Thanks for everyone who participated in the Tech Tuesday survey.  I learned a lot.  First off, not surprisingly, my audience overall is fairly tech savvy with 2/3rds knowing at least how to code up HTML and over ½ having done some programming.

Second, I was happy to find that the level of difficulty was about right and that if anything I should maybe add some harder “bonus” material.

Third, post length seems spot on with nearly 80% liking it as it is and roughly 10% each asking for shorter and for longer, so I will stick with roughly the existing length.

Fourth, it appears that all three proposed topics are of essentially equal interest (more on that claim below).  The only thing that is clear is that all three of these are in fact good topics for my audience as only 10% wanted  to hear about something different altogether.

Now I am going to restart Tech Tuesday next year with a sequence on Web Technologies.  It does seem like a timely topic that would be useful for everyone including Congress to understand.  Not that I expect to have members of Congress in the audience, but once the series is up, I will include a link to it whenever I write to one of my representatives.

So I claimed above that there really isn’t a meaningful difference between the votes for each of the topics.  How do I figure that?  Let’s leave out the “other” category and focus on the three remaining ones.  There were 78 respondents for those categories, which conveniently happens to be a multiple of 3.  A perfectly even distribution would be 26 for each of the three possibilities which would clearly provide no signal about preference.  

So how much of a preference signal am I getting from the actual voting outcome of 24, 26, 28?  One way to think about this is as follows.  If people had no preference and were simply throwing darts, how often would the outcome show a difference of 4 or more between two topics?  Now there is a mathematical way of figuring that out, but that’s a bit involved.  Instead, I wrote a short piece of code to run a Monte Carlo simulation (you can look at the HTML for this post and it includes the Javascript code).

Thanks for everyone who participated in the Tech Tuesday survey.  I learned a lot.  First off, not surprisingly, my audience overall is fairly tech savvy with 2/3rds knowing at least how to code up HTML and over ½ having done some programming.

Second, I was happy to find that the level of difficulty was about right and that if anything I should maybe add some harder “bonus” material.

Third, post length seems spot on with nearly 80% liking it as it is and roughly 10% each asking for shorter and for longer, so I will stick with roughly the existing length.

Fourth, it appears that all three proposed topics are of essentially equal interest (more on that claim below).  The only thing that is clear is that all three of these are in fact good topics for my audience as only 10% wanted  to hear about something different altogether.

Now I am going to restart Tech Tuesday next year with a sequence on Web Technologies.  It does seem like a timely topic that would be useful for everyone including Congress to understand.  Not that I expect to have members of Congress in the audience, but once the series is up, I will include a link to it whenever I write to one of my representatives.

So I claimed above that there really isn’t a meaningful difference between the votes for each of the topics.  How do I figure that?  Let’s leave out the “other” category and focus on the three remaining ones.  There were 78 respondents for those categories, which conveniently happens to be a multiple of 3.  A perfectly even distribution would be 26 for each of the three possibilities which would clearly provide no signal about preference.  

So how much of a preference signal am I getting from the actual voting outcome of 24, 26, 28?  One way to think about this is as follows.  If people had no preference and were simply throwing darts, how often would the outcome show a difference of 4 or more between two topics?  Now there is a mathematical way of figuring that out, but that’s a bit involved.  Instead, I wrote a short piece of code to run a Monte Carlo simulation (you can look at the HTML for this post and it includes the Javascript code).