Re: Sample vs Population Question


Tim McCr8 wrote:
> > while the sample might be "statistically viable" as well as the response base
> > this research methodology is *self-selective.* This means that you
> > might hit the target and you might not -
>
> Thanks. I guess this is where I get hung up. Because I understand that
> my responders are self-selected. But aren't all responders--even to the
> most perefctly drawn sample--self-selected? I though that's why
> non-response is such a big issue.

True, but some self-selection may be more selected than others?
(Appologies to G. Orwell) It might be a good idea to see if you can
find any literature on response rates for professional groups similar
to your own. IIRC ,you have a 20% response rate that sounds okay but
I am not used to this type of survey. I did work on a couple of studies
where our overall reponse rate was a bit over 80% but this was
probably an unusual population and we did three waves of mailing
/callings.


> I understand why this is an issue in questions of public policy, public
> health or even political attitudes. Our survey was concerned with none
> of these big issues. All we wanted to do was represent our findings as
> a snapshot of professionals working across the industry. Almost by
> definition we "hit the target."

Well not really. Clearly you have a subset of the professional group
but I suppose that there may be some reason why one distinct subgroup
is more likely to respond. Probably, it is not that likely or at least
that important but it always needs to be kept in mind.

>
> I get hamstrung because of my training as a direct marketer. We thrive
> on self-selction because the economics don't work otherwise. As I said
> in the first post, as a marketer I'd jump for joy at 5%
> response--especially if the 5% looked (demographically) exactly like my
> target population.

See above :) My boss was getting a bit worried because we were
slipping below 90%! This was not a marketing exercise however.
However if you are saying that the respondents look demographically
just like the population why do you think that you are getting a sig
interaction between age and salary reporting? It may well be there
but if it was a bit unexpected I'd be inclined to think that it could
be equally a true difference or a change difference.

>That this same thinking doesn't work for surveys
> (and again, I am talking about this type of limited survey to a known,
> finite audience that we can confirm is reflecive--demographically--of
> the entire population) makes my head spin.

You can generalize - that is one presumes the point of doing the
survey. I, anyway, was just suggesting that saying that to call those
who did not respond to the salary question a different group based on a
sig t-test absent some other reasons, either data based or theoretical
based may be capitalizing on change.
John Kane, Kingston ON Canada

.

Relevant Pages

  • Re: Sample vs Population Question
    ... > might hit the target and you might not - ... Our survey was concerned with none ... target population. ...
    (sci.stat.edu)
  • Re: One Inch Punch
    ... >>If you hit something and tighten all the muscles at that moment.. ... acting like the mucles tightening. ... its not adding anymore power to the collision. ... target, it goes too slow to keep up a shockwave effect. ...
    (rec.martial-arts)
  • Re: DoomRL style targting
    ... attackers the chance to fire past blocking creatures to hit a further ... can continue onwards to the desired target. ... intentionally attempting to fire past a creature. ...
    (rec.games.roguelike.development)
  • Re: Fallout 3: Now Im looking forward to it.
    ... melee is governed by other parts of the SPECIAL attributes. ... First, you have rifle and are 100 meters from your target, crouched down ... you hit, and pull the trigger. ... Also, a lot of these martial artists guys would probably argue about your statemant about hth nature, im not one of them so i wont;) ...
    (comp.sys.ibm.pc.games.rpg)
  • Re: One Inch Punch
    ... >punch with a round punch.. ... >If you hit something and tighten all the muscles at that moment.. ... the target AND push on or follows thru. ... only we have different reason for tightening and stopping at the ...
    (rec.martial-arts)