Pretest as Pretest
I was on the whole quite pleased with how the survey went. People seemed genuinely interested in both the survey and the project in general, and asked (via both e-mail and the MU*) frequent questions. I e-mailed back the requested clarifications and on occasion found myself drawn into interesting discussions of the terminology I chose. It was these commentaries from my participants, in fact, that showed me my terminology wasn't clear enough, and was perhaps splitting the categories a little too finely. For my general design I would eliminate the category since, as several participants pointed out, being truly OOC is extremely unlikely to occur. Most users don't wish to create a separate character with their name, and to log on as them every time the player goes OOC. Also, on some MU*s with a heavy emphasis on RP, multiple characters are actively discouraged.
I also found there were several versions of the category of masked. There were those that used masked as a means to exercise and enjoy some aspect of their personality they would feel uncomfortable expressing iRL, which was the category I initially thought all masked incidents would fall into. Closely related to this category (due to the necessity of having a character to be masked by) were those who used masked for humorous intent, as in the example given at the beginning of this paper. However, there were also those that were masked simply by nature of having coded in various 'attributes' or 'properties' for their character, such as the computer replacing 'barks' or 'murrs' for every instance of 'says.' These individual players were OOC, but the computer treated them as if they were still IC. Finally there were those who were extremely good at 'getting IC' and who, when faced with being OOC, still had the mindset of the character fresh in their mind, i.e. using the pronoun 'she' when the player is male and the character is female. These respondents chose to maintain the character's mindset so as to be able to go quickly and easily back IC. As a consequence, on the next survey I would use more clarification and simplification in describing my questionnaire terms.
I also had one incident that, while it worked out well in the pretest, warned me of a potential problem in the general design. I spoke to two individuals whom I knew personally were in a relationship. One requested a survey, but since the other individual did not I did not send her a survey. I was, therefore, both surprised and pleased to receive one from her. However, had I not personally known both the individuals in question I could easily have been fooled into accepting two surveys from one individual. Thus in the general design I would add a paragraph to the opening letter stating clearly that the survey was based on PPS sampling. I must admit, it would be a bit of a mental strain to turn away a request for a survey, however.
Finally, I had one incident during the pretest that made me question my choice of a self-administered survey. I accidentally e-mailed an older, incomplete version of the questionnaire to Maki, one of my volunteers, and discovered this quite late in the project. I immediately contacted Maki and asked him if he'd be willing to let me administer the questionnaire to him on the MU*. He agreed and 'joined' me there, and I gave him the survey questions one at a time, giving him time to answer between questions. I noticed a few things fairly quickly. First, I was constantly worried about whether or not to respond. I did not want to sound cold and distant, but I also worried that a constant litany of 'okay' and 'uh-huh' would get irritating after a while. Also, I wanted my respondent to feel comfortable, and not to feel intimidated. Secondly, it was quite nice to be able to have immediate feedback -- if Maki didn't understand a question he asked for clarification and I gave it. This meant both that there were no misunderstandings due to my wording, and that no questions were missed. Also, if I wasn't sure of Maki's response I too could get immediate feedback, from him. Thirdly, I found myself wording the questions more clearly as I was typing them in. I found this initially odd, since I'd thought I'd written them quite carefully, until I realized that by the time I got to surveying Maki I knew where the weaknesses in my survey were. I was trying to compensate for them, to make the survey both more understandable and more enjoyable for my volunteer.
When we were done I noted the time (it took only 55 minutes) and curiously asked Maki which he'd preferred -- the self-administered questionnaire or the one I'd given. He answered that while he'd liked the survey, he'd found the administered questionnaire fun. It could be simply that Maki likes talking with me, but it's also possible the immediate feedback of an administered questionnaire would provide superior response rates and quality of information than a self-administered, e-mailed survey. If this is the case then many more interviewers would be needed.
Pretest as General Design
I sent out 14 surveys, and 12 were returned, giving me an 80% response rate. Of the twelve respondents, three were female, eight male, and one self-identified as a male-to-female transsexual. The median age was 26, with a mean of 28 and two-thirds years. Two of the respondents were Canadian; all the others were from the United States. The median time spent MU*ing was between 3 to 3-1/2 years, while the mean was (very roughly) approximately four and a third years. On the average, they spend 3-1/2 hours per on-line session, for about 20+ hours per week, and on the average slightly more than half that time is spent socializing. However, individual sessions range from 1-2 hours, to 6-12 hours per session, with a reported median of 2-4 hours. The reported time spent on-line per week varied from less than 10, to 30-40, with a median of approximately 20. Also, as Alois noted, most of these reported times are probably 'guesstimates.'
Of all the respondents, two (both young males) joined MU*s solely for the socialization alone, and one of them has since stopped MU*ing due to increased social contacts in his life. A combination of RP and socialization was mentioned by all other respondents as the reason for joining the MU*s, and of those ten, two also mentioned professional contacts/networking as part of their reasons.
For the purpose of this paper I cannot analyze all the questions in my survey. Instead I am focusing on a handful, in order to thoroughly examine them. Questions #13: 'Have you ever played a character that was essentially 'you' in-game?' #17: 'How did you handle personal issues (e.g. extreme shyness, being blind, or having unpopular personal beliefs) while playing essentially yourself?' #37: 'Are you ever masked? If so, when do you feel is the right time to be masked?' and #49: 'Do you feel more or less confident in face to face encounters than you do on MU*s?' will demonstrate whether personality exploration is occurring, how, what sort, and why. Question #14: 'Have you ever played a character with your own name?' will show that a separation between self and character is being maintained. An understanding (or lack thereof) of the distinction between OOC and IC is shown by the response to #38: 'While gaming, how do you show you are Off-Stage?' Questions #22 and 23: 'Have you ever played a character that was not your gender? Did you feel you succeeded in 'passing' as someone of the other gender?' will give a quick view of one of the more common types of identity play, and #29: 'Have you ever played a character that was not your race?' will indicate whether or not RL hierarchies are being maintained on the MU*. Questions #45-48, 52, and 53 are used to give a rough gestalt of the general feeling the respondents had, in regards to the time they spend on the MU*.
When asked if they'd played a variant on themselves as a character, all the respondents but three answered in the affirmative. All three were male (at 25, 26, and 39 years) and of those three, one had been playing for less than a year, while the other two had been playing for more than 10 years. The only thing they appeared to have in common was the length of their on-line sessions, which ran about 2 hours on the average. All three did not feel they had worked out personal issues via their characters, but this was not a characteristic unique to only them, as 82% answered this question similarly. One of them denied ever being masked, an attitude shared by only one other (female) respondent. Furthermore, these three had wildly varying answers in regards to whether they were more confident expressing themselves IC on occasion: one answered yes, but the other two felt the question was inapplicable to them. Indeed, 42% of my respondents felt the same amount of confidence expressing themselves whether they were IC or OOC.
None of the respondents, however, had played a character with their own name, although two males reported using variants on their own names for characters. All of the reported using some textual aid to denote separation between OOC and IC. All of my respondents but one (an older female) reported having attempted gender-swapping, although of those 11 two felt they had not succeeded in passing. However, I would have to regard this statistic with caution. While I was genuinely surprised to discover the RL gender of one of my respondents, in at least one case where a respondent reported passing I noted privately that I had easily been able to tell the player's gender. 'Passing' is, I think, too general a term to use, as players on MU*s will have widely varying abilities to notice cross-gender play. In regards to attempting cross-race play, 42% reported attempting it. However in each of these cases the respondents noted that this was attempted because the race was 'right' for the character or the character's background, and the players wished to play the characters as convincingly as possible. None of the players felt they'd 'passed' convincingly, not having enough personal knowledge to do so, or noting that the race in question was supernatural, and thus there was no way to tell if the portrayal was convincing or not.
In regards to the ten respondents that reported playing masked, there were broad overlaps of reasons given (i.e. to maintain IC mindset and to use it in a humorous fashion). Of the 30% that reported doing so for humorous purposes one individual also mentioned maintaining the IC mindset; all were male but their ages and gaming experience varied wildly. Others (78%) used this reason also, while one individual mentioned that since he was logged in as the character of course he'd appear masked while speaking OOC. Finally, 50% of the 10 individuals who masked mentioned being masked as a means to better express themselves, either due to player shyness or a feeling that the character's attributes were superior to their own for expressing themselves. Of this 50% none were female, and age and gaming experience showed no correlations.
Finally, when asked about whether the time spent MU*ing had been beneficial, 58% reported (with varying degrees of enthusiasm) that the time they'd spent gaming had helped and enriched their lives. The reasons given varied from finding more friends and learning more about life, to the benefits of being able to try on someone else's point of view. Of the others, three respondents (25%) felt the question was inapplicable to them (the two respondents that reported never being masked fell in this category). The final 2 respondents (17%) believed their lives were better, but weren't sure this was due only to time spent on the MU*s.
It would appear initially there were no real correlations between being masked and any other of my categories. However, after staring at the data for far too long, it occurred to me that I should separate out those that felt they could use their characters to express themselves more fully from those that did not, and see if there were any relationships to be discovered in that group alone. Checking for those that did not use their characters to express themselves at all meant that I separated out the oldest female and the longest MU*-gaming male.
It is clear there is no correlation between personal confidence (either IC or OOC) and the player using the character to more fully express themselves. However, the oldest female, while having MU*'d only 7 months, reports having gamed for many years, and does not use her MU* character for self-expression. The oldest males do, but they also have very little MU* experience, despite many years spent RPing. It is the most experienced male MU*er that also reports not using his character whatsoever for personal expression. Thus I find a tentative link between using the character for personal expression, and either age or length of time spend MU*ing. This would indicate that the individuals that have the least 'need' to explore their characters (due to extensive life/MU*ing experience) are also the ones least likely to do so.
Keeping this tentative link in mind, I feel I can now guardedly verify my hypothesis. If asked 'Do MU* participants use their pseudonymous virtual personas while OOC to express aspects of their personalities they do not feel comfortable expressing iRL?' I would reply yes, with the proviso that this occurs only if the player is still experimenting with MU*ing or has not reached the age of about 35-40 years. Of course, the number of respondents is extremely small, so any conclusions must be tentative, and further study would be required for verification.
There is, of course, a problem in comparing my research to the literature review, since there was no direct match-up between them and my study. Also, I was not studying the nature of pseudonymous deception, and thus I did not have any real means of comparison with Donath's (1999) study. However, some of the general conclusions arrived at in the various studies covered by the literature closely match those I found. For example, like Geierman (1996), Nakamura (1995), and Rodino (1996), I found that use of female-reading pseudonyms encouraged more (and more aggressive) responses than male-reading pseudonyms. However, unlike Geierman's (1996) and Nakamura's (1995) studies, these responses did not map precisely to RL hierarchies. They were more in the nature of gender as performance, as Rodino (1996) suggested, i.e. a male player with a female character strongly pursuing a male character played by a male player. In fact, identity as performance, and to deliberately disrupt RL hierarchies, was a strong theme running through all my respondents' replies, just as Rodino (1996), Bruckman (1993) and MacKinnon (1997) suggested. Nor did I find the strong insistence on stereotypical racial portrayals which Nakamura (1995) found on LambdaMOO. However, this may be due to my respondents: firstly, I was questioning them concerning their portrayal of themselves, and secondly the MU* they're used to playing on insists on non-human characters. In such a case RL hierarchies are a bit tougher to implement. For a single example, what species do you map as Caucasian?
My research results also agreed with Jaffe et al (1995), Lee (1997), and Chester & Gwynne's (1998) results in regards to pseudonymous and deliberate identity creation being ultimately beneficial to communication. While it is true not all of my respondents mentioned being masked as increasing their sense of freedom of expression, several did state it so, and none found it detrimental. Finally, Meyer & Thomas' (1990) portrayal of the net Underground as postmodernist is reflected in my research also. The deliberately created, symbolic identities of my respondents defiantly (re)define themselves within a mutually supportive fringe subculture, with little or no attempt to fit within or be defined by public or mass culture.
Last Updated: Tue, March 28, 2000