or Connect
EpicSki › The Barking Bear Forums › On the Snow (Skiing Forums) › General Skiing Discussion › Any intermediate skiers out there?
New Posts  All Forums:Forum Nav:

Any intermediate skiers out there? - Page 2

post #31 of 46
Nolo - I'm going to try to respond to both Oboe's and your most recent post in one reply since (a) They is some overlap, and (b) I'm heading out of town later today for the entire weekend to attend a wedding. So, unfortunately I am short on time, yet want to respond. Let me give it a try.

Originally posted by oboe:
PhysicsMan, it hasn't been my read that she's looking for self selection. Whee do you get that?
First, Nolo stated at least one well defined, fairly strict criterion in her post of June 20, 2002 01:58 PM:

Let's say I am looking for a gallery of pictures of people who would fall within the definition of ATS 5-7.

Then, she went on to say (in her post of June 20, 2002 03:34 PM) that she was considering self-selection:

I need the people who apply for the study to do a good job selecting themselves.

To me her first statement strongly suggests that inadvertent admission of people outside of the range ATS 5-7 would dilute and/or skew the results of her study. As an example (rigged to be obvious), suppose ATS 1-4 respond well to teaching method A and very poorly to B & C, 5-7 to teaching method B (but not A and C), and 7-9 to method C (but not A & B).

Then, if sufficient numbers of people that are not 5-7's are inadvertently included in her study (by self selection bias or for any other reason), her results will be less conclusive, along the lines of, "Well, pretty much any of the teaching methods worked with intermediates, at least to some degree."

Nolo then went on to ask:

"How would passing the job of sample selection to the EpicSki community de-skew the sample? "

Perhaps some confusion arose because in my previous post, I used the term "de-skew" as a catch-all for both sample homogeneity
(ie, spread, consistency, standard deviation) and sample bias (ie, unintentional offsets in the mean value of variables due to phenomena like "grade inflation").

It seems well established from the comments in this thread that different groups of skiers have very different ideas of who would get classified as 5-7's. SS's seem to often engage in "grade inflation" and pass along their ideas of levels to their customers. PSIA uses another set of definitions, people on PowderMag another, and "strict-grading" people like TomB, Oboe and me use yet another set. Thus, the possibility of achieving any sort of sample homogeniety based on people catagorizing themselves seems highly unlikely.

This is why I suggested not using initial self-selection, but rather employing a set of impartial, knowledgable assistants as the gatekeepers for admission to your study. Not only would they not be biased by personally knowing the people selected for the survey, but they are also *much* more likely to have a consistent definition of ATS 5-7. SS instructors (around the country) that you could enlist to hand out invitation / instruction forms would be an obvious choice. BTW, you should keep track of how many forms are handed out and compare this figure to how many people respond to them, since this will give you an upper bound on any non-randomness creeping in at this stage of the sample generation process.

If using instructor colleagues is not an option open to you for some reason, I would argue that almost any set of people acting as gatekeepers would do a better job than self-selection as long as: (a) you provided them with some consistent and simple instructions; (b) the people they approached were strangers to them; etc. etc. The instructions that you provide to the "lay" gatekeepers could even be the set of photos of intermediates that you were originally considering as guidance in the self-selection process. Similarly, since the design of your survey was discussed here on Epic, I would say that anyone who is an EpicSki member should be automatically rejected from your sample (but they would be fine as gatekeepers - unless you need to blind them to other aspects of your survey).

Obviously, even if you pass along the initial gatekeeping job to SS colleagues or EpicSki denizens, there will enevitibly be some remaining degree of self selection for reasons such as differences in refusal to participate, drop-outs,
etc. as a function of skiing level, gender, and a myriad of variables, but the whole point of a good research design is to attack the big problems first, then work your way down to the smaller ones.

"People are more or less intermediate, but the idea of finding a uniform sample is pretty far-fetched."

Problems like this have arisen in every survey since the dawn of time. One doesn't just suddenly realize that this is a problem and then lay down and die. Rather, you do the best you can during the experiment design phase to get a narrow-dispersion sample consistent with your goals, but within mundane constraints like time & money. First you define what a truly optimal sample should be for your particular purpose. Next, you come up with an overall plan for sampling which will deal with many issues including specific inclusion criteria, insuring various types of randomness (eg, gender, geographic, etc.), etc. Finally, you figure out how to best implement that plan (eg, issues like where do you get your assistants, how do you train them, etc.)

Obviously, if your study is to be statistically acceptable, even before you do any of the above, you should have thought about *VERY* clearly stated hypotheses (and nulls) that you want your survey to distinguish between. As a very concrete (but hypothetical) example, do you want your experiment to tell you whether the rate of picking teaching method "A" in question 3 is significantly different from (a) the rate of picking teaching method "B", (b) the rate of picking *any* other teaching method, (c) an answer picked randomly, etc. Consideration of your specific and precisely constructed hypotheses will tell you (in part) how to construct your optimal sample, how large it has to be for specified levels of significance, etc..

"Do we focus our efforts on selecting the right selectors?"

You should not only focus your efforts on this particular issue, but also on the myriad of other issues that arise in scientifically valid research designs, some of which I mentioned above (eg, for starters, construction of good hypotheses, rigorous estimation of needed sample size, etc.).

I don't do survey design for a living so I can't help you with specific literature references. However, it is a huge field with much literature available to help non-experts who have to occasionally design and execute a survey as part of their "main job". I would suggest hitting your local university library and pulling a couple of books to get some ideas. FYI, often this subject is shelved with subjects like epidemiology and statistics.

Gotta run. Lets talk again next week. Please feel free to PM or email me (physicsman000@yahoo.com) if you want to continue this discussion off Epic.

Tom / PM

PS (in edit) - Just so I have some idea of the limitations on the scope of your project because of time and money consideratons, are you presently thinking in terms of a sample size of 10, 100, 1000, or more? Obviously, experimental designs appropriate to large samples are often not suitable for small samples. Also, is this project likely to be published in a specialist peer-reviewed journal (or constitute part of an academic thesis), published in a skiing journal with little statistical review, or not published at all (eg, a preliminary/feasibility study, just for your own information, etc.). Clearly, the ultimate goal impacts the amount of care that will go into the design.

[ June 21, 2002, 01:19 PM: Message edited by: PhysicsMan ]
post #32 of 46
Nolo - one of our ski resorts has started doing a video snow report. Lots of nice shots of skiers - will this help?

Look Here

Hope that helps [img]smile.gif[/img]
post #33 of 46
Thread Starter 
Thanks, PM. I will drop you a line next week.


Thanks for a great lead!

[ June 24, 2002, 04:20 PM: Message edited by: nolo ]
post #34 of 46
Ah, well, that's NOT me, and I don't ski like that, so I guess I just need skierdude to get in contact and I'll show you what "Intermediate" really means!!!

post #35 of 46
Thread Starter 

You will want to leave your "banana turns" at home.

post #36 of 46
I realise there has to be a joke in her but...
How do you make a banana turn?

I mean, what is a banana turn?

post #37 of 46
Thread Starter 
I had just read your joke about bananas. The banana turn is reportedly an exercise used in PMTS training.

Sorry, it was a lame attempt. I will leave humor to you who know how, and slink back to my work.
post #38 of 46
The Banana Turn...

[ June 21, 2002, 03:43 PM: Message edited by: ryan ]
post #39 of 46
I now think I know what you're referring to. (ingore the PM)

Should we be expecting SCSA to go blind? (If you are implying what I think you are implying)

post #40 of 46
The Emperor's New Skis! I don't see no picture, dude!

BTW, some of us intemediates don't have balls. Its a girl thang...... [img]tongue.gif[/img]
post #41 of 46
I can't see your photos, and I've just been asked by our lovely friends at Microsoft to sign up to .NET.

Oh, and I will post once I have some photos of me skiing.

post #42 of 46
At a guess, it's the security setting on your .NET webspace, cause the error I'm getting is...

"You are not authorized to view this page
You might not have permission to view this directory or page using the credentials you supplied.


If you believe you should be able to view this directory or page, please try to contact the Web site by using any e-mail address or phone number that may be listed on the www.msnusers.com home page.

You can click Search to look for information on the Internet.

HTTP Error 403 - Forbidden
Internet Explorer "

post #43 of 46
Here are my lame intermediate turns on 9990 and Apex Ridge at The Canyons for your enjoyment and dissection:

Please let me know if you can't see them, but I think I've finally figured this posting-pics business out!
post #44 of 46
Thread Starter 

The pictures are perfect. Thank you for going to the trouble to post them.

May I copy them?
post #45 of 46
Originally posted by nolo:
May I copy them?
You may. However, if they end up in a best-selling coffee-table book, my pitbull lawyer (oboe) may have something to say about it.
post #46 of 46
Thread Starter 
Thanks, James. It's for purely academic purposes of which EVEN Oboe would approve.

You're not lame at all. I'd say you have a bright future ahead of you, judging from how you're going for it in the top photo...

You're coming out for the Bears Thing in Utah?
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: General Skiing Discussion
EpicSki › The Barking Bear Forums › On the Snow (Skiing Forums) › General Skiing Discussion › Any intermediate skiers out there?