What do you mean with having a CI of 11? 11 what?
It's been a long time since I *somehow* chose a sample size of 75 from a homogenous, hard to reach population of 3,156.
Why would I do this, I have no idea. I am digging through my notes from a year ago....
Now trying to pound a round peg into a square hole.
If I have 95% confidence, and my confidence interval is 11, what does that mean in layman's terms? This is a very homogenous, elite population.
Thank you!!!!
What do you mean with having a CI of 11? 11 what?
Hi - thanks for taking me and my problem on :-)
I don't understand confidence interval at all - I used a calculator on line and plugged in my population and sample size, and the 95% confidence rate and it told me I have a confidence interval of 11.3?
What's your estimate then? On which scale? You should read in a basic statistics book about confindence intervals...
Sorry, I am an EdD student with limited stats knowledge. So maybe I should start over! I have a population of 3,400. I did random sampling, but perhaps it was also convenience sampling, because I did not want to send the survey out to 1,000 people. I wanted to send it to 72. I realize my results are not able to be generalized across the whole group. I was pretty much doing a pilot study on a new topic. Of those 72 surveyed, 12 said they would interview with me as part of my qualitative research. I did a mixed methods study. I also pulled in others to interview.
How do I explain my rationale when defending? I know I brought sampling error in because I chose not to go out to 1,000 people or whatever it would have taken to hit the "right" number. My survey contained proprietary information and I did not want it going to the masses. This is an elite, hard to access population - if they knew the survey was sent to the masses, they likely would have immediately hit "delete."
Any thoughts? Does my rationale make sense? The qualitative part of my dissertation is 60 pages long and explores a ton of data as informed by the surveys then built opne through elite interviews. The quantitative part was important, of course, but not the moneymaker for the dissertation.
I see that almost 100 people have read this thread and only Kristian was kind enough to respond :-) Thanks for trying! I look forward to hearing your thoughts on my approach. And yes, maybe I am now trying to pound a round peg into a square hole. ;-) But I need to make it work - doesn't have to be perfect, just defendable is my mantra right now!
I guess I know that online calculator, and if I have guessed correctly, that website explains the CI too. But if not, a 95% CI of 11.3 for a prevalence means that if you had a 20% prevalence, there is 95% possibility that the true prevalence in the main population is some value within your calculated value (20%) plus and minus 11.3% (then your range of confidence interval becomes 9.7% to 31.3%). For example they survey people's opinion on a president candidate and say Obama has 55% +/- 5% vote. This +-5% means the 95% CI.
In that very website or other online "power calculators", calculate your test power. If that was above 80%, you are very good to go! Otherwise, if it was below 80%, note that we are limited by our time and budget. So you can always say I sacrificed my test power for putting an end to my distress (i.e. my convenience)! (That will reduce the value of the statistics of your research, but it is completely understandable that sometimes it is IMPOSSIBLE to collect more data).How do I explain my rationale when defending? I know I brought sampling error in because I chose not to go out to 1,000 people or whatever it would have taken to hit the "right" number. My survey contained proprietary information and I did not want it going to the masses. This is an elite, hard to access population - if they knew the survey was sent to the masses, they likely would have immediately hit "delete."
Yes definitely your rationale makes sense, except that you did not want to select 72. You did want to select much more but all available were these 72 (because of the limitations you stated above plus because many of them did not respond). Therefore, in your thesis mention that you reached for more than 72 (for example 300 etc.) but could get only 72. This is completely reasonable then.Any thoughts? Does my rationale make sense? The qualitative part of my dissertation is 60 pages long and explores a ton of data as informed by the surveys then built opne through elite interviews. The quantitative part was important, of course, but not the moneymaker for the dissertation.
No worries. Nothing wrong with your thesis. Good luck.I see that almost 100 people have read this thread and only Kristian was kind enough to respond :-) Thanks for trying! I look forward to hearing your thoughts on my approach. And yes, maybe I am now trying to pound a round peg into a square hole. ;-) But I need to make it work - doesn't have to be perfect, just defendable is my mantra right now!
--------------------------
yeah it made me smile. Those are elite people who in my own experience know almost nothing about management or their specific fields, and yet try to control their organizations!
I agree with ted. Also make your defense as short as possible. Arrogant committee members don't like listening to lectures longer than 5 minutes. Up to 10 minutes is Ok, but longer than that is a real threat.
"victor is the reviewer from hell" -Jake
"victor is a machine! a publication machine!" -Vinux
I don't think I saw whether the CI is around a mean, proportion, rate, or other ...
in any event, you can compute power using a calculator -- note there is some leeway in the inputs, and you can see just how sensitive your inputs are to help build your case, I think
for one thing, many studies turn out to be "pilot" studies due to financial restraints, but the pilot at least sets the stage for a future larger study, the sample size of which is computed using info from the pilot study. I feel confident your advisors will see your level of activity and thought, and I can't imagine a rationally-acting professor giving too much trouble over this, especially considering that you've said the quantitative part is minor compared to the rest
Also ... "elite" ... I'm intruiged :-) maybe after you pass your defense you could come back and let us know a little more about your study, who are these "elite" :-)
The mathematical explanation of a statistical procedure is really just pseudo-code, which we can make operational by translating it into real computer code. --B. Klemens
Thanks to both of you, my old friend Victor and new friend Ted! this is so much clearer now! I really appreciate your time. I thought I had a great argument and my committee was picking it apart, I know they are just getting me ready - I am working on the verbiage for defense (March 13th, starting to get that panicky, nauseous, excited feeling!)
The elite population is college presidents - nothing sexy like movie stars or Nobel Peace Prize winners, LOL!!!! But they are so hard to get to, surrounded by their "gatekeepers" ugh! And uber busy! Many don't even read their own email...I wonder if someone wipes their chin at lunch if their soup dribbles? I know you're both laughing right now. :-)
Thanks again to both of you!
Ha!
Sounding interesting, best of luck!
Take them snacks, and refreshments, and if you can get a room where you can control the lighting, dim the lights, defenses right after lunch (when they're more sleepy) sometimes works well ;-) and try to find the most comfy chairs in your department for them to sit in (I was able to find 4 "captains" chairs that weren't being used for my committee members)
The mathematical explanation of a statistical procedure is really just pseudo-code, which we can make operational by translating it into real computer code. --B. Klemens
I love it, comfy chairs and snacks, yay! Great plan! :-)
I just worked through that section of the dissertation and feeling good about getting up and defending it, thanks again!
Tweet |