Understanding contribution of content to public document repositories

Mani Subramani (msubramani@csom.umn.edu)

Naren Peddibhotla

(npeddibhotla@csom.umn.edu) Information and Decision Sciences Department

Carlson School of Management University of Minnesota Minneapolis, MN 55455

Last revised: April, 2005

Abstract There are many publicly accessible repositories on the Internet that are created by firms and professional bodies. Such repositories are created by the uncompensated efforts of individuals contributing content e.g. book reviews or review of attractions at tourist destinations for the benefit of unknown others who may be considering reading the books or planning visits to these destinations. While the potential value of such publicly accessible online repositories is recognized, there has been little prior work examining the motivations of individuals to make contributions to them. We draw on social psychological theory and theory on prosocial behavior and volunteerism to examine this phenomenon and highlight the factors motivating individuals to make contributions to online repositories. We examine these questions using data on the profiles of contributors of reviews to a publicly accessible review repository. We find that contributors have a varied set of motivations for contributing – altruism, reciprocity, social affiliation, utilitarian benefits and enjoyment - that have also been seen in other online contexts. In addition, we observed that people also contributed to repositories because it gave them an opportunity for self-development. We also saw that a potential contributor will be more likely to contribute in an area with fewer prior contributions, and that repositories include content both from experts and non-experts. Keywords: Document repositories, qualitative data analysis, and prosocial behavior.

1

Understanding contribution of content to public document repositories Introduction A variety of public repositories of content in different subject areas are available on the

Internet. Such repositories hold knowledge in the form of documents contributed by

various participants. Some of these repositories involve contributors who do not have an

affiliation to the same organization – e.g., medical practitioners contributing reviews at

www.cochrane.org, individuals contributing knowledge on various topics at

www.wikipedia.org and researchers offering solutions to problems faced by underserved

communities at www.thinkcycle.org. There are also repositories built by business

organizations to facilitate and promote usage of their own products or services or those of

third parties – e.g., reviews of holiday cruises at www.cruisereviews.com, product

reviews at www.dooyoo.co.uk and www.amazon.com. The content in such repositories

are often considered central to the value delivered by online retailers.

In the IS literature, researchers have focused on participant behavior in online forums

such as an email network (Constant et al. 1996), an email listserv (Butler 2001) and

bulletin-boards (Wasko and Faraj 2000). Compared to such forums, the act of

contributing to a repository has both similar and some distinguishing characteristics.

Like email and bulletin boards there are few, if any, social cues (Sproull and Kiesler,

1986) as compared to face-to-face interactions. Moreover, extrinsic rewards for

contribution are often nonexistent and at best, minimal. However, unlike email and

bulletin boards, repository contributions are usually made without any appeals or

requests of help for information. In addition, contributors usually get very little feedback

on whether and how their contributions are helpful to others. While repositories usually

2

provide mechanisms for users viewing the contributed documents to provide feedback on

the quality of the material, providing such feedback is almost always voluntary and it is

often not provided.

Contributors play a critical role in the success of such repositories by contributing content

to them. A greater understanding of user motivations to make such contributions is thus a

very important requirement for their successful implementation. Therefore, our study

addresses the research question is: what are the factors motivating individuals to make

repository contributions? We conduct an exploratory study by suggesting some

propositions based on prior theory and analyze qualitative data in the form of the profiles

of reviewers at a large online public product review repository.

The rest of the paper is organized as follows. In the next section, we present a few

theoretical perspectives that can throw some light on the phenomenon and present our

propositions to address our research question. Following that, we describe our research

methodology including the research site, data collection and analysis methods. Then, we

present the results of our analysis along with our interpretations. We conclude with a

discussion of our findings, limitations of our study and suggestions for future research.

Theoretical perspectives

The content contributed to a repository can be considered a public good since once it is

given to a person, it is costless to provide it to everyone else (Bergstrom et al,. 1986). It is

also difficult to prevent someone who does not pay for the content from using it. An

individual may engage in free-riding by never contributing to the repository but at the

3

same time enjoying the contributions made by others. We therefore first explore how the

contribution of content might be examined using the public goods framework.

According to economic theory, a rational person will contribute to a repository only if he

is provided incentives to compensate for his effort. More importantly, such incentives

will work only when the person is recruited for the task, when he is involved in a stable

relationship with the repository and when his performance can be monitored (Olson

1967). In public repositories, however, there is no active recruiting. Moreover, because of

the large number of participants, it is not possible for a central authority to monitor each

individual and provide tailor-made incentives to induce contribution. As against this

individualistic view of the contributor acting alone, other researchers consider the

contributor as interacting with others within a social context. Public goods theory

suggests that a rational person will not contribute unless he has at least an expectation

that his efforts would be reciprocated by others (Thorn and Connolly 1987). The broader

literature has also identified conditions under which such exchange will occur. However,

the features of contexts studied in prior research are not commonly found in public

repositories. Prior research has found, for example, that social exchange will occur when

individuals know each other, at least in a limited way (Takahashi 2000). We next turn to

prior research in the IS literature using the lens of a social context with respect to

knowledge sharing in email networks and Usenet bulletin boards.

Examining the use of email for help seeking and giving within one organization, Constant

et al. (1996) suggest that citizenship behavior, the desire to benefit the organization was a

motive for helping in the online forums created by organizations. In a public repository,

4

however, the participants are not affiliated to any organization and so would not have the

commitment to an organizational purpose that employees would have. The study of

online helping in Internet newsgroups by Wasko and Faraj (2000) suggests that altruism,

generalized reciprocity and community interest created by ongoing interaction of the

members of these online groups are important motivations. However, unlike in the case

of repositories, participants of a bulletin-board share an understanding and common

interests in the topic that the group is devoted to. For the most part, this occurs because a

bulletin board is created with an intention toward discussion on a specific topic and there

are norms on what kinds of questions can be posted there. In a document repository, on

the other hand, contributors can come from different backgrounds and have different

subject interests.

Contributing to a document repository leads to benefits for others whether or not there are

any benefits for the contributing individual. A rich stream of research in social

psychology has studied such a class of activities called prosocial behavior. Prosocial

behavior covers the broad range of actions intended to benefit one or more people other

than oneself – behaviors such as helping, comforting, sharing, and cooperation (Batson,

1998). Using dispositional and situational factors, and drawing on established social

psychological theories, prior studies have examined different types of planned and

spontaneous prosocial behavior (see Batson 1998 for a complete review). See table 1

below for a comparison between different types of prosocial behavior and repository

contribution.

5

Versus Contributing content to a repository

1 Spontaneous helping (e.g., helping an accident victim)

Involves immediate reaction to a situation requiring help.

Involves planning of one’s time and effort.

2 Planned helping involving kinship or long-term friendship (e.g., shopping for groceries)

Help is specifically asked for

Help is not specifically asked for

3 Contributing time and money (e.g., volunteering, donating money to charity)

What is given away is gone

Content that is contributed can be used in more than one place for more than one purpose

4 Contributing knowledge to a person (e.g., teaching a person, replying to an email query)

Knowledge is given only to the person you choose.

Anybody can access your contribution.

(Table 1) Much of the research on prosocial behavior has studied situations in which an individual

is unexpectedly called upon to help for a brief amount of time, often called spontaneous

helping (e.g., passengers helping an individual who falls in a subway car, Piliavin et al.

1975). The literature has also examined planned helping in the context of kinship

relationships, and other situations such as donating to one’s alma mater, donating blood

and volunteering. We draw upon results from some of this research on prosocial behavior

and prior IS research on helping in online contexts to formulate our understanding of

contribution to public repositories.

Altruism is one of the factors that have been highlighted as motivating helping in the

context of volunteering (Clary et al. 1998). Altruism is defined as helping without

expecting a direct reward. It has been seen as an important motivator for participants at

online forums such as email networks and bulletin boards (Constant et al. 1996; Wasko

and Faraj 2000). Altruism is valued by individuals and contributing to repositories might

6

provide a means to express such a value. People might contribute in order to simply help

others by giving them the knowledge they possess. Therefore, our first proposition is –

Proposition 1: Repository contributions are likely to be driven by altruistic motives.

While helping driven by altruism reflects helping without direct benefits, another factor –

career interests – suggests that helpers provide help because of the recognition of the

broader link between helping and the advancement of their careers (Clary et al. 1998),

and increase in their own economic utility. The positive reputational benefits such as

recognition by peers could also play a role in motivating contributions as well (Shamir

1991). Prior research on online forums has found evidence of such motives driving

participation (Wasko and Faraj 2000; Lakhani and Wolf 2005). Thus we also expect that

people would contribute to public repositories because they see an opportunity to earn

direct rewards (monetary or otherwise), to earn a reputation or to leverage the activity for

benefits in their jobs. We club these driving forces under the term “utilitarian” motives

and propose that:

Proposition 2: Individuals are likely to contribute to repositories because of utilitarian

benefits they obtain.

Social factors such as the norm of reciprocity and gratitude towards the collective for

assistance received in the past also influence individuals to help, for example in a context

such as donating to one’s alma mater (Diamond and Kashyap, 1997). The need for social

affiliation and the need for professional self expression are the other social factors linked

to helping (Constant, Kiesler, Sproull 1994). Such motives have been observed in online

7

forums such as email networks (Constant et al. 1996), bulletin boards (Wasko and Faraj

2000) and open source software development (Lakhani and Wolf 2005). In public

repositories, people may similarly contribute in order to reciprocate other contributors or

socially bond with them by doing similar activities.

Proposition 3: Repository contributions of individuals will be driven by considerations of

generalized reciprocity and social affiliation.

Intrinsic motivation that provides enjoyment for an individual has also been identified as

a reason why people engage in a certain activity, such as helping (Deci, 1972). Intrinsic

motivation has been defined indirectly by excluding that which is not extrinsic, i.e.,

something that is within an individual. In the broad literature in this area, intrinsic

motivation is synonymous with the playful enjoyment that an individual feels when

performing an activity (Deci and Ryan 1985). In open source software development, for

example, people contribute bug fixes partly because of the thrill in writing code to solve a

specific problem (Lakhani and Wolf 2005). A similar set of intrinsic motives seems to

drive participation in email networks (Constant et al. 1996; Butler et al. 2002) and

bulletin boards (Wasko and Faraj 2000). In the case of public repositories too, people

might contribute because they enjoy submitting documents containing what they know.

The joy could potentially arise from constructing arguments and composing prose in the

text of the document. So we propose that:

Proposition 4: Individuals are likely to contribute to repositories if they enjoy the

activity.

8

An individual may not actually provide help in a given situation despite being motivated

to do so for any of the above reasons. Studies of factors that lead individuals to step

forward to provide assistance suggest that the nature of assistance provided by

individuals is linked to their perception of a sense of responsibility to act (e.g.,

Berkowitz, 1978). Experiments in situations, where onlookers witnessed incidents where

they could step forward to help, indicate that the sense of responsibility that any one

person feels for helping is inversely related to the number of other potential helpers (e.g.,

Piliavin et al. 1975). The overall conclusion of these studies is that the presence of

potential helpers diffuses the level of responsibility any individual feels to provide

assistance. This has been termed the bystander effect (Latane and Darley, 1970).

Similarly in public repositories, an individual might be less inclined to contribute on a

topic on which there are many existing contributions.

Proposition 5: The less the number of other contributions to a particular domain, a

potential contributor is more likely to write a contribution in that same domain.

Finally, prior research on prosocial behavior has also shown that people with greater

competence were the people who helped more (e.g., Clary and Orenstein, 1991). The

mere perception of competence in an individual had been found to increase the likelihood

of helping. Promoting the belief that one has the ability to help another individual

facilitated the expression of empathy for that individual (Barnett et al. 1985). Empathy

has been identified as a key antecedent of the altruistic motive (Batson et al 1981), which

in turn leads to helping. Discussions of knowledge repositories within organizations

consider it almost axiomatic that contributions would be made by experts (Davenport and

9

Prusak 1998). Similarly, we might also expect that for public repositories, only the

knowledgeable or competent individuals would step forward to contribute.

We therefore advance our final proposition: Proposition 6: Individuals making contributions to repositories are likely to be content

experts.

Research methodology Research site We examined the above issues using qualitative archived data on reviewers at the product

review repository at Amazon.com. By focusing on contributions to one repository, we

expect to eliminate confounds from the characteristics of the user interface and features

provided by different technology underlying repositories.

The Amazon.com repository is visited by over 14 million users every month. Customer

reviews of products displayed on the screen are just one click away – drawn from the

online repository by a search when the user clicks for more information. Amazon

provides the facilities for any individual to sign up and contribute reviews of books,

music, videos and other products sold on the site. In writing a book or music review and

posting it on the Amazon.com site, reviewers share their opinion on these products with

other users of the repository – information that they believe may be useful in helping the

users decide if the book or CD is worth reading or listening to. The reviews are

moderated – Amazon.com employs a small group of editors who delete inappropriate or

offending content from reviews. Amazon.com allows a reviewer to develop and track a

referent group by creating a “Favorite People” list. When one of those people in this list

10

writes a review or recommends something, Amazon puts it in the individual’s customized

“Friends and Favorites” home page. This allows the reviewer to keep track of members in

his or her referent group and their opinions. Amazon also provides a bulletin board for

discussion among reviewers giving a reviewer and his referent group another forum to

communicate.

Contributing reviewing to the Amazon repository is not compensated – it is entirely

voluntary. However, Amazon ranks reviewers using a composite of the number of

reviews submitted and the average number of helpful votes received by reviews. The

possibility of recognition as a valued reviewer represents the only formal incentive

offered to contributors. Amazon.com also provides all reviewers the facility to create a

personal page on the site with a photograph and up to 4000 characters of information on

themselves.

Data collection and analysis We collected detailed profiles of the set comprising the most prolific reviewers and the

most helpful reviewers (the top 1000). The profiles contained rich personal details that

contributors revealed about themselves. This often included personal details such as

where they lived, how old they were, details about their families and pets, descriptions of

their professional careers, information on their hobbies, their interests, their passions and

pet peeves, the factors motivating them to write reviews, personal life history, their

favorite books and the music that they liked1. Providing profile information is voluntary,

some contributors had opted to provide only their name and nothing else. Most

1 Amazon.com provides a standard template for the profile information but most contributors sampled had chosen to use their own format.

11

contributors however provided a broad range of details about themselves and their

motivations. A representative profile is in Appendix –1.

We also conducted semi-structured telephonic interviews with two of the top 50

reviewers (see Appendix 3 for the interview questions). We obtained the email addresses

of these two reviewers from their posted profiles, and requested them for an interview

after explaining the purpose of our study. Each interview lasted about an hour. We

recorded one of the interviews, and transcribed both soon after the interview was

completed.

Next, we read each profile and coded them into categories. The names of the reviewers

were recorded going top to bottom and the categories were noted going left to right.

In analyzing the data, we followed the techniques of open coding and axial coding

advocated by Strauss and Corbin (1998). We used open coding to categorize the text in

the reviewer profiles into categories suggested by prior theory. We identified keywords

suggesting the different categories of motives such as, for example, reciprocity and

enriched this set with keywords we encountered in instances of behavior driven by such

motives (see appendix 4 for the rules we followed for coding responses into categories

derived from prior theory). We also used explanations suggested by the data in the

profiles. In creating categories derived from the data, we often backtracked to earlier

profiles to check if any of those profiles could be coded into a new category created

based on a new profile.

12

Second, after coding the data, we grouped the categories that reflected similar concepts

and themes, consistent with the notion of axial coding. These steps highlighted the core

phenomenon of the motivations underlying contribution of reviews and we used the

linkage between the categories to infer the theoretical explanation.

An alternate coder then coded a random set of profile texts drawn from the sample

independently using the categories drawn up by the first coder. This was done to verify

that those profile texts could indeed be classified under the categories chosen by the first

coder. Next, the alternate coder examined a random set of categories and examined the

profiles listed under each of the categories to confirm if the profiles warranted

classification under that category.

These procedures also ensured validity (ensuring accurate interpretations), and reliability

(consistency of the interpretations) of the conclusions drawn from the data (Yin, 1994).

Results Descriptive statistics

At the time of the data collection, we counted 814,948 people who had contributed at

least one review at Amazon. Of these we studied the profiles of the top 1009 reviewers2.

The descriptive statistics of the population as wells as our sample of Amazon reviewers

are as follows:

2 We had set out to sample the top 1000 reviewers. However, many individuals shared the same position at rank number 986. We considered all these individuals and thus ended up having 1009 reviewers in our sample.

13

Population Sample Number of reviewers

814,948 1009

Number of reviews

Total 2,082,765 152,065

Mean per reviewer

2.56 150.71

Median per reviewer

1 107

Mode per reviewer

1 64

Highest per reviewer

3146 3146

Lowest per reviewer

1 18

Std. deviation 13.34 186.29 Number of helpful votes

Total 7,898,974 892,383

Mean per reviewer

9.69 884.42

Median per reviewer

0 612

Mode per reviewer

0 563

Highest per reviewer

21872 21872

Lowest per reviewer

0 192

Std. deviation 299.70 1306.47

(Table 2)

From the above table we see that the individuals in our sample are very prolific

contributors, with an average number of reviews per person that is about sixty times the

entire population average. The top five reviewers each contributed over 1000 reviews. On

the other hand, most of the individuals in the population are “one-review” contributors

with no new contributions after their initial effort. Unlike these casual contributors, the

individuals in our sample have contributed a minimum of eighteen reviews.

14

The reviewers in our sample are also contributors of the most useful reviews, with an

average number of helpful votes per person that is about ninety times the average per

person for the entire population. Among our sample of top 1009 contributors, reviews

received an average of about six helpful votes per review, with a minimum of about one

helpful vote per review on average per individual. In contrast, the reviews contributed by

most individuals in the population have not been rated as useful by others.

Even among the top 1009 reviewers, there is a large variability both in the number of

contributed reviews, and the number of helpful votes per reviewer. Amazon determines a

reviewer’s rank based on a combination of the number of contributed reviews, the

number of helpful votes and the number of “not-helpful” votes. So it is possible for an

individual who has contributed more reviews than another person to be ranked lower than

the latter because of fewer helpful votes. Whereas an individual can control the number

and quality of reviews he contributes, the number of helpful votes is determined by

others.

Out of the entire population of reviewers, 82,569 (10.13%) had written profiles about

themselves, as compared to 900 (89.20%) in our sample. Each profile in our sample

contained 115.46 words on average, with the maximum being 735 words.

The following table shows the background of some of the reviewers in our sample of top

1009 contributors, based on our reading of their profiles. From table 3, it is evident that

the contributors in our sample of top 1009 reviewers came from a wide variety of

backgrounds, and are not biased toward a narrow group in terms of what they do apart

from writing reviews.

15

Background / profession Number (% of reviewer sample) Priest / evangelist / aspiring priest 18 (1.78%) Performing artist / performing arts related job / aspiring to such a vocation

22 (2.18%)

Moviemaker / aspiring moviemaker 8 (0.79%) Musician (both amateur and professional) / aspiring to be a musician

36 (3.57%)

Armed forces 16 (1.59%) Librarian / library aide 15 (1.49%) Psychiatrist / Psychotherapist / Counseling 18 (1.78%) Publishing industry / music industry / literary business / aspire to be in one of these

28 (2.78%)

Owner / worker in a bookstore / video store 7 (0.69%) Academic job / home school teaching / aspiring teacher

90 (8.92)

Planning for teaching career 12 (1.19%) Journalist / journalism related work / aspiring to be a journalist

28 (2.78%)

Editor 34 (3.37%) Lawyer / legal assistant / aspiring to be in legal field

18 (1.78%)

Tourist guide / travel agent 2 (0.2%) Consultant 33 (3.27%) Trainer / instructor 16 (1.59%) Mentoring role 2 (0.2%) Nurse 3 (0.3%) Commentator 2 (0.2%) Involved in book club / literary group / music forum / writers’ workshops

23 (2.28%)

Writer of books / plays / songs 74 (7.33%) Writer in newspapers / periodicals / other publications; technical writer / screenplay writer

96 (9.51%)

Made music CDs 5 (0.50%) Web content writer 74 (7.33%) Actor / graphic artist / artist of any kind / aspiring to be an artist

21 (2.08%)

Employee of Amazon 1 (0.1%)

(Note: Some reviewers might fall in more than one category above)

(Table 3)

Motivations to contribute

16

Altruistic motive: Our coding of the reviewer profiles in the sample revealed the following altruistic

motives:

S. No. Type of response / observation Number of

occurrences Percentage of occurrences (%)

1 “Become noble” 1 0.1 2 “Because it is my duty / because

of social compassion” 5 0.50

3 Share experience derived from the book

58 5.75

4 To promote the reading habit 3 0.3 5 Enabling others to pick and

choose books, movies etc. 77 7.63

Total 136 13.48 (Note: The total figure above does not equal the sum of occurrences of individual categories since some profiles contain more than one of the above types of responses.)

(Table 4) By recommending or criticizing products sold on Amazon, a reviewer is saving the

readers of his / her review the cost involved in searching for good products. By sharing

one’s experience with the use of a product, a reviewer is making it easier for a reader of

his / her review to make a purchase decision. A person who is encouraging others to read

in general is showing potential readers the value they can gain from doing so. An

individual who is contributing reviews is thus providing a benefit to others. In their

profiles, such people who expressed the sentiment of helping did not indicate any

apparent benefit they obtained in return for contributing reviews. At an abstract level, a

reviewer sees this altruistic motive as being directed toward the entire community of

potential readers of his review. The following excerpt from a reviewer profile shows an

example of such an expression of altruism:

17

“Wanting to help is the primary reason I write book reviews on Amazon.com. Most

people don't have enough time to read so I want to be sure that people know which books

can help them, and which cannot, and for what purposes. I also like to point out the

bigger questions that sometimes the books and other reviewers fail to articulate.”

The data in the above table thus show that a small proportion of individuals (13.48%)

contribute reviews due to altruistic motives. This proportion is similar to the prevalence

of altruistic motives in a study by Wasko and Faraj (2000), who found 9.8% of

participants of online networks being driven by altruism. Similarly, participants in

another online network studied by Constant et al. (1996) gave a 16% weight to altruistic

motives among all motives for replying to email queries.

One of the reviewers whom we interviewed disclosed an altruistic motive when he said:

“When I write a review, I am saying: ‘here is my opinion’. I am trying to help others in a

purchase decision.” Knowing that his reviews were being found useful was important to

him: “I check my reviews everyday to see my vote count. It gives me an indication of

someone having read it.”

In addition, what’s interesting is that this reviewer indicated that the altruistic motive got

him started into contributing reviews, though it seems to have declined in importance as

time passed by: “Contributing reviews is like a social act. Is it altruism? Partly, I also

write reviews to motivate others to write reviews. In my initial days of writing reviews, it

was more altruistic. Later you get free stuff for writing reviews.” This suggests that

individuals perhaps need not have sustained altruistic motivation to be willing to make

18

sustained contributions to a repository. It might be enough to altruistic occasionally.

Other motivations can make it worthwhile for continuing to contribute.

The other reviewer whom we interviewed made a distinction between simply giving an

opinion in a review vs. helping others in their purchase decision: “Reviewing for a

popular book such as a Tom Clancy novel is an expression of opinion, a part of an

ongoing discussion. This is not of value in the same way as reviewing children’s books

where I am interested in telling teachers what you will get in a particular book.” In other

words, altruism is not just about doing something (giving opinions, criticizing, expressing

feeling about a product) that one would otherwise do in other contexts (e.g., among

friends). It is more about going out of one’s way and making an effort to address

someone’s need (explaining what’s in a product for a potential user, giving a balanced

view).

As a counterpoint to the above findings in computer-mediated communication (CMC)

contexts, we can examine the role of altruism in situations that are very similar to

contributing to a repository. For example, volunteering involves individuals actively

seeking out opportunities to help, committing time and effort to the activity, and

continuing with helping others over a long period of time. Clary et al. (1998) identify six

functions that volunteering activity can serve for an individual, one of which relates to

individuals expressing values related to altruistic and humanitarian concern for others.

They find that the relative weight of altruistic motivation (measured by the mean score on

their volunteer function inventory scale) as compared to other motivations is 25.37% for

people actively involved in volunteering and 20.75% for those who are not. Though these

19

are approximations of the relative strength of the altruistic motive, we note that this

motive is more important in volunteering contexts as compared to repositories.

Contributing to a repository is also similar to alumni making donations to their alma

mater. It involves a deliberate and planned decision to help anonymous others by

providing funding to an institution over time. In their survey of alumni, Diamond and

Kashyap (1997) found that the perceived need of the alma mater has a relative weight of

25.81% as compared to other motives for donating to their alma mater. The perceived

need of the recipient as felt by an individual results in the altruistic motivation according

to the empathy-altruism hypothesis (Batson et al., 1981). We again see the higher

incidence of the altruistic motive with respect to other motives here as compared to a

repository context.

Utilitarian motive: In contrast to altruistic motives, fewer reviewers stated utilitarian motives to explain why

they wrote reviews, as shown in table 5:

Type of response / observation Number of

occurrences Percentage of occurrences

In order to obtain free copy of book

3 0.3

To see one’s writing in print 3 0.3 For monetary gain 1 0.1 To be recognized as a reviewer by authors by including blurb on book

1 0.1

To promote one’s business 11 1.09 To be recognized a top reviewer / expert in a particular field

10 0.99

Total 29 2.87 (Table 5)

20

Each of the above motives involves obtaining a material benefit in return for contributing

reviews. The benefit could be tangible (in the form of free products, or money), or

reputational wherein the reputation leads to tangible benefits in other ways to the

contributor. A socially undesirable quality such as selfishness would not tend to be

concealed as a motive when writing one’s profile because of response anonymity

(Aquilino 1998). So we can consider 2.87% of the individuals to be contributing in

exchange for personal gain. This proportion corresponds to the 3.1% of participants in the

study by Wasko and Faraj (2000), who indicated personal gain as a reason for

contributing to a Usenet group. It is also similar to the figure of 5.7%, which is the weight

given to selfishness among other reasons for answer queries by participants of another

online network studied by Constant et al. (1996). In their study on volunteering, Clary et

al. (1998) found that active volunteers rated the importance of the “career functional

motive” (that gives career-related benefits) as 10.80% among various motivations to

engage in volunteering. This figure was 17.54% for individuals who were not involved in

volunteering. The low number of occurrences of utilitarian motives in our study suggests

that altruism is more predominant compared to utilitarian motives in determining the

willingness of a reviewer to contribute.

Thus, we find that altruistic motives can be a potential reason why people contribute to a

repository though to a lesser extent as compared to traditional helping contexts, and that

they might be more important than utilitarian motives.

21

Reciprocity: The reciprocity effect is also evident in the profiles of the top 1000 reviewers (see table 6

below). Individuals indicated that they wrote reviews in return for the benefit they had

received from other reviewers in the past. We coded reciprocity as a motivation by

mentions of norms of conduct, obligations and trust and from statements reflecting affect

towards the community.

Type of response / observation Number of

occurrences Percentage of occurrences (%)

Benefited from reviews written by others

42 4.16

Helping others in repayment for help received in the past

13 1.29

Writing reviews in the expectation that others will help me

1 0.1

Total 49 4.86 (Note: The total figure above does not equal the sum of occurrences of individual categories since some profiles contain more than one of the above types of responses.)

(Table 6) The above responses indicate that an individual can contribute a review either because he

had benefited from reviews written by others in the past, or in order to be able to receive

the benefits of reviews that will be written by others in the future. It is also apparent that

an individual contributes not for the benefit of another person from whose review he had

benefited in the past, but for the potential benefit of all the participants in the review

repository.

As some reviewers admitted in their profiles: “I got started reviewing because I enjoyed reading lots of other people's reviews to

choose a book.”

22

“I have consulted Amazon's public reviews for years before making a purchase and I

decided to start giving back to the Amazon community”

“I know I read these reviews prior to buying any book and they have been excellent help,

so if I can steer someone to one they will enjoy, well, then I've paid my dues.”

One of the reviewers whom we interviewed disclosed the reciprocity motive for

contributing reviews when he said: “I browse reviews written by others when I am buying

something. So, in a way, I write reviews because of the altruistic efforts of others

benefiting me.” Beyond merely reacting to the benefits he received from the reviews

contributed by others, he was also active in encouraging others to contribute: “I also

write reviews to motivate others to write reviews.” This suggests that an individual might

contribute to not only reciprocate others whose contributions he benefited from, but also

to potentially benefit others who might be expected to reciprocate him in the future.

However, only a small percentage of individuals indicate a reciprocity motive for their

contributions. There is thus little evidence of people contributing owing to a sense of

reciprocity to other reviewers. Further, this is to a lesser extent than found in past

research on helping in online communities. Wasko and Faraj (2000) found that 13.4% of

participants of Usenet groups participated out of a sense of reciprocity. Similarly,

Constant et al. (1996) had found that members of an email network gave a weight of

11.8% among reasons for responding to requests for help.

Reciprocity has also been studied as a motivation in non-CMC contexts such as alumni

making donations to their alma mater. Diamond and Kashyap (1997) found reciprocity

23

has a relative weight of 13.98% as compared to other motivations in determining an

individual’s obligation to support his alma mater in various ways. Again, we see a much

more important role for reciprocity in determining a willingness to contribute, as

compared to the context of a repository. This is perhaps indicative of the distinction

drawn by Yamagishi and Cook (1993), who compared “group-generalized” exchange vs.

“network-generalized” exchange and found greater sharing of resources in the latter than

in the former. In group-generalized exchange, group members contribute to a common

pool and each member receive benefits as a result of this pooling. In a community of

three individuals “A”, “B” and “C”, for example, A’s giving of resources to the pool

would make the resources available to both “B” and “C”. “A” has no direct interaction

with either “B” or “C”. Contributing to a repository has this feature as the knowledge that

is contributed once can be used repeatedly. In the network-generalized exchange,

participant “A” gives resources directly to another individual “B” who does not repay

“A” directly. Individual “A” instead receives resources from a third participant “C”, who

received resources from “B”. In this particular exchange cycle, A’s resources are given

only to “B” and not to the entire group. Similarly for B’s giving of resources to “C” and

C’s giving of resources to “A”. Donating to one’s alma mater, responding to queries on a

newsgroup or email network all have features similar to this form of exchange.

Finally, it is interesting to note that while reciprocity has been important in research

dealing with public goods (e.g., Thorn and Connolly, 1987), our findings show that it is

not a major factor in explaining why people contribute reviews.

Bonding with others (social affiliation):

24

The reviewers also considered it important to bond with a community of reviewers,

customers, publishers and authors as mentioned in the following reasons for reviewing

given in their profiles:

Type of response / observation Number of occurrences

Percentage of occurrences

Bond with others with similar interests in books / music / movies etc.

49 4.86

Bond with others with similar experiences NOT connected with books / music / movies etc.

2 0.2

Reviewing to receive feedback 129 12.78 Share experience derived from a product*

58 5.75

Influence the lives / thoughts of others

9 0.89

Contribute to the literary cycle 1 0.1 To do service to authors with praise and feedback

4 0.40

Total 208 20.61 (Note: a. The observation marked “*” has also been interpreted as an altruistic motive in table 4 above, b. The total figure above does not equal the sum of occurrences of individual categories since some profiles contain more than one of the above types of responses.)

(Table 7) By contributing reviews an individual can interact with other participants at the

repository, whether directly by one-one interactions or indirectly by responding to

reviews contributed by others. The responses in table 7 show that reviewers might

consider themselves as part of a group of like-minded individuals. A comment that is

indicative of this is shown in the following excerpt from a reviewer profile:

“This is so cool that Amazon permits us book lovers the space to share our thoughts

about what we're reading. I love to peruse other peoples' thoughts on the books I'm about

to buy and enjoy exchanging comments and ideas with other readers.”

25

What individuals seem to value most is the interactions with others who are not part of

their regular circle of acquaintances, as expressed by one of the reviewers whom we

interviewed: “Most pals, buddies know about my writing reviews. They do not look at my

reviews. Feedback from readers of my magazine reviews is usually from people whom I

know. What is noteworthy is the feedback from customers at Amazon: people who do not

know you. I get mail from people all over the world.” This suggests that a repository that

spans or is accessible by a large cross-section of individuals will be more attractive to

potential participants.

This incidence of the motivation to socialize with others is similar to another CMC

context. In their study of Usenet groups, Wasko and Faraj (2000) found that 18.7% of

individuals participate because they want to be part of a community of practice. In those

newsgroups, participants get to exchange multiple viewpoints, draw upon the experience

of peers, and also maintain and advance the community as a whole.

Among non-CMC contexts, Clary et al. (1998) found that individuals who were active in

volunteering gave a 10.21% weight to social motivations as compared to other

motivations for engaging in volunteering activity. For individuals who were not involved

in volunteering, the figure was 11.4% in determining an intention to engage in

volunteering. Diamond and Kashyap (1997) found that alumni gave a weight of 34.41%

to “individual attachment” (indicating connection and unity with other alumni) among

other motivations in determining the obligation they felt for supporting their alma mater.

Contributing reviews to a repository is different from interactions on a newsgroup or

other communication media, or the above face-to-face contexts. As noted earlier,

26

contributions are made without specific requests for help from particular individuals. Yet

we find 20.71% of individuals indicating that they contribute reviews to socialize with

other participants of the repository. These findings suggest that the act of contributing a

review to a repository is more than just an interaction with a database. The high incidence

of the social motivation in the contexts of both repository contributions and alumni

support for their alma mater is interesting. It seems to reinforce our above inference that

the activity of contributing reviews is a way for an individual to connect with others like

oneself. This is apparently less of a case in volunteering if the activity allows contact

mostly with the recipients of the volunteer’s help and less with other volunteers like him.

Intrinsic motive We also found that some reviewers were writing reviews for the enjoyable experience

that the activity provided, as described in the following reasons given in their profiles:

Type of response / observation Number of

occurrences Percentage of occurrences

For a creative release 3 0.30 For fun / relaxation 20 1.98 Total 23 2.28

(Note: The total figure above does not equal the sum of occurrences of individual categories since some profiles contain more than one of the above types of responses.)

(Table 8) For a task-oriented situation such as writing a review, these individuals feel a sense of

“flow” (Csikzentmihalyi 1975) or enjoyment from an activity that is equivalent to

playing. For example, some of the reviewers said in their profiles:

“… and reviewing books is remarkably cathartic”.

27

“I have taken great pleasure and profound joy in writing reviews about the works of

artists in the entertainment field.”

“Since I write reviews for my own enjoyment, I tend mostly to review books I enjoyed…” “I am doing this for fun and imagine that, besides myself and I, no one else will ever read

this.”

One of the reviewers whom we interviewed said “Reviewing is fun. I do it for my own

enjoyment. If others enjoy it, that is fine.”

In prior studies of online forums, 6.5% of people indicated such enjoyment as a reason

for participating on Usenet groups (Wasko and Faraj 2000), and 9.5% for an email help

network (Constant et al. 1996).

There is thus a relatively infrequent occurrence of the enjoyment motivation among

contributors to repositories as compared to the other types of online forums. Self-

determination theory (Deci and Ryan, 1985), which posits that an individual needs to feel

competent and in control, may suggest a reason for this observation. According to this

framework, the intrinsic enjoyment motivation arises when a person succeeds on a task

that requires some skill. This is the case with solving the problem posed by someone on

an email network or a bulletin board, or using a new computer technology (Venkatesh

2000). Writing a product review, on the other hand, does not address a specific problem

posed by others. The level of detail and the sophistication of the argument in a review is

entirely defined by the writer, thus making a non-task-oriented situation. As a result,

there would be a lesser scope for deriving joy from the activity.

28

Choice of content contributed: We also found that people tend to choose areas to contribute in based on the contributions

made by other reviewers. We coded profiles into the categories in table 9 based on

statements saying that the reviewer chose to write about products that nobody (or a few

people) had reviewed, or had provided inadequate reviews in their opinion.

Type of response / observation Number of occurrences

Percentage of occurrences

Bring awareness about the books, CDs / genre of books, CDs that customers may not otherwise learn about

46 4.56

No one else has reviewed those CDs / books etc.

16 1.59

Add information that other reviewers have not covered

11 1.09

Provide a complete assessment (as compared to professional reviewers)

4 0.40

Total 64 6.34 (Note: The total figure above does not equal the sum of occurrences of individual categories since some profiles contain more than one of the above types of responses.)

(Table 9) In the above responses, contributors are stating that they tend to contribute reviews about

a particular book, CD, toy etc. (or in a particular area such as, e.g., art in the Middle East)

for which few (or no) other individuals have made contributions. Even if others have

contributed reviews, an individual tends to contribute his / her own review only if he / she

feels that the existing reviews in that area are inadequate in some way.

An example of this can be seen in the following excerpt from a profile:

“I try to review items that have no reviews, especially if there is no publisher description.

I want to encourage readers to try some of the more obscure titles that I enjoy. You may

29

note that most of my reviews have no feedback ... that is, at least in part, because few

people wander into these wonderful bycorners of literature/theology/folklore.

Do I read more conventional books? Yes, but if there are already reviews that

approximate what I think needs to be known, I don't bother to review them.”

One of the reviewers whom we interviewed also stayed away from contributing to areas

that others had contributed in, when he noted: “For new music CDs, I review everything.

But for catalog material (stuff that is on my shelves at home), I do not write reviews if

others have written reviews.”

There can be many individuals who can potentially contribute a review for a particular

product for the benefit of customers. Whether a particular individual steps forward to do

so can depend on the number of other individuals who have already taken upon

themselves to contribute a review. If an individual sees that others have already

contributed adequately in terms of covering all issues about the product in question, he

might feel less inclined to make his own contribution. The final word might not have

been written in all the existing reviews about the product. There might be even be many

gaps or flaws in those reviews. Yet a potential individual might not be keen about

contributing his own review simply because there are reviews contributed by others as

noted in the above responses in some profiles. This diffusion of responsibility is

consistent with prior theory that the greater the number of bystanders who witness an

emergency, the less likely is one of them to help the victim (Latane and Darley, 1970).

30

This effect might not be operative for individuals who contribute reviews for the benefit

of other reviewers (instead of potential buyers of the product). A reviewer whom we

interviewed said: “Reviewing is both like writing for a database and for other people.

People who look at reviews have already made up their minds. I think most votes that

people give reviews come from other reviewers. I myself vote for other reviewers’ reviews

before I write my own review. Most of the stuff I review has already been there for a

while (i.e., has not been recently published). I think half the voters are buyers and the

other half are reviewers.” So if an individual assumes that his reviews are being read and

appreciated by other reviewers, then his review can be part of an ongoing discussion, an

ongoing joint effort to extend the knowledge about a subject. Thus, the intended target for

one’s contribution can determine whether an individual is dissuaded or encouraged to

make a contribution in an area with many existing contributions.

Who contributes? Tapping helpers versus tapping experts:

In the context of product reviews, a contributor can be said to be an expert if he / she has

educational qualifications or experience (as part of a professional career or in terms of

usage of current product and / or products similar to the one being reviewed) in the

subject that he / she is reviewing in. As expected, we did find instances of reviewers

writing about subjects in which they were experts.

For example one reviewer stated this in his profile: “I also have a long time interest in history, and developed some expertise in medieval

history while assisting my brother in preparing the family genealogy.”

31

In the course of our detailed reading of the profiles of the top 1009 reviewers, we found

an individual with a PhD in statistics who reviews only books on statistics and a house

construction / repair contractor who writes only about tools and hardware. Many of the

reviewers also stated that they were writers of books/plays/songs, musicians, performing

artists, moviemakers, thereby indicating some level expertise in their respective areas in

which they wrote reviews. For example, among the reviewers we had editors (3.35%),

consultants (3.35%), and authors / composers (7.3%). One of the reviewers whom we

interviewed, who reviews music CDs, had been involved with evaluation of music

records in a 20-year career as a disc jockey and a 10-year experience as a music reviewer

in print magazines.

On the other hand, we also observed that many reviewers wrote on topics that were not

focused on one area of specialty. Amazon classifies the various products it sells into

broad categories such as books, music, videos, toys, and electronics. Each of these

categories has subcategories as shown in the table below:

Sub-categories in books Sub-categories in music Sub-categories in video Arts & Photography Alternative Rock Action & Adventure Biographies & Memoirs Broadway & Vocalists Animation Business & Investing Children's music Anime & Manga Children's Books Classic Rock Art House & International Computers & Internet Country Classics Cooking, Food & Wine Dance & DJ Comedy Entertainment Folk Cult Movies Gay & Lesbian Hard Rock & Metal Documentary Health, Mind & Body International Drama History Jazz Horror Home & Garden Latin Music Kids & Family Horror Miscellaneous Military & War Literature & Fiction New Age Music Video & Concerts Mystery & Thrillers Pop Musicals & Performing

Arts

32

Nonfiction R & B Mystery & Suspense Outdoors & Nature Rap & Hip-hop Science Fiction & Fantasy Parenting & Families Rock Special Interests Professional & Technical Television Reference Westerns Religion & Spirituality Romance Science Science Fiction & Fantasy Sports Teens Travel Sub-categories in toys Sub-categories in

electronics Birth-12 months Brands Accessories & Supplies 12-24 Months Dolls Computer Add-ons 2 Years Electronics Handhelds & PDAs 3-4 Years Favorite Characters Office Electronics 5-7 Years Games 8-11 Years Grownups 12-14 Years Interests Action Figures Sports & Outdoor play Activities & Learning Stuffed animals Age Ranges Vehicles & Die-cast Arts & Crafts

(Table 10) To obtain data on the number of product categories in which individuals wrote reviews,

we sampled 50 reviewers ranked 1-10, 101-110, 251-260, 501-510, and 1000-1009. For

each individual, we went to his / her profile page, and noted the number of product sub-

categories for which he had written reviews using the counts provided by Amazon. We

tabulated these counts and calculated the number of sub-categories that each individual

had written reviews for. Table 11 presents a summary of our findings:

33

Statistic All sub-categories3

Sub-categories in Books

Sub-categories in Music

Sub-categories in Video

Sub-categories in Toys

Sub-categories in Electronics

Maximum 52 21 15 17 15 2 Minimum 1 0 0 0 0 0 Average 13.72 7.72 2.4 2.4 1.1 0.1 Std Deviation

10.21 6.18 3.98 3.90 2.79 0.36

Median 12 6.5 0 0 0 0 Mode 7 7 0 0 0 0 Count 50 46 20 23 10 4

(Table 11) Most individuals had written reviews for books, whereas only four out of the fifty had

written reviews for electronics items. Within each category, we see a large variation (in

terms of range and standard deviation) in the number of product sub-categories in which

reviews were written.

Expertise is usually built after continued working in one or a few related domains

(Ericsson, 1996). As a result, the fewer the number of categories in which a reviewer

writes reviews, the greater the expertise that he or she can be said to have in those areas.

The large variation in the number of categories reviewed by reviewers indicates that there

is a mix of both experts and non-experts. One of the reviewers whom we interviewed has

a PhD in rhetoric but writes reviews in 52 categories covering books, toys, music, video

and electronics.

In addition, a significant proportion of the reviewers cannot be deemed to be experts in

one or more areas according to their own admission. Many people indicated that they

3 The column “All categories” has data for all categories of products – books, music, video, toys and electronics, considered as one group. Data in this column cannot be derived from data in the columns for individual categories. For example, there is an individual who has contributed in a maximum of 52 different categories all considered together. But he is not the same as the individuals who contributed the maximum in the individual books, music, video, toys and electronics categories.

34

were striving to attain expertise in areas such as performing arts, movie-making, and

music. For example, the table below shows the number of individuals who aspired to be

experts in writing skills and did not see themselves as experts currently.

Type of response / observation Number of

occurrences Percentage of occurrences in sample

Aspiring to be a published writer 76 7.53 Hope to be recognized as a top reviewer / expert in a particular field

10 0.99

Improve one’s writing 10 0.99 Total 91 9.02

(Table 12) The above reviewers are writing reviews in order to give advice as opposed to

demonstrating their deep knowledge about the subject. This is illustrated by the following

comments made by some reviewers in their respective profiles:

“I don't profess to be an expert at anything. In fact, I've always been something of a

professional diletante (sic). I write the reviews to help others make some kind of

judgement about whether they want to spend the money I did on a book.”

“I am not an expert in any of those subject areas, but I don't feel I'm boasting when I say

that my reviews of non-fiction books in these areas are more than just rantings”

“I don't claim to be an expert, but I am familiar with a great deal of music, and hope I

can identify a cd's strong points and identify the type of listener who will enjoy it.”

Thus the repository of reviews is being built by both experts and non-experts and not just

by experts alone.

35

Other findings: We also coded reviewer profiles into the following category that we did not anticipate

from prior theory:

Contribution for self-development: Some individuals indicated motives directed toward developing oneself to explain why

they contributed reviews as shown in the table below:

Type of response / observation Number of

occurrences Percentage of occurrences

Sharpen thoughts by articulation 6 0.60 Improve one’s writing 10 0.99 Chronicle what I have seen, heard or read

9 0.89

Developed expertise on a subject and seeking to build on it

16 1.59

Total 39 3.86 (Note: The total figure above does not equal the sum of occurrences of individual categories since some profiles contain more than one of the above types of responses.)

(Table 13) The above responses indicate that the activity of contributing a review an individual can

improve one’s writing skills, organize and clarify the content of one’s thoughts about the

subject, and provide an avenue for enjoyment. Unlike the other-oriented motives

discussed earlier, the writing of reviews is itself a reward to the contributor. As one of the

reviewers said in his profile:

“I write reviews on Amazon.com's website for two reasons: first, to clarify and organize

my own thoughts and second, to assist others in selecting worthwhile reading material”.

One of the reviewers whom we interviewed elaborated on these sentiments: “I write

reviews to keep myself active and thinking (especially when something pops up related to

36

my teaching class)… My reviewing helps me with the critical thinking required for

school…Reviewing kept me active. It got me back into teaching. It wasn’t a problem

getting into teaching again. The mind was not rusty because of the reviewing I have been

doing. Taking things apart, being critical is important…” So there is more to enjoying the

activity of contributing reviews than just the playful fun involved.

Prior research has viewed self-oriented motives in terms of “intrinsic” motivation, in the

form of the joy one experiences when doing an activity. Wasko and Faraj (2000) found

that 6.5% of participants in the Usenet groups they studied participate because they enjoy

the experience. Similarly, in the email network that Constant et al. (1996) studied,

participants gave a weight of 9.5% to the joy one experiences when solving problems

faced by others, among reasons why they answer queries from others.

However, our findings indicate a more subtle reason for contributing reviews. Much like

the volunteers studied by Clary et al. (1998), individuals might contribute reviews to

“exercise knowledge, skills, and abilities that might otherwise go unpracticed”. In their

study, active volunteers indicated a weight of 19.35% to this “understanding function”

among various motivations for their volunteering activity, and non-volunteers gave a

weight of 19.82% among motivations that would determine their intention to engage in

volunteering. Each review can be an opportunity to develop one’s expertise in a particular

genre of products (corresponding to the sub-categories we discussed earlier), by

collecting one’s thoughts and organizing them in the form of written content. This ties in

with our earlier finding that contributors of reviews have varied levels of expertise.

Summary of findings:

37

We found evidence of well-established drivers that explain motivations for reviewer

contribution: altruism, reciprocity, social affiliation, utilitarian benefits and enjoyment.

We also found evidence of a situational factor that affects the content of reviewer

contribution: the bystander effect. But perhaps the most interesting findings were that

non-experts too contribute to the review repository and that self-improvement motives

can also drive individuals to participate in the repository.

The above findings can be depicted in the following theoretical framework (see figure 1

below):

(Figure 1)

Discussions and Conclusions In trying to understand the reasons why people might contribute to document

repositories, we study a large, public, and well-known repository of product reviews. In

this first-time field exploration of the phenomenon, we use the profiles that the

contributors of the reviews have written about themselves as a source of data to glean

their motivations. Our findings are discussed below:

Self-oriented factors:

• Self-development • Enjoyment

Other-oriented factors

• Altruism • Reciprocity • Social affiliation • Utilitarian benefits

Willingness to contribute

38

First, our above findings suggest that a variety of motivations can drive individuals to

contribute to a repository. This is consistent with the functionalist perspective in

psychology, which holds that people perform a certain activity because it serves one or

more functions (Snyder and Cantor, 1998):

a) Value-expressive: an activity may allow an individual to express his / her personal

values and to his / her concept of self. Contributing to a repository can be a way of

expressing one’s values about altruistic concern for others;

b) Utilitarian: a certain behavior may result in rewards from the person’s external

environment. So a person might contribute to a repository in order to receive monetary or

other rewards;

c) Social adjustive: doing a certain thing may lead an individual to better fit in with his /

her peer group. Contributing to a repository, for example, can offer a way for an

individual to be and socialize with other participants there, and reciprocate the benefits he

receives from them;

d) Knowledge: by engaging in particular task, an individual might have a new learning

experience, and be able to exercise one’s knowledge, skills and other abilities. Writing a

contribution for a knowledge repository can be just one such example of an activity that

allows a person to put down, organize and improve one’s thoughts about a specific topic.

Unlike the perspective of economic theory (in particular of public goods), we see that it is

not just reciprocity or utilitarian motives that drive contribution to repositories. The

results also clarify the nature of the intrinsic motivation that has long been studied in IS

39

literature. In the context of repositories, it springs out of the knowledge motive described

above.

It is likely that a particular feature of a knowledge repository system will be critical in

motivating particular individuals to contribute because that feature serves a function that

is salient for them. Having a range of such features increases the likelihood that at least

one feature will serve a motivational function of different individuals and therefore

encourage them to contribute.

Second, a potential contributor will be more likely to contribute in an area that has little

or no other contributions from others. This observation qualifies the finding of Thorn and

Connolly (1987) whose model of an information repository leads to reduced contribution

as the overall number of participants increases. Our study, instead, finds that contribution

decreases only with increase in number of contributors in a given domain. This also

throws some light on the observation by Butler (2001), who found that the number of

participants can be both positively and negatively related to the amount of

communication activity going on in an online community.

By providing a wide variety of topics in which individuals can contribute to the

repository, and highlighting those that have little or no contributions, a repository system

might be able to induce more contributions. The system might even leverage the

bystander effect by having a “call for contributions” for a period of time in which

potential contributors are not informed how many contributions have already been made

in a particular topic. So long as a potential contributor does not know if many other

individuals have contributed, he / she is more likely to contribute.

40

Third, the presence of a large number of non-experts among the reviewers on Amazon

suggests it does not take too much expertise to sustain a thriving repository. It is likely

that the non-experts are using the forum to organize and archive their own thoughts, and

thus improve their own knowledge. This is attested by the statements listed under the

“contribution for self-development” section above.

In addition, each individual has different requirements for knowledge on a particular

topic. Individuals vary in their absorptive capacity to assimilate knowledge of different

kinds on the same subject (Cohen and Levinthal, 1990; Szulanski, 1996). For example,

for a novel by Stephen King, an individual who is new to the genre would look for

different kinds of information as opposed to someone who is familiar with that line of

books. Such a variation in need for information calls for reviewers with varying levels of

expertise.

This observation among reviewers also goes against the finding from the model

developed by Thorn and Connolly (1987). Their model suggests that information

asymmetry leads to reduced contribution. Sender asymmetry exists when “some

participants have access to better information than others” (p. 518), i.e., when some are

more expert than others. User asymmetry exists when “some stand to benefit more from

information of a given quality than do others” (p. 518), i.e., when participants have

different needs for knowledge. Assuming that only the “hope for reciprocity” is a

consideration for participants, Thorn and Connolly argue that an individual will

contribute only when he expects to receive information of identical quality from others.

At Amazon, differences in expertise among reviewers would lead to sender symmetry.

41

Different shoppers are looking for different information in reviews leading to user

asymmetry. However, our study shows that since participants contribute because of

considerations apart from reciprocity, asymmetry in knowledge levels can actually induce

more contributions to the repository.

To encourage individuals with varying levels of expertise to contribute to the repository,

the repository system might be designed to help them organize their contributions in

terms of topics and keywords. It might also aid potential contributors by providing a

template for making contributions.

Limitations: First, we conducted this exploratory study at only one site, the Amazon.com repository of

product reviews.

Secondly, we did not ask the reviewers specific questions about why they contributed.

Rather, we culled out their motivations from whosoever had mentioned their motivations

in their profiles. As such, we might have missed out some factors that might also be

relevant to the contribution activity.

Future research: The findings in our study call for further theoretical development of the factors causing

individuals to contribute to repositories. They can then be tested on participants of

knowledge repositories in organizational settings. A similar theoretical model might also

be used to study knowledge sharing in other forms of online networks, such as bulletin

boards, and knowledge directories.

42

43

References Aquilino, W. S. (1998) Effects of interview mode on measuring depression in younger

adults. Journal of Official Statistics, 14, 1, 15-29.

Bandura, A. 1977. Social learning theory. Englewood Cliffs, NJ: Prentice-Hall.

Barnett, M. A., Thompson, M. A., and Pfeiffer, J. R. (1985). Perceived competence to

help and the arousal of empathy. Journal of Social Psychology, 125, 5, 679-680.

Batson, C. D. (1998). Altruism and prosocial behavior. In D. T. Gilbert, S. Fiske, and G.

Lindzey (Eds.), The Handbook of Social Psychology (4th ed.), pp. 282-316. New York:

McGraw-Hill.

Batson, C. D., Duncan, B.D., Ackerman, P., Buckley, T., and Birch, K. (1981). Is

empathic emotion a source of altruistic motivation? Journal of Personality and Social

Psychology, 40, 290-302.

Bereiter, C. and Scardamalia, M. (1993). Surpassing ourselves: An enquiry into the

nature and implications of expertise. Open Court Publishing: Peru, IL.

Bergstrom, T., Blume, L., and Varian, H. (1986). On the private provision of public

goods. Journal of Public Economics, 29, 25-49.

Berkowitz, L. (1978). Decreased helpfulness with increased group size through lessening

the effects of the needy individual’s dependency. Journal of Personality, 46, 299-310.

Butler, B. S. (2001) Membership size, communication activity, and sustainability: A

resource-based model of online social structures. Information Systems Research, 12, 4,

346-362.

44

Butler, B., Sproull, L., Kiesler, S., and Kraut, R. (2002). Community effort in online

groups: Who does the work and why? In S. Weisband and L. Atwater (Eds.) Leadership

at a distance. Mahwah, NJ: Lawrence Erlbaum.

Charles, S.K. "Knowledge management lessons from the document trenches," Online,

(26:1), 2002, pp. 22-28.

Clary, E. G. and Orenstein, L. (1991). The amount and effectiveness of help: The

relationship of motives and abilities to helping behavior. Personality and Social

Psychology Bulletin, 17, 58-64.

Clary, E. G., Snyder, M., Ridge, R. D., Copeland, J., Stukas, A. A., Haugen, J., and

Miene, P. (1998). Understanding and assessing the motivations of volunteers: A

functional approach. Journal of Personality and Social Psychology, 74(6), 1516-1530.

Constant, D., Kiesler, S., and Sproull, L. (1994) “What’s Mine Is Ours, or Is It? A Study

of Attitudes About Information Sharing,” Information Systems Research, (5:4), 1, 400-

421.

Cohen, W. M., and Levinthal, D. A. (1990). Absorptive capacity: A new perspective on

learning and innovation. Administrative Science Quarterly, 35, 128-152.

Constant, D., Sproull, L., and Kiesler, S. (1996) The kindness of strangers: The

usefulness of electronic weak ties for technical advice,” Organization Science, 7, 19-135

Davenport, T. H. and Prusak, L. Working Knowledge: How Organizations Manage What

They Know, Harvard Business School Press, Boston, MA, 1998.

Deci, E. L. (1972) Intrinsic motivation, extrinsic reinforcement, and inequity. Journal of

Personality and Social Psychology, 18, 105-115.

Deci, E. L., and Ryan, R. M. (1985). Intrinsic motivation and self-determination in

human behavior. New York: Plenum.

Diamond, W. D., and Kashyap, R. K. (1997). Extending models of prosocial behavior to

explain university alumni contributions. Journal of Applied Social Psychology, 27(10),

915-928.

45

Ericsson, K. A. (1996) The acquisition of expert performance: An introduction to some of

the issues. In K. Anders Ericsson (Ed.) The road to excellence: The acquisition of expert

performance in the arts and sciences, sports and games, Lawrence Erlbaum: Mahwah, NJ.

Festinger, L. A. (1954). A theory of social comparison processes. Human Relations, 7,

117-140.

Harris, M. B. and Huang, L. C. (1973). Competence and helping. Journal of Social

Psychology, 89, 2, 203-210.

Lakhani, K. R., and Wolf, R. G. (2005) Why hackers do what they do: Understanding

motivation and effort in free/open source software projects. In J. Feller, B. Fitzgerald, S.

Hissam, and K. R. Lakhani (Eds.) Perspectives on Free and Open Source Software, MIT

Press (forthcoming).

Latane, B. and Darley, J. (1970)The Unresponsive Bystander: Why Doesn’t He Help?

Prentice Hall, Englewood Cliffs, NJ, 1970.

Nonaka, I. (1994). A dynamic theory of organizational knowledge creation. Organization

Science, 5, 1, 14-35.

Olson, M. (1967). The logic of collective action. Public goods and the theory of groups.

Cambridge, MA: Harvard University Press.

Piliavin, Irving M; Piliavin, Jane A; Rodin, Judith. (1975). Costs, diffusion, and the

stigmatized victim. Journal of Personality and Social Psychology, 32(3), 429-438.

Shamir, B. (1991) “Meaning, self and motivation in organizations,” Organization

Studies, 12, 1991, pp. 405-424.

46

Snyder, M. and Cantor, N. 1998. Understanding personality and social behavior: A

functionalist strategy. In D. T. Gilbert, S. Fiske, and G. Lindzey (Eds.), The Handbook of

Social Psychology (4th ed.), 635-679. New York: McGraw-Hill.

Sproull, L. and Kiesler, S. (1986). Reducing social context cues: Electronic mail in

organizational communication. Management Science, 32, 11, 1492-1512.

Strauss, A. L. and Corbin, J. M. Basics of Qualitative Research: Techniques and

Procedures for Developing Grounded Theory, 2nd ed., Sage Publications, Thousand

Oaks, CA, 1998.

Szulanski, G. (1996). Exploring internal stickiness: Impediments to the transfer of best

practice within the firm. Strategic Management Journal, 17, 27-44.

Takahashi, N. (2000). The emergence of generalized exchange. American Journal of

Sociology, 105, 4, 1105-1134.

Thorn, B. K. and Connolly, T. (1987). Discretionary data bases: A theory and some

experimental findings. Communications Research, 14(5, 512-528.)

Venkatesh, V. (2000) Determinants of perceived ease of use: Integrating control, intrinsic

motivation, and emotion into the Technology Acceptance Model. Information Systems

Research, 11, 4, 342-365.

Wasko, M. M. and Faraj, S. (2000) “’It is What One Does’: Why People Participate and

Help Others in Electronic Communities of Practice,” Journal of Strategic Information

Systems, 9, 2000, pp.155-173

Yamagishi, T., and Cook, K. S. (1993). Generalized exchange and social dilemmas.

Social Psychology Quarterly, 56(4), 235-248.

Yin, R. K. (1994) Case Study Research: Design and Methods, Sage Publications,

Thousand Oaks, CA.

47

Appendix 1:

List of questions forming a template for reviewers to respond to when writing their personal profile:

• Where do you live? • What you do for work? For fun? • What's your all-time favorite movie or book? • What are your passions? Your pet peeves? • If someone handed you an Amazon.com gift certificate, what would you buy? • If you had a month off--no work, no responsibilities--what would you do?

Appendix 2

Mr. D Reviewer Rank: 23

A very good reason to write reviews for Amazon.com is to share ideas with others. As a writer I find I have a much better understanding of what I have read when I write about it. Additionally, when I write about a book I enter into a dialogue, however distant, with the author. Most books I read are excellent, and even the flawed ones have much that is valuable. Writing a book is a long and sometimes lonely process full of uncertainty and doubt. When the book is finally published after several years of work there is a sense of accomplishment immediately followed by a let down. Most books are published and quickly go out of print. Their authors sometimes wonder, was it worth all the time and energy, especially when they get superficial reviews from professionals who often do not have time to actually read the books they review. 'It fell stillborn from the press' is an old book business cliché and lament, and it is also sadly often the truth.

48

Appendix 3:

Interview Questions

Thank you for agreeing to this interview. If a question is not clear to you, please let us know and we will clarify it for you. You are free to decline an answer to any of these questions.

1. Introductory questions (5 minutes): a. Can you tell us something about yourself:

i. your personal background, educational background etc. ii. your interests

iii. your profession b. How long have you been writing reviews? c. Do you participate in the discussion boards at Amazon? d. How much time, on average do you spend writing reviews?

2. Questions to understand details about writing reviews: (20 minutes) a. To what extent do you feel posting a review is like interacting with a

database as opposed to being like an interaction with other people. When you write your reviews, do you have a mental picture of the other people at the Amazon website (customers, other reviewers)?

b. Have you received any feedback on your reviews – from others online? From friends, colleagues and other persons offline?

c. What do you think you get in return for or as a result of writing the reviews? What do you think are the benefits to others from the reviews?

e. Do you ever compare the opinions of other reviewers on the same book you have reviewed? Do you do this before or after you write a review?

f. Do you like the competition that goes with the ranking system? Have you experienced any censorship of reviews by Amazon?

g. Why write reviews for Amazon, and not for other websites such as, say, Barnes and Noble, Books-A-Million, or other book review sites?

h. When writing your reviews, do you expect that the reader will know about some of the concepts you use (e.g., genre of the type of novel being reviewed etc.)?

3. Closing questions: (5 minutes) a. Can you describe one or two things that you like very much about

Amazon (not necessarily connected with writing reviews)? b. Can you describe one or two things that you dislike very much about

Amazon? c. May we contact you again by email in connection with this project if we

have any questions? Thank you so much for your time and sharing your thoughts on the reviews you write. Have a great day!

49

Appendix 4:

Coding rules Altruism

• Explicit statement of wanting to help others without expecting any benefit for oneself when doing the activity.

• Expressing sentiments of doing one’s duty. • Indications of wanting to attain a goal larger than writing a review for others to

read. Utilitarian motives:

• Mention of actual or expected receipt of any tangible benefit (money, products etc.) from others in exchange for contributing reviews.

• Mention of actual or expected receipt of any non-tangible benefit (e.g., reputation) from others in exchange for contributing reviews.

Reciprocity

• Stating that one is writing reviews as some form of repayment to others for the benefit one has gained by reading the reviews contributed by others.

• Mention of having benefited by reading reviews written by others. Social affiliation

• Saying how the writing of reviews enabled the person to connect to the community of reviewers or content producers.

Enjoyment motive

• Mention of how the writing of reviews is a joyful or playful experience. • Statement saying that review writing provides some form of pastime or relaxation.

Choice of content contributed

• Statement indicating why the individual chose to contribute reviews for certain products, areas or categories.

• Statement indicating why the individual chose NOT to contribute reviews for certain products, areas or categories.

50

Expertise of contributor

• Mention of commonly acknowledged indicators of expertise such as educational qualifications, experience, certifications, etc., with respect to the domain in which he contributes reviews.

• Self-declaration by individual indicating a degree of expertise in any area concerning the topic(s) on which he contributes reviews.

Self-development

• Statement indicating that the writing of reviews organizes, develops or improves one’s knowledge about a subject.

51