We consulted with a friend who actually knows some statistics to arrive at some of these values, and while they aren't necessarily as large as we'd like, our friend is now thinking of getting a Masters degree in Applied Statistics at Cornell. Maybe he'll be able to lie better then.
The primary number that we hoped to discover from the survey was total readership. We know the number of copies of each issue downloaded from three main sites, America Online, sumex-aim, and GEnie. Three issues carried the survey form, so all of our percentages had to be divided by three to get an accurate number (or so our friend said, maybe he wasn't lying hard enough). Responses from America Online and sumex-aim accounted for approximately 5% of the number of issues downloaded from those sites, whereas GEnie had a lower percentage response of about 2%. Since it's impossible to send us email from GEnie directly, a 2% response rate is excellent. A friend at American Demographics Magazine said that a 4% return rate on those little white cards in magazines that are pre-paid and easy to fill out is good, so our 5% is even better, considering that our survey was longer than a little card. Applying that 5% to the 127 people who responded from Usenet, it seems that 2,540 people get TidBITS from comp.sys.mac.digest. Next applying the 5% to the total 229 respondents, we come up with 4,580 readers overall. That's forgetting the number of people who get TidBITS from someone else, and to judge from the surveys, few of those people responded. Of the 229 people, 34% said they redistributed TidBITS and the numbers of additional readers they gave add up to 911. So now we're up to 5,491, which is pretty good, considering all the defunking that you have to do to read TidBITS these days. The final possible addition is that if 34% of respondents (or 76 people) distributed 911 copies (about 11 copies per person), then if the 34% of 4,580 people (assuming that people who receive TidBITS from someone else don't redistribute again) or 1,557 people each distribute about 11 copies as well, that will be a whopping 17,129 copies to add to our previous subtotal of 5,491, to give a grand total of 22,620. Still with me?
Even if our method of calculating redistribution is wrong (which it probably is, since the people who responded to the survey are the most likely to be the people who redistribute, thus artificially inflating the redistribution percentage), we're still happy with between 5,500 and 20,000 readers as of January, 1991. That number rises constantly, to judge from the amount of email we get asking for information about TidBITS. The real trick is going to be switching to an implicitly-tagged text format, because then we'll legitimately be able to count the entire readership of comp.sys.mac.digest as TidBITS readers, and it will be fun to add 37,000 more readers just like that. We also hope to set up a LISTSERV when we move to text-only, which will increase the number of readers who were otherwise unable to download TidBITS.
Of those 229 respondents, we received email from 209 and snail mail from 20. Interestingly enough, almost half of the snail mail we got was from countries other than the US. TidBITS is read in 18 countries, including: Australia, Canada, Denmark, England, Finland, France, Germany, Israel, Italy, Japan, the Netherlands, New Zealand, Portugal, Scotland, Sweden, Switzerland, the USA, and Wales. If you want to be picky, Scotland and Wales are part of Britain, which would lower the number to 16. In theory we could figure out the number of states in the US that have TidBITS readers, but that would be a bit more difficult and fairly meaningless anyway.
Most people got TidBITS from Usenet, sumex-aim, America Online, and GEnie, in that order, but a number of BBS's had multiple respondents, including the Memory Alpha BBS in Ithaca, Tom's BBS in Boston, the AMUG BBS in Atlanta, and the Twilite Clone BBS, whose location I don't know. A number of other BBS's had a single respondent. A few respondents get TidBITS from some of our less popular redistribution sites, like LISTSERV@RICEVM1.BITNET and MACSERVE@PUCC.BITNET and the British National Public Domain Software Archive. Still fewer respondents report that they get it from a friend or on a local network, which seems to imply that people who read TidBITS twice removed from the nets tend not to deal with the nets at all. Even though CompuServe can send mail to the Internet, we only received one response from CompuServe. What's with those people?
Where do TidBITS readers come from, other than the woodwork? We didn't ask this question, but it was often obvious from the email address or a signature. TidBITS readers come primarily from higher education and big business, not surprisingly, since those two sectors are best connected. A lot of people find TidBITS on small local bulletin boards as well, which means that areas like K-12 (not usually known as lower education, for some reason :-)) schools, dealers, and small businesses read TidBITS. Most of the major universities showed up, as did large companies like Apple, Claris, IBM (yep, even IBM), Toshiba, Sony, Motorola, and government bodies like NASA and various branches of the military.
What are TidBITS readers like? They are knowledgable (heck, to figure out the nets you've got to be bright) and interested. A good percentage of them are also packrats, since 61% of respondents use the TidBITS Archive. I'll bet that number would be higher if the archive were faster and smaller, but it still holds a lot of information and is bound to be large and slow to a certain extent no matter what. On average, people said that they rated 8 points out of 10 with regard to their knowledge of the Macintosh, but only 5 points out of 10 with regard to HyperCard. As of the time of the survey, only about 66% of respondents had HyperCard 2.0 - that number is surely higher by now, although we found a decent bit of HyperCard animosity reflected in other survey answers.
On the normal 1 to 10 scale, people only rated themselves 2 in terms of how often they used the contact information and 2 in terms of how often they looked up the references in other magazines. However, many people said something to the effect of, "I don't use it much, but it's very handy when I do need it. Don't discontinue contact information or references!" OK, we won't. Approximately half of the respondents indicated that they might be interested in writing articles at some point, but that conflicts a bit with the number of articles we've received from people.
And what of everyone's favorite question, "What is your favorite color?" It provoked many strange and uncountable answers, so we wimped out and asked Double Helix to count each entry in which a color name appeared. So if the answer was "Blue, no, red, auuugh!" (a common answer), both blue and red would be counted. If someone said "Not blue!" I munged the word so that it wouldn't count. See below for the results.