Louie+Pausch+--+Research+Paper

=Social Norms in Second Life: An Empirical Approach=

The empirical section of our research netted results that were both predictable and surprising. We decided to test three dependent variables—whether or not a Second Life user had copied a non-copy object, whether or not a Second Life user would copy a non-copy object if they had the technology, and whether or not a Second Life user would report another user to Linden authorities if they had knowledge that another user was copying non-copy objects.

The research we conducted used six independent variables to test the dependent variables. Three of the independent variables we used were qualitative, while the remaining three were dummy variables (yes or no questions). The qualitative variables were gender, age, and months playing Second Life. The dummy variables were whether or not the user lived in the United States, whether or not the user was a student, and whether or not the user used p2p software (such as Limewire, Kazaa, etc).



We selected these variables because we felt that they would give us the best snapshot possible of why users held their beliefs and took certain actions pertaining to copyright. Indeed, the results we came across we feel gave us insights as to how and why users feel the way they do with respect to the texts assigned and the framework developed within this class. As the class discovered, the theme for the semester is abstraction and disconnection. By understanding generational factors (age), experience with technology and Second Life (play time and p2p usage), location, and contextual factors (whether or not the user is a student), we are able to discern how individuals have been able to come together to form a community or public within Second Life in order to practice cultural transmission. Hopefully the equations we have unearthed will help describe how people use the objects they create in Second Life.

Let us first observe means for the data. The number of surveys done given by our group was 130. The average age within these 130 was 28. 53% of the respondents were male (one of our respondents gave the gender of “shemale” and that value was left null). The average respondent had used Second Life for 11 months. 63% of our respondents were from the United States. 39% were students, and 42% had used p2p software. 1.5% had copied non-copy objects within Second Life (a total of 2 observations). 22% said they would copy non-copy objects if they had the software. 39% of observations said they would report avatars if they knew they were copying non-copy objects.

The equations we developed for this project were as follows: These equations help us to view the correlation between the variables and the dependent variables.

The standard errors in these equations were as follows: would copy had a standard error of 0.40, have copied had a standard error of 0.12, and would report had a standard error of 0.49. The standard errors for the individual coefficients were all generally low. There was almost no problem within our data when it came to error. As far as t and F statistics are concerned, the values were remarkably low and nearly all of our data was significant. The high volume of interviews which Team America managed to give managed to drive error to a very low value. This does nothing but add credibility to our equations.

The dependent variable ‘have copied’ is perhaps the most difficult of the dependent variables to grasp. The regression equation we managed to write for it is almost useless, because out of our 130 respondents, only two reported ever having successfully copied a non-copy object. Both were interviewed by American Blackburn. The fact that so few respondents reported successfully copying non-copy objects, it makes more sense to search for facts within their two answers rather than to use the equation to determine which independent variables correlate to whether or not a user has copied a non-copy object. First, the age differential between the two individuals who reported having copied non-copy objects was 15 years and fell right in the center of our range (our range was 15-60, and the two individuals ages were 20 and 35). Both of these individuals were male. Both of these individuals were in the top 20% in experience with Second Life. Besides the gender variable, all of the other gender variables netted one ‘yes’ answer and another ‘no’ answer.

The question being faced now becomes “what does this mean?” Descriptive statistics means absolutely nothing if we cannot draw useful inferences from them. Unfortunately, our powers of interpretation are severely hindered with such a low number of positive responses. Nonetheless, we can still drive at things which may identify why they have copied non-copy objects. First, they are both males who have been playing Second Life for a long time. This may be explained because only those who have used Second Life for a long time have experienced the “copy-bot” incident. The copy-bot incident gave nearly all Second Life users the ability to copy non-copy objects. The scandal was quickly wrapped up. However, users who wanted to copy non-copy objects had the ability to. It is interesting to note that only two of the respondents reported having copied non-copy objects. However, instead of making the inference that 1.5% of Second Life users have copied non-copy objects, we should probably instead infer that closer to 15% have copied non-copy objects. This is because when the sample is constricted to users who have used Second Life for two years or more (when the Copy Bot incident occurred), the percentage of users who have copied non-copy objects increases to 15%. As will be shown soon, the fact that they were both males has no relevance because of the other two equations.



Due to the extremely low number of positive responses to the have copied question, we must rely on the ‘would copy’ data to draw inferences from the framework described in our class. Luckily, this data is much more significant and many more observations about correlation can be made from our empirical data. First, there were 29 positive responses (people reporting that they would indeed copy non-copy objects if they had the ability). This translates into a 22% rate of positive response. This higher rate allows us to draw many more inferences from our data.

Looking at the data to discover which variables correlate most directly to whether or not a user would copy a non-copy object nets both intuitive and non-intuitive results. First, the independent variable most closely correlated to whether or not a user would copy a non-copy object is whether or not a user is based in the United States. That independent variable is negatively correlated at a rate of 23%. This means that if a user is not in the United States, they are 23% more likely to respond yes to the question about whether or not they would copy a non-copy object. The second most correlated independent variable is the use of p2p software. This variable is positively correlated at a rate of 16% (meaning that a user would be 16% more likely to respond ‘yes’ if they used p2p software). Gender and whether or not a user is a student have a slight positive correlation of less than 5% (gender netted a 4.2% correlation for females and the question of studentship netted a 1.5% correlation). Age and time using Second Life netted a correlation of 0.1%, which is very insignificant.

From this data, a plethora of information can be drawn that can be put into the framework developed in our class. At first glance, it may be surprising that the value most highly correlated with the ‘would copy’ dependent variable is whether or not the user lives in the United States. However, upon closer examination, this becomes much clearer. Lawrence Lessig, in his book [|Free Culture], describes a multitude of attacks made by the US government upon the public domain experiment (Lessig 158-162). As Lessig rightfully describes, this is more than an issue of copyright, is a matter of corporatism. By destroying our means of cultural transmission via much stronger copyright laws, Americans have become significantly more abstracted from their existence. The most egregious example described by Lessig in his book is the Digital Millennium Copyright Act, which passed unanimously in the US Senate and without much of a fight in the US House. However, [|when this law was proposed in the rest of the world], it met with much harsher resistance.

The two countries outside of the United States which have attempted to pass DMCA-like laws have met with mixed success. The first example is Canada. [|The Canadian government attempted to pass a law which was even more restrictive than the American DMCA.] The Canadian law would have [|included absolutely no fair use protection or backup provision]. [|Boingboing] called the Canadian bill the “worst copyright law in the developed world.” When it was introduced, however, it met fierce resistance. [|Dr. Michael Geist], a law professor from the University of Ottawa, started an online movement which eventually received national attention and derailed the bill. He wrote extensively on the project, including a 30 day blog-series called “30 days of DRM,” created a [|youtube video] which enjoyed thousands of views, and a [|facebook group] which grew to more than 10,000 members after just one week of existence. The bill was pulled by the Canadian government less than a week after it Geist began making news in the [|Canadian press] and in [|American technology blogs.]



[|Switzerland’s government had more success in passing a DMCA-style bill]. However, this is only because the Swiss government acted in secret to pass this bill. According to news articles and blog posts, the Swiss DMCA passed with very little public information about it being released. In Switzerland, however, [|50,000 signatures trigger a referendum for any law passed by the Congress]. [|This campaign eventually materialized, but with little success]. From the information gathered, it appeared that the leader of the campaign was slightly political naïve (and also worked for a company that produced DRM). In the future, it is highly likely that Switzerland will have a referendum on their DMCA. At any rate, [|the Swiss DMCA is much less draconian than the US law] and the Canadian bill, because it includes a provision legalizing the distribution (via p2p and other software) of non-copyrighted material.



It can now been seen that the world as a whole has not become subjects to the permission culture at work within the United States. It is therefore not surprising that the location of a user correlates directly to whether or not these users would copy an object designated non-copy.



The connection between the use of p2p software and whether or not a Second Life user would copy non-copy objects is quite a bit more intuitive. Individuals who use p2p software, for legal or illegal purposes, have an appreciation for cheap and easy cultural transmission. This is the main prerequisite for Dewey’s idea of the “Great Community” (Dewey 158). These individuals who value cultural transmission via the use of p2p networks obviously also have an interest in cultural transmission in all ways including copying non-copy objects in Second Life. The logical conclusion of this train of thought is that individuals who use p2p software would be more likely to be wiling to copy non-copy objects in Second Life.

The slight correlation between whether or not a Second Life user would copy a non-copy object and whether or not a user is a student and whether or not a user is female are a little bit more difficult to describe. The gender correlation may be able to be discarded because there is no consistency between this result and gender results in the other equations. As has already been noted, both of the individuals who had successfully copied non-copy objects were males. As shall be shown, females are statistically more likely to report individuals who have copied non-copy objects. Therefore, since all three of these reports contradict, the gender variable can be thrown out.

Although the correlation between studentship and the social norms relating to copyright in Second Life is weaker than some other independent variables, it still is consistent throughout the entire regression analysis. This can be explained because college campuses often are the forefront of the copyright wars being raged in the United States, especially with the RIAA.

Lawrence Lessig includes an anecdote in [|Free Culture] about a student who felt the wrath of the Recording Industry Association of American. Jesse Jordan, the student in Lessig’s anecdote, ends up having to pay his entire savings of $12,000 to the RIAA in order to settle the lawsuit against him in 2003 (Lessig 48-52).

Jordan was one of the first students to have the RIAA sue them. The RIAA has [|continually threatened students with lawsuits], sending out several thousand “pre-litigation settlement letters” to students. These events usually occur in waves and are sent to colleges all over the country. The efforts of the RIAA have been well documented and are well known, but the important part of the campaign with respect to our work is that the RIAA is specifically targeting college students. A spokeswoman for the RIAA [|legitimizes their strategy]:

"It should be clarified that our college campaign is in addition to the lawsuits we file against individuals using commercial ISPs to illegally download and distribute music. Second, college students have reached a stage in life when their music habits are crystallized, and their appreciation for intellectual property has not yet reached its full development. These two points coupled together present challenges to those who would like to be compensated for their creative works. Understanding the value of intellectual property is important to the future job market for many of these students--industries that rely on copyright protection employ more than 11 million workers nationwide and continue to grow."

Obviously, therefore, students would react differently to copyright within Second Life. However, unfortunately for the RIAA, students appear to be reacting negatively to their campaign. Their scare tactics have either failed or backfired. If our research is correct, students are still more likely to be open to copying non-copy objects.



The final dependent variable we tested was whether or not a Second Life user would report another user whom they knew was copying non-copy objects. Unlike the other two dependent variables, this variable tests for a viewpoint of copyright which more than likely opposes Free Culture or the “Great Community” of Dewey, and which agrees with hyperindividualism in Friedman’s framework (Dewey 184).

The results were again fairly predictable. Unlike the ‘would copy’ variable, none of the correlation coefficients reported were greater than 10%. The only variable which achieved a 10% correlation was whether or not a user used p2p software. I count as significant any variable which has a correlation of greater than 1%. Therefore, time using Second Life and age can be disregarded completely in this regression equation. Age had a correlation coefficient of 0.00%, and time on Second Life had a correlation coefficient of 0.2%. Recall that we threw out gender because it had no consistency across equations, and could not be generalized overall. Citizens of the United States were 4.1% more likely to report copiers of non-copy objects than the average user, and students were 5.7% less likely to report copiers of non-copy objects than the average user.

This fits beautifully into the framework developed in the previous equation. Studentship, use of p2p software, and whether or not a Second Life user determines the ‘would report’ data just as it did the ‘would copy’ data, except for in the opposite direction. Instead of use of p2p software and studentship correlating positively to the regression equation, it correlates negatively. Likewise, the location of a Second Life user correlates positively to the dependent variable ‘would report’ whereas it correlated negatively to ‘would copy.’

The end of our research leaves us with an empirically sound framework. Three of our six independent variables significantly affected users’ views on copyright—whether or not users are from the United States, whether or not users use p2p software, and whether or not users are students. The other three independent variables—age, gender, and time using Second Life—have no significant bearing on what users’ views of copyrights are. Studentship and the use of p2p software correlate positively to more progressive views on copyright, namely that these users would be more willing to copy non-copy objects if they possessed the technology and also they would be less willing to report users who were known to be copying non-copy objects. Location within the United States correlates negatively to progressive copyright views. In other words, if a user is from the United States, they would be less willing to copy non-copy objects if they had the technology, and they would be more willing to report user whom were known to be copying non-copy objects.

This framework wraps around the theory developed about Second Life by Smiley Clawtooth and Christian Fall. The fact that the theoretical framework we have been discovering throughout the semester has been proven empirically lends credibility to what we have been doing all semester within Second Life. I am glad that this empirical research has been done, because statistics lead directly to theoretical credibility.