved because people were happy they got to co-work.
Perceived Proximity in Virtual Work: Explaining the Paradox of Far-but-Close
Jeanne M. Wilson, Michael Boyer O'Leary, Anca Metiu, Quintus R. Jett
“ By understanding what leads to perceived proximity, we also believe that managers can achieve many of the benefits of co-location without actually having employees work in one place.”
“As team members discover that they share a certain social category, they establish a common ground from which they can work “
“the more two people identify with some social category, entity or experience (e.g. profession, gender, ethnicity, common political views, shared trauma, etc.), the more common ground they will have between them and, thus, the more proximal they are likely to feel.”
They use the example of two mothers of young children. I suspect this effect is much stronger if the far-apart people share a trait they don’t share with people nearby.
“Absent a shared identity, people have a strong tendency toward faulty attributions about others’ motives”
This is consistent with other research showing people are more likely to commit the fundamental attribution error against distant people.
“As one subgroup sought to differentiate itself from the other, less communication between the two groups ensued, resulting in even fewer opportunities to discover and develop a shared identity. “
“an organization with strong structural assurance might have very rigorous hiring standards. As a result, employees of the organization would feel comfortable communicating with distant co-workers — safe in the knowledge that they were dealing with competent and reliable professionals”
This was definitely present at Google
“e, the technological infrastructure for open-source projects (version control systems, specialized topic-based discussion forums, and so on) provides rich and stable support for complex work and interpersonal interactions between developers who rarely if ever see each other “
High openness to experience -> less assumption of bad faith of colleague’s part.
“ With experience, members of distributed groups learn to communicate frequently (Jarvenpaa and Leidner 1998), start tasks promptly because of time delays (Iacono and Weisband 1997), disclose personal information (Moore et al. 1999) and explicitly acknowledge receiving messages (Cramton 2001). In essence, experienced tele-workers learn effective norms and routines to be productive in this specific context. ”
“feeling close reduces the uncertainty and ambiguity of working at a distance and improves at least two of the three dimensions of group effectiveness: (1) capability to work together in the future; and (2) growth and well-being of team members”
Having a co-located co-worker visit another site helps your relationship with people at that site, even if you don’t visit yourself.
Too much closeness leads to negative feelings, such as worries about surveillance.
When does the medium matter? Knowledge-building experiences
and opportunities in decision-making teams
Bradley J. Alge,a,* Carolyn Wiethoff,b and Howard J. Kleinc
“t media differences existed for teams lacking a history, with face-to-face teams exhibiting higher openness/trust and information sharing than computer-mediated teams. However, computer-mediated teams with a history were able to eliminate these differences. These findings did not extend to team-member exchange (TMX). Although face-to-face teams exhibited higher TMX compared to computer-mediated teams, the interaction of temporal scope and communication media was not significant. In addition, openness/trust and TMX were positively associated with decision-making effectiveness when task interdependence was high, but were unrelated to decision-making effectiveness when task interdependence was low”
Why This Startup Won't Let the Team Work From Home
Randy Frisch, CEO of Uberflip
“Our dependence on instant messaging and shorthand updates have decreased our likelihood to brainstorm, opting for a quick WTF or LOL, rather than an attempt to build off a crazy idea.”
Marissa Mayer: ““People are more productive when they’re alone, but they’re more collaborative and innovative when they’re together. Some of the best ideas come from pulling two different ideas together.””
How do virtual teams process information? A literature review and implications for management
Petru L. Curs¸eu, Rene´ Schalk and Inge Wesse
“Virtual teams are better in exchanging information and in overcoming information biases. However, they encounter problems in using and integrating information. A greater pool of knowledge leads to higher memory interference in virtual teams. Nevertheless, in virtual teams, processes such as planning and coordination, are less effective and the emergence of trust and cohesion is more difficult to achieve. These opposing effects limit the potential of better knowledge integration in virtual teams.”
“the development of trust, cohesion and a strong team identity is one of the most difficult challenges for managers of virtual teams. “
“Especially in the initial phases of the team project, meeting face-to-face will let the team members get acquainted with each other. Direct contact is essential for the development of trust and cohesion”
Georgia Institute of Technology, College of Management, 800 West Peachtree Street NW,
Atlanta, GA 30332-0520, USA
Lucy L. Gilson
Department of Management, School of Business, University of Connecticut, 2100 Hillside Road,
Unit 1041, Storrs, CT 06269-1041, USA
M. Travis Maynard
Department of Management, School of Business, University of Connecticut, 2100 Hillside Road,
Unit 1041, Storrs, CT 06269-1041, USA
“ researchers have noted the tendency of VTs to possess a shorter lifecycle as compared to face-to-face teams”
“thee number of ideas generated in VTs has been found to increase with group size, which contrasts with results found in face-to-face groups”
“, the addition of video resources results in significant improvements to the quality of a team’s decisions (Baker, 2002)”
“compared to men, women in VTs perceived their teams as more inclusive and supportive, and were more satisfied. Also, in a study of e-mail communication among knowledge workers from North America, Asia, and Europe, Gefen and Straub (1997) found that women viewed e-mail as having greater usefulness, but found no gender differences in levels of usage. Bhappu, Griffith and Northcraft (1997) examined the effects of communication dynamics and media in diverse groups, and found that individuals in face-to-face groups paid more attention to in-group/out-group differences in terms of gender than those in VTs.”
“found that collocated teams reported a significantly lower number of difficulties with various aspects of project management (such as keeping on schedule and staying on budget) than did virtual or global teams”
“Several studies have demonstrated that participation levels become more equalized in VTs than in face-to-face teams (Bikson & Eveland, 1990; Kiesler, Siegel & McGuire, 1984; Straus, 1996; Zigurs, Poole & DeSanctis, 1988). The most commonly cited reason for this is the reduction in status differences resulting from diminished social cues”
“or by team members compared to interactions within face-to-face contexts. In particular, Siegel et al. (1986) found that uninhibited behavior such as swearing, insults, and name-calling was significantly more likely in CMC groups than in face-to-face groups.” This is worse in men.
“In general, lower levels of satisfaction are reported in VTs than in face-to-face teams (Jessup & Tansik, 1991; Straus, 1996; Thompson & Coovert, 2002; Warkentin et al., 1997)”
“Finally, satisfaction in VTs appears to be affected by a team’s gender composition. In particular, all-female VTs tend to report higher levels of satisfaction than all-male VTs”
“For negotiation and intellective tasks, face-to-face teams have been found to perform significantly better that CMC teams, whereas there were no differences found on decision-making tasks (Hollingshead et al., 1993). “
“A type of task in which CMC groups seem to outperform face-to-face groups is brainstorming and idea-generation because there is no interruption from other group members, in effect allowing all members to “talk” at the same time. “
Go (Con)figure: Subgroups, Imbalance, and Isolates in Geographically Dispersed Teams
Michael Boyer O'Leary, Mark Mortensen
“we find that the social categorization in teams with geographically based subgroups (defined as two or more members per site) triggers significantly weaker identification with the team, less effective transactive memory, more conflict, and more coordination problems.”
“imbalance in the size of subgroups (i.e., the uneven distribution of members across sites) invokes a competitive, coalitional mentality that exacerbates these effects; subgroups with a numerical minority of team members report significantly poorer scores on identification, transactive memory, conflict, and coordination problems. In contrast, teams with geographically isolated members (i.e., members who have no teammates at their site) have better scores on these same four outcomes than both balanced and imbalanced configurations.”
For virtual teams to perform well, three enabling conditions need to be established:
“Shared understanding is the degree of cognitive overlap and commonality in beliefs, expectations, and perceptions about a given target. “
“Integration is the process of establishing ways in which the parts of an organization can work together to create value, develop products, or deliver services. “
“ integration refers to organizational structures and systems, while shared understanding refers to people’s thoughts.”
Mutual trust (or collective trust) is a shared psychological state characterized by an acceptance of vulnerability based on expectations of intentions or
behaviors of others within the team
Chapter 2: Knowledge Sharing and Shared Understanding in Virtual Teams Pamela J. Hinds, Suzanne P. Weisband
“ In teams without a common understanding, team members are more likely to hedge their bets against the errors anticipated from others on the team, thus duplicating efforts and increasing the likelihood of rework.”
“A common occurrence on teams is that members think they have come to agreement, but the agreement they believe they have reached is viewed differently by different team members”
“shared understanding among team members has these benefits: • Enables people to predict the behaviors of team members • Facilitates efficient use of resources and effort • Reduces implementation problems and errors • Increases satisfaction and motivation of team members • Reduces frustration and conflict among team members”
Need an understanding of both the task and the process
“telephone may be better than text-based systems for detecting and resolving misunderstandings.”
“Weisband (forthcoming) found that teams that shared information about where they were and what they were doing performed better than teams that did not share this information”
Site visits better than off-site retreats
Chapter 3: Managing the Global New Product Development Network A Sense-Making Perspective
Susan Albers Mohrman, Janice A. Klein, David Finegold
“Sense making in the new product development system is by necessity virtual. It occurs simultaneously within and across different levels and elements of the organizational system. “
This chapter feels buzzwordy as hell and I’m going to skip it
Chapter 4: Building Trust: Effective Multicultural Communication Processes in Virtual Teams
Cristina B. Gibson, Jennifer A. Manuel
“Collective trust is a crucial element of virtual team functioning. “
“collective trust can be defined as a shared psychological state in a team that is characterized by an acceptance of vulnerability based on expectations of intentions or behaviors of others within the team “
“This larger study began in July 1999 and had two phases: in-depth qualitative case analysis with two different virtual teams from each of eight organizations and a comprehensive quantitative survey that will be administered in each firm and analyzed as to statistical predictors of virtual team effectiveness.”
On virtual teams “ trust is harder to identify and develop, yet may be even more critical, because the virtual context often renders other forms of social control and psychological safety less effective or feasible. “
“Some controls actually appear to signal the absence of trust and therefore can hamper its emergence. Institutional controls can also undermine trust when legal mechanisms give rise to rigidity in response to conflict and substitute high levels of formalization for more flexible conflict management (Sitkin and Bies, 1994).”
“interdependence is also critical in establishing trust in virtual teams.”
Compliments build trust
Chapter 8: Exploring Emerging Leadership in Virtual Teams Kristi Lewis Tyran, Craig K. Tyran, Morgan Shepherd
“A leader is said to emerge in a team when the team as a whole reaches a consensus that they perceive the emergent leader to be their leader”
“ We found agreement among team members that leaders emerged in nine of the thirteen virtual teams participating in our study. “
“For traditional teams, trust in a leader’s ability to facilitate team task and relationship interaction effectively has been found to be a critical factor in achieving the consensus necessary for a leader to emerge”
Types of trust:
Can do the tasks to achieve the goal
Leaders were more trusted, especially in task competence
May be harder to emerge as a leader in virtual teams, in part due to difficulty building trust
Leaders sent more messages than average but not necessarily the most messages
“virtual team performance was not clearly related to emergent leadership within the team”
This seems useless
Chapter 10: Overcoming Barriers to Information Sharing in Virtual Teams, Catherine Durnell Cramton, Kara L. Orvis
Skipping the description of problems, assuming it’s similar to Cramton’s other work
“Establish procedures for information sharing within the virtual team. It may be helpful for leaders to distinguish among task, social, and contextual information and to design procedures appropriate to each type of information. For example, task-related information may be best shared at a weekly dial-in conference call in which representatives at each location are guaranteed airtime. It probably is a good idea to designate a facilitator for such conference calls so that time is managed, focus is maintained, and new information is made salient to all.”
Weekly calls include a few moments where everyone says how they are. I think this is incorrect and it makes me doubt their other suggestions. It will chew up a lot of time, people won’t listen, and isn’t enough time to actually share anything relevant.
Longer time frames
Awww, they’re worrying about the cost of phone calls. So cute.
“Build the virtual team’s social identity”
Chapter 14: Influence and Political Processes in Virtual Teams by Efrat Elron, Eran Vigoda
“ Our study relied on ten semistructured interviews with members of virtual teams” Two companies based in US and Canada, with lots of subsidiaries.
“We have witnessed a rapid growth in studies that developed well-grounded models and theories of influence and politics in organizations (Bacharach and Lawler, 1980; Gandz and Murray, 1980; Ferris and Kacmar, 1992; Kipnis, Schmidt, and Wilkinson, 1980; Mintzberg, 1983; Pfeffer, 1992; Yukl and Falbe, 1990)”
“ People in teams that rely heavily on e-mail are therefore more careful with writing than speaking because of the permanency effect of the written word. Thus, political behavior takes on a more careful and covert form when documented.”
Requests to take off-topic discussions offline curtail political behavior
“ Noted one participant, “Virtual teams are task oriented. You do not have enough chances to read and understand this politics, if it’s there at all. In fact, I don’t feel that I actually have enough opportunities to be exposed to such activities in my virtual team. We don’t have that much time left for politics; we need to work.””
“Many multicultural team participants try to be “on their best behavior” when in contact with members from other cultures because they feel that they are not only private individuals but also representatives of their country and culture. As a result, they tend to use tactics that are more acceptable socially”
“ In general, these findings indicate that VOP may be significantly lower than operational politics in conventional teams. Members of virtual teams report lower intensities in the use of influence and political behaviors compared with similar activities of conventional face-to-face groups. In addition, the most dominant influence attempts used are rationality, consultation, and assertiveness, which are considered among the most socially acceptable tactics and also the most effective ones (Yukl and Tracey, 1992). The use of less acceptable tactics such as sanctions, exerting pressures and threats, and blocking information was denied by all our interviewees. “
Okay, I can’t take the seriously after that, skipping the rest
Chapter 15 Conflict and Virtual Teams Terri L. Griffith, Elizabeth A. Mannix, Margaret A. Neale
“Media effects is a phrase used to describe all the outcomes that result from the use of a particular communication medium. “
“practice also suggests that even teams located in the same building may communicate largely by e-mail. In fact, in the organization described in this chapter, all teams spent an equal amount of time working face-to-face on the team task (approximately 13 percent). E-mail use did vary by whether all the members were colocated, but not to a great extent (34 percent of all task communication in colocated teams versus 45.76 percent in distributed teams).”
“What is technically a conflict about the task, for example, may be taken personally and thus be experienced as relationship conflict.”
They looked at twenty eight teams on a spectrum of virtualness
“ We found that virtual teams had greater levels of process conflict than traditional teams, but only when also controlling for the effects of trust. We did not find differences in the levels of task or relationship conflict “
Difference was only .3 out of 5
“Neither trust nor team identification is significantly related to the distribution of the team members.”
“regardless of virtualness, relationship and process conflict have significantly negative effects on performance as rated by the team’s manager. Task conflict, however, does not seem to produce these effects.”
“trust was higher the more the team communicated using e-mail and the more they worked together face-to-face. “
Situation Invisibility and Attribution in Distributed Collaborations
Catherine Durnell Cramton, Kara L. Orvis, Jeanne M. Wilson
Fundamental attribution error again
Chapter 11: Leadership in Virtual Teams:
Zaccaro, S. J., Ardison, S. D., & Orvis, K. L. (2004)
“Zaccaro, Rittman, and Marks (2001) argued that leaders contribute to team effectiveness by influencing five team processes: motivational, affective, cognitive, coordination, and boundary spanning.”
Trust is the expectation that others will behave as you expect = ability to predict others’ behavior
“Team effectiveness is grounded in members being motivated to work hard on behalf of the team. “ that seems incomplete at best
“, team members are more likely to contribute more of their individual efforts to collective action when they trust one another”
“trust, particularly knowledgebased and identification-based trust, evolves from a long period of face-to-face interactions”
“t teams with high levels of initial trust devoted almost half of their early communications during the first 2 weeks of the teams’ existence to discussions of their families, hobbies, and weekend social activities...such communications were not sufficient to maintain trust.”
“trust in temporary teams develops when team members have clearly defined roles with specified and accurate behavioral expectations. “
“When virtual teams form with members having strong professional identities, then swift trust can emerge from stable role expectations and subsequent actions are likely to confirm these expectations,strengthening team trust.”
“effective team coordination depends upon the emergence of a shared mental model”
Skipping the stuff on crampton 2001 since I’ve read the source
“team leaders need to help their teams set performance objectives that are aligned with the strategic requirements operating in the team environment”
The following is presented as an example of how organizing social support is hard, not that this is a bad idea: I find shared meals really stressful, and I know other people do as well. The fact that the default bonding activity is shared meals seems really bad to me. Between eating disorders, medical food restrictions, and simple preferences, just choosing a restaurant or menu can be really fraught.
"Massive drinking" was reported by a friend on my FB wall. I agree with you that it's fraught, but not necessarily more so than the "stand up if this applies to you" game, which one person on FB objected to pretty strongly.
My experience at Wave was that there were things they did that felt supportive and welcoming, that at my previous job would have felt invasive and required an emotional performance. I'm not sure there is a substitute for a healthy workplace.
How have you created social support at work? What things were attempted but backfired?
Examples of things that worked:
"One thing was part of a training we got on diversity and inclusion. As a warm-up sort of opening-up thing, with the whole company in one room, the presenter did a "stand up if..." For a whole bunch of things. You were allowed to just sit the whole time if you didn't want to share, but it was a large group and all of the things were fairly common experiences, that got increasingly taboo. Started out light like, stand up if you have a pet, and got heavier, like, stand up if you've lost a loved one to suicide. And SUDDENLY everyone was a WHOLE PERSON who has loved ones and pain and has been through some shit, and lots of them have been through some of the same shit as you. That said, this workshop happened at a company that was already pretty great about social bonds, and I feel like a lot more people might opt out if my current company tried it...."
"I worked with there was a critical mass of musicians, and a coworker booked some time for us in a music studio to jam together. We had a wide range of musical abilities, but generally were able to play together and have a really good time. It really brought me closer to my coworkers, especially the one who organized it."
Letting a large but racist donor walk away.
Group lunches for identity or interest groups.
A liberal attitude towards employee usage of conference rooms in off hours, e.g. for D+D games.
Explicit channels for compliments towards peers.
Examples of things that backfired:
Branded clothing not available in your size or cut (most commonly, a lack of women's shirts at tech companies)
Scheduling parties outside of working hours.
Almost anything from the above list, if you experience it as invasive rather than supportive.
I don’t often argue the merits of bednets versus cash transfers, which means I get intellectually sloppy knowing I won’t be challenged.
OK, but in that case wouldn't it be better to stick around people with opposing points of view?
This seems like a pretty severe misreading to me. Ozy is saying that they want to hone their arguments against people with expertise in a particular field rather than a different field, which is perfectly reasonable.
I'm in no way associated with EA Funds (although I do contract with CEA), but I can take a guess. Several EA orgs pay for assistants and certain other kinds of help for academics directly, which makes me think that the straightforward interpretation of the statement is true: Nick wanted to fund time savings for high impact people, and academics can't accept money to do that, although they can accept donated labor.
Saying MBSR will have an effect for 38 years after treatment seems extremely generous. Do you have any data either way on how long the benefits of mindfulness last for? I've seen stuff saying CBT works for 5 years on depression(/anxiety) without much of a drop, but 38 years is very long.
The only studies I remember (and it's been a while since I did this) checked in at 18 months and still showed significant improvement. I'd be very interested if you have a source for the CBT numbers, since that seems directly applicable.
What is 'time cost of initial work'?
Time to attend the class.
What does 'negative years of life after treatment' refer to?
guesstimate doesn't have a sigmoid function, so I used a constant minus a lognormal range.
The effects of MBSR on deprssion/anxiety (or is it just anxiety? I haven't check the studies yet) you report are much weaker than I expected.
All of the studies I read were on mild-to-moderate depression or anxiety, and ended with people having no-to-mild depression/anxiety on average. It wouldn't surprise me if in reality some people got very large benefits and others got none, which would make this a somewhat better intervention because people who didn't benefit could stop.
you seem to be assuming the DALY weight of each hour meditating is 0.5, which is roughly as bad as it is to be anxious anyway, no? Surely time meditating can't be that painful.
I think treating it as bad as death is a ceiling we won't reach, but I do think many people find meditation actively unpleasant, especially if they have a mood disorder, relative to doing something distracting. If meditation were more fun than watching TV we wouldn't have to justify it with long term health.
Another worry comes when I ask myself "would this be a good thing for EAs to fund?" It seems anyone with access to the internet could self-teach mindfulness if they really wanted to.
The costs of learning are so low relative to the costs of practicing they end up not making a big difference in the model.
It seems a bit weird for EAs to be paying for the medical treatments of other people in the developed world.
Then my concern is one of cultural barriers and that take up of mindfulness would be quite low (intuitively, this seems like a bigger problem for mindfulness than CBT).
My intuition is the opposite: CBT was developed in a very specific cultural context and seems to take as an assumption that nothing is actually wrong. Mindfulness meditation has been practiced for... long enough that getting an actual date is infeasible.
If the true obstacles aren't money but awareness or motivation, that suggests the better things for EAs to do might be paying for public campaigns that advertise mindfulness, e.g. via Developed Media International. My concern is then neglectedness: there are(/will be) companies trying to market mindfulness to people for a profit. If this is true, EAs might want to leave this to the market to provide and do something else. I'm not quite sure how to think about this either.
You're the second person to argue for this (other was on my personal blog), and I hear the argument. I think there's a slippery slope of what to control for here- if I include sleep, I'd also want to look at how happy people were when meditating relative to the activity it displaced.
For (2): I'll update the doc to reflect this, but in order of increasing abstraction:
'1. Copy StrongMind's model to other poor countries. A SM representative I talked to said they would be happy to help get other charities off the ground.
'2. Research how to optimize mindfulness, make it palatable to the general public, and expand the circle of people it helps/has been demonstrated to help. I met an organization working on this for business at a conference and am quite excited at the potential for increasing both happiness and productive capacity, although I know nothing about the implementation.
'3. Figure out how to measure subjective experience so we can better compare interventions, within and without mental health.
According to the person I talked to, StrongMind's chose their intervention because it was the only one that had experimental backing equivalent to what J-PAL does.
Unorganized thoughts on CBT for the bottom billion:
Poor Americans complain that it implicitly assumes the problem is in your head and is unhelpful if there's an actual problem in your life. I'd guess this would be worse for the bottom billion.
I'd expect significant changes to have to be made to adjust to local cultures.
The maximum cost would be in the ballpark of StrongMinds. The minimum cost would be much lower, I expect CBT to be more amenable to scaling via phones.
I am pretty skeptical of automated CBT. I think it follows the letter of CBT but is missing some implicit step that is actually what is helpful. This is mostly an intuition.
OTOH, it could have a very low success rate and still be very helpful if enough people tried it and the cost of trying was low.
CBT has been getting less effective as time goes on, potentially because the best insights have made their way into popular culture. I wouldn't expect that to be true for the bottom billion, which would increase its effectiveness.
So I think my ultimate answer is "CBT is a good jumping off point to look for something that would indeed be very cost-effective in the third world."
I think costly signaling is the wrong phrase here. Costly signaling is about gain for the signaler. This seems better modeled as people trying to indirectly purchase the good "rich people donate lots to charity.". Similar to people who are unwilling to donate to the government (so they don't think the government is better at spending money than they are) but do advocate for higher taxes (meaning they think the government is better at spending money than other people are). They're trying to purchase the good "higher taxes for everyone".
If Elie had given the money elsewhere, that would have been strong evidence that the other thing was valuable and underfunded relative to GW top charities.
Only if GiveWell and the EA Fund are both supposed to be perfect expressions of Elie's values. GiveWell has a fairly specific mission which includes not just high expected value but high certainty (compared to the rest of the field, which is a low bar). EA Funds was explicitly supposed to be more experimental. Like you say below, if organizers don't think you can beat GiveWell, encourage donating to GiveWell.
I'm shocked that no one has commented on Elie Hassenfeld distributing 100% of money to GiveWell's top charity. Even if he didn't run GiveWell, this just seems like an extra step between giving to GiveWell. But given that one of the main arguments for the funds was to let smaller projects get funded quickly and with less overhead, giving 100% to one enormous charity with many large donors is clearly failing at a goal.
I would guess that $300k simply isn't worth Elie's time to distribute in small grants, given the enormous funds available via GoodVentures and even GiveWell direct and directed donations. It seems to me the obvious thing is to is have the fund managed by someone who has the time to do so, rather than make another way to give money to GiveWell.
I'm concerned in almost the opposite direction- that having the doctor as gatekeeper to something the patient legitimately needs, with the threat of taking it away if the patient doesn't look sick enough, corrupts the doctor-patient relationship and the healing process.
Hypothesis: there's lots of good, informal meta work to be done, like telling convincing your aunt to donate to GiveWell rather than Heiffer International or your company to do a cash fundraiser rather than a canned food drive. But the marginal returns diminish real quickly: once you've convinced all the relatives that are amenable, it is really hard to convince the holdouts or find new relatives. But the remaining work isn't just lower expected value, it has much slower, more ambiguous feedback loops, so it's easy to miss the transition.
Object level work is hard, and there are few opportunities to do it part time. Part time meta work is easy to find and sometimes very high value. My hypothesis is that when people think about doing direct work full time, these facts conspire to make meta work the default choice. In fact full time meta work is the most difficult thing, because of poor feedback loops, and the easiest to be actively harmful with, because you risk damaging the reputation of EA or charity as a whole.
I think we need to flip the default so that people look to object work, not meta, when they have exhausted their personal low hanging fruit.