>Sorry long meeting.
>
>Seriously, diagram your model and the relationships of the different
>predictor variables. Is it essentially a regression model, or are you
>planning on introducing an age related measurement.
>
>A simple chart with age on the x axis may not be for the best.
>
>Your theoretical model always determines your analysis. Or in CF related
>terms you always should plan out your app before doing any coding.
>
>larry
Larry,
I was just kidding about leaving me hanging.
I conducted this survey because I thought that there expected value of the
age on the CF-Community list would be around 39-44 based on the number of
anti-Reagan posts. I was wrong in many ways. First, I published the wrong
expected value. The expected value I originally published was 44. That
was a top line estimate but I knew it would never go that high. I expected
the mode to be around 40 because as a tech list, I wanted to revise the
numbers to the downside. I have my expected value and the actual value. I
actually never done a regression model, but I can look at the numbers and
see that I just have to shift everything to the left on the X-axis to get a
relative fit. I had anticipated a positively skewed distribution. Which
is what I got. I doubt any of my statistical analysis will wow anyone.
I was also wrong about the avg age of the members of this list. None the
less, I have re-learned a valuable lesson my last political science
professor taught me. Almost anything in popular politics can be explained
by party politics.
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings]
