The Wicked Wop Posted August 9, 2004 Share Posted August 9, 2004 Poll 6 wasn't going the way I was hoping, so I'll pose the question more this way. Say you were the victim of a phishing scam, hacking, or virus that caused you to lose $1000 through an E-Commerce transaction.....if so would you give up or severly limit your online buying, banking habits? Link to comment Share on other sites More sharing options...
Kilmer17 Posted August 9, 2004 Share Posted August 9, 2004 Dont like the results so you change the question? How Democratic of you. Link to comment Share on other sites More sharing options...
The Wicked Wop Posted August 9, 2004 Author Share Posted August 9, 2004 Thats a good idea kilmer I'll try and remember that next time. I think adding a dollar figure to the equation helps establish some kind of threashold people are willing to except, where the previous poll was a little open ended. Link to comment Share on other sites More sharing options...
mjah Posted August 10, 2004 Share Posted August 10, 2004 Um, you aren't actually using these results for an actual real-life thesis, are you? Do you plan on making generalizations about the online community as a whole from these polls, or are you going to (correctly) limit your conclusions to self-selected non-poll-averse volunteers from profesisonal sports message boards? Link to comment Share on other sites More sharing options...
The Wicked Wop Posted August 10, 2004 Author Share Posted August 10, 2004 There will be no generalizations about the online community.....the intended purpose of these polls is to see how (generally proficient and adept) computer users view security and privacy matters. The intended purpose is show how great a threat computer viruses and hacking type of attacks can be. Link to comment Share on other sites More sharing options...
The Wicked Wop Posted August 11, 2004 Author Share Posted August 11, 2004 Need a few more votes fellas........if you haven't done so already please take a second a chime in Link to comment Share on other sites More sharing options...
Kilmer17 Posted August 11, 2004 Share Posted August 11, 2004 You know I was just busting your chops right? Link to comment Share on other sites More sharing options...
Mr. S Posted August 12, 2004 Share Posted August 12, 2004 i think it was good you had an option this time for 'limited or restricted use' or something. Gives a better idea. Link to comment Share on other sites More sharing options...
The Wicked Wop Posted August 12, 2004 Author Share Posted August 12, 2004 Of course kilmer....wouldn't expect anything else:D Link to comment Share on other sites More sharing options...
mjah Posted August 13, 2004 Share Posted August 13, 2004 There will be no generalizations about the online community.....the intended purpose of these polls is to see how (generally proficient and adept) computer users view security and privacy matters. So you're not generalizing to people online. Rather, you're generalizing to reasonably competent computer users everywhere. By asking a small group of football forum participants. How will you demonstrate that these participants represent the larger group you're trying to talk about? And it looks like you have 23 votes as I type this. Out of curiosity, have you taken any statistics classes? Link to comment Share on other sites More sharing options...
The Wicked Wop Posted August 13, 2004 Author Share Posted August 13, 2004 Well Mjah, yes I have taken 2 statistics classes, and there is precisely 24 votes. The simple fact is what the hell would you suggest as far as getting peoples opinions on security matters.....its there choice to vote and express their opinions. True this is a football forum, but it hardly means that the participants are biased one way or another to the questions.....maybe if you posted more than 67 times you would realize this. The fact is asking people on this boards opinions to survey questions is probably a better measure then asking a bunch of old women down by the grocery store. The fact is this is a thesis for a CIS class, that class requires that we do polls and surveys to help back up our claims. True I will give you that 24 responses to this one question does proves a trememdous amount, but at the same time it shows a variety of opionion and perspectives on the issue. It also requries that we document research and facts by others, as well as show our own results. By the way have you ever seen an actual survey.....probably not, I've actually worked on about 5 surveys that I have produced in my profession, and have another survey in my paper as well, you know what the whopping total of respondents was in those factual surveys (last survey I just completed - 66 [with 15 coming from the same company]; the one cited - 89) not huge numbers considering how many companies there are in the world. Link to comment Share on other sites More sharing options...
mjah Posted August 13, 2004 Share Posted August 13, 2004 what the hell would you suggest as far as getting peoples opinions on security matters..... No disagreement there. It's hard. I would suggest that posting a question for a self-selecting respondent group is probably going to disproportionately attract those "generally proficient and adept" computer users who are most concerned about e-commerce fraud, thereby skewing your results. Truly unbiased survey design can be very tricky. Personally, I'd work on contacting people randomly, instead of counting on them to come to me via the forum. At the very least, I'd post the same exact questions on multiple forums that deal with different interests. True this is a football forum, but it hardly means that the participants are biased one way or another to the questions.....maybe if you posted more than 67 times you would realize this. It doesn't matter if I have 67 posts or 67,000 posts. The fact is, it's not enough to say "People around here are hardly biased." You have to be able to establish concretely that you've taken reasonable steps to ensure that they aren't biased, relative to the larger population you're interested in. Otherwise you can't generalize. And the nonrandom sample of respondents automatically makes this very difficult to do. Here's another question: what percentage of your respondents have been women? I'll bet it's pretty small. It may be significantly smaller than the female proportion of "generally proficient and adept" computer users to whom you want to generalize the results. Or is gender not a significant driver of the answers respondents give? In that case, better find some results that support that claim. The fact is asking people on this boards opinions to survey questions is probably a better measure then asking a bunch of old women down by the grocery store. Is that your only other option? By the way have you ever seen an actual survey.....probably not, Surveys? Seen them, read them, designed them, distributed them, analyzed them, critiqued them, published them. If I can find an online link to my MIT thesis work, I'll be sure to send it to you. I've actually worked on about 5 surveys that I have produced in my profession, and have another survey in my paper as well, you know what the whopping total of respondents was in those factual surveys (last survey I just completed - 66 [with 15 coming from the same company]; the one cited - 89) not huge numbers considering how many companies there are in the world. I have no idea what the general setup was with your previous surveys, because you didn't really give much information about them. But as you know, for statistical purposes it is often true that: 66 >> 24 89 >> 24 I've done surveys with small sample sizes too, for two reasons: 1. For one reason or another, it was nearly impossible to find many willing respondents among the large target population. 2. The target population was really small, so even 100% participation would have resulted in a small number. In case (1), it's very tricky and often impossible to generalize the results. In case (2), it's not much of a problem to generalize 30 respondents to a total population of 35. It seems evident to me that surveying competent computer users meets neither of these criteria for using a small sample size. Look, I understand that you're trying to balance the ideal of a perfect survey against the reality that you need to actually administer and analyze the thing without a team of 15 people working for six months. And I realize that nobody is forcing you to apply every major principle of survey design to your CIS class thesis. I'm just cautioning against claiming that the results apply with any accuracy to a group that is probably about five or six orders of magnitude larger than the respondent group. Also, I thought this survey was in support of a thesis for a post-undergrad degree, not for a single class. Link to comment Share on other sites More sharing options...
The Wicked Wop Posted August 13, 2004 Author Share Posted August 13, 2004 I don't disagree with some of your assessments, but the fact is this is a small concise thesis. This is not a thesis where you have to get 1000 repsondents to prove a point and quantify or document that they are unbiased. There are also interviews invovled with IT professionals, personal experiments and experience, and survey questions to individuals on how they percieve the threats. I have read and seen numerous Thesis, acutally its termed a DSP at my school, that include small scale questionares to friends and fellow employees that would be far more biased then asking computer users on a message board for their opinion. Whether that opinion backs up my claim that computer viruses and hacking attacks are a threat is up to them. The fact is, the particpants on this board appear to come from most walks of life, with verying interests, experiences, and professions. The other constricting factor, is that this is just an 11 week class, with only 2 weeks devoted to this type of research....not trying to make excuses, just trying to be realistic. The fact is I will know more about how the professor views my work by this weekend, if he says its weak, then maybe I'll try and administer some advice from you Link to comment Share on other sites More sharing options...
chomerics Posted August 13, 2004 Share Posted August 13, 2004 Originally posted by mjah No disagreement there. It's hard. I would suggest that posting a question for a self-selecting respondent group is probably going to disproportionately attract those "generally proficient and adept" computer users who are most concerned about e-commerce fraud, thereby skewing your results. Truly unbiased survey design can be very tricky. Personally, I'd work on contacting people randomly, instead of counting on them to come to me via the forum. At the very least, I'd post the same exact questions on multiple forums that deal with different interests. It doesn't matter if I have 67 posts or 67,000 posts. The fact is, it's not enough to say "People around here are hardly biased." You have to be able to establish concretely that you've taken reasonable steps to ensure that they aren't biased, relative to the larger population you're interested in. Otherwise you can't generalize. And the nonrandom sample of respondents automatically makes this very difficult to do. Here's another question: what percentage of your respondents have been women? I'll bet it's pretty small. It may be significantly smaller than the female proportion of "generally proficient and adept" computer users to whom you want to generalize the results. Or is gender not a significant driver of the answers respondents give? In that case, better find some results that support that claim. Is that your only other option? Surveys? Seen them, read them, designed them, distributed them, analyzed them, critiqued them, published them. If I can find an online link to my MIT thesis work, I'll be sure to send it to you. I have no idea what the general setup was with your previous surveys, because you didn't really give much information about them. But as you know, for statistical purposes it is often true that: 66 >> 24 89 >> 24 I've done surveys with small sample sizes too, for two reasons: 1. For one reason or another, it was nearly impossible to find many willing respondents among the large target population. 2. The target population was really small, so even 100% participation would have resulted in a small number. In case (1), it's very tricky and often impossible to generalize the results. In case (2), it's not much of a problem to generalize 30 respondents to a total population of 35. It seems evident to me that surveying competent computer users meets neither of these criteria for using a small sample size. Look, I understand that you're trying to balance the ideal of a perfect survey against the reality that you need to actually administer and analyze the thing without a team of 15 people working for six months. And I realize that nobody is forcing you to apply every major principle of survey design to your CIS class thesis. I'm just cautioning against claiming that the results apply with any accuracy to a group that is probably about five or six orders of magnitude larger than the respondent group. Also, I thought this survey was in support of a thesis for a post-undergrad degree, not for a single class. Damn Mjah, what do you do for a living? Link to comment Share on other sites More sharing options...
mjah Posted August 13, 2004 Share Posted August 13, 2004 WW -- Fair enough. I wasn't trying to insult your choices or anything. Sounds like you have it under control, and as you said -- you certainly know more about what's expected than I do. I thought you were doing these surveys as part of a major, major thesis project (like a degree capstone thing or master's/PhD), which is why I jumped on it. I'm gonna tick off an answer to your survey, too. chomerics -- I'm an operations consultant for mid-sized capital firms and fairly large companies. Apparently they don't keep me busy enough, because I end up going off on rants on sports forums. :dunce: Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.