New Study Confirms:
SHORT SKI METHOD STILL SUPERIOR TO STANDARD TEACHING METHOD FOR ADULT BEGINNERS
The controlled study presented here, brings new evidence about the superiority of the Short-Ski Method or Graduated Length Method (GLM) over the standard American Teaching System (ATS). The study was performed under the following scientific aspects:
· A control group and an experimental group were utilized.
· All test criteria were objectively measurable.
· Measurements were taken both at the beginning of the ski course and at the end of the ski course. Each skier's relative improvement within the combined groups was recorded. In this way the difference between the two methods could be isolated from the difference between the skiers' abilities .
· The statistical significance level was established.
The short-ski method or Graduated Length Method (GLM) was developed by Clif Taylor in Vermont in the 1960's. The strategy is to start on very short skis (65-80cm) and use the parallel turn from the very beginning. The skier progresses to gradually longer skis as he improves.
The method employs very little technique explanation and is characterized by a great deal of action and self discovery. GLM was very popular in the United States especially at Killington, Aspen Highlands and anywhere, that Taylor went to promote his method. Short Ski methods were also developed in Austria by Karl Koller, in France (Ski Evolutif) by Pierre Gruneberg & Robert Blanc and in Germany (Ansteigenden Ski Längen--ASL) by Martin Puchtler. Since the introduction of Kneissl's 63 cm Big Foot several authors have proposed beginner teaching methods using it. Kurt Schock was probably the first. One major difference between the Big Foot Method and GLM is the method used to familiarize the learner to longer skis. For the transition phase the Big Foot Method uses a Big Foot on one leg and a full length ski on the other.
This study was carried out using the method proposed by Puchtler 9 . The Skischule Nordbayern (Since Puchtler's death in 1995 under the direction of Karlheinz Fischer) still uses this method successfully. I learned to ski at this ski school and am currently a short-ski ski instructor there. Although many ski schools offer Big Foot Courses, there are currently only 5 Ski Schools in the world offering a true graduated length program to beginners: Skischule Nordbayern in Bischofsgrün, Skischule Montana in Grafrath, Ecole de Ski International in Alp d'Huez, Ecole de Ski Francaise and Arc Aventures in Les Arcs.
There have been 6 controlled studies 1,2,3,4,5,12,13 which compare the short-ski method with the standard ski teaching method. All studies except one indicate, that the short-ski method is superior to the standard method. None of these six studies presented statistical significance. Sturm1 and Pfeiffer2/Fry4 attempted to subjectively establish innate (before the study) similarities and differences between short-ski participants and control group participants. None of the studies included tests both before and after the instruction period.
To justly compare the two methods, measurements were made both before and after the instruction period. Skills, that will be measured afterwards, are not yet present before ski instruction takes place. For this reason two different sets of skills must be measured. The Pre-Test measures coordination and courage. The Post-Test, a composite test, measures speed and control. Because these two tests measure two different things on two different scales, it is necessary to find a common unit of measure. The ranks of the the raw scores on these tests serve this purpose.
In addition to the pre test and final test, observations were made to estimate success on the first day.
A pre-instruction test must measure skills, which are possible after only 90 minutes on skis. This test must measure coordination as well as courage. Gliding straight down the fall line on a very slight incline, stepping back and forth over a rope or marked line (as described in the DSLV SkiLehrplan 7 p. 109 ) is an objective test for measuring these two facors.
Success on first day
Perhaps the most important measure of a ski teaching method is the taste of success on the first day. Four objective measurements for the first day are:
1. Distance skied in meters
2. Vertical meters skied
3. Number of direction changes executed
4. Number of falls per 10 direction changes
Post-Test or Final Test
Two elements in skiing summarize the goals of most skiers: speed and control. These two elements are necessary for most skiers to have fun. The following events provide an objective means of testing these elements:
1. Slalom race on a green slope. (measuring speed and control)
2. Free Run Race on a green slope. (measuring speed)
3. Number of direction changes
over a short stretch of blue slope. (measuring
A direction change is defined as both skis turning across the fall line and changing edges.
Each participant's ranks in these three events are added together to yield the Sum of Post-Ranks (column Sum of Post Ranks in Table 2). The lower scores are better.
Technique is not an end in itself, but rather a means to reach the two more important goals: speed and control. Modern Ski Teaching methodology even discourages teaching a normed ski technique. Technique is also a very subjective matter and so a technique test would lend itself to manipulation or claims of manipulation. Speed and control, on the other hand are simple to measure . For these reasons we choose not to use a subjective judgment of technique as part of the Post-Test.
NARRATIVE OF THE STUDY
The AFRC Ski School at Garmisch is an ideal environment for such a study because of its consistent high volume of full-week guests and the high degree of professionalism demonstrated by its instructors. The week I was there in February 1997, there were 350 skiweekers. The Ski School is a Professional Ski Instructors of America (PSIA) Ski School and hosts PSIA Certification Examinations each year. The Ski School uses the standard American Teaching System as presented in the PSIA literature 10 . The Ski School Director, Shred (Leigh Plowman, a PSIA examiner) is very open and helpful and was interested in this study. The instructors are well trained, having a minimum of 12 days internal training. By mid February even the rookie instructors have 50 days of practical experience. At the end of each Skiweek the ski school hosts a final Slalom race. This race provided part of the composite Post-Test as described below.
On registration day first-time skiers were asked if they wanted to participate in the study by learning to ski with the short-ski method. 12 out of the first 15 candidates chose to do so. The remaining 17 adult first-timers were assigned to the control group. The two groups were very similar. The ages ranged from 17 to 44. There were no apparent disabilities. None of the the participants were particularly athletic. Most had office jobs.
All first-timers were fitted out normally in the AFRC ski rental shop with ski boots and standard length beginner skis (160 cm). On the first morning the short-ski group as well as the control group were taught using their standard length beginner skis and the standard teaching method.
The instructors for the control group were BSam (Brian Samway) and Squirrel (Carl Swanback). BSam is a rookie ski instructor with a level 1 PSIA Certification. At the time of the study he had about 50 days of practical experience. Squirrel has many years of experience as an instructor and as a Ski School director. He has a level 2 PSIA Certification. The short-ski group was taught by myself and I have a Level 1 German Ski Instructors Association (DSLV) Certification and many years of ski teaching experience.
After 90 minutes of instruction the Pre-Test was performed. The short ski group was still on 160 cm skis. A 25 meter rope was stretched out along the fall line on a slight incline. The participants were instructed to ski along the rope stepping back and forth across the rope as many times as possible. A step is counted, if it is made first with one foot then the other and if each ski lifts off the ground. The number of clean steps across the rope provided the raw score for the Pre-Test. (column Pre-Test Score in Table 2). (column Pre-Test Score in Table 2). The control group was slightly better than the short-ski group at this stage. See the collation of the two groups in the left half of illustration 2.
At lunch time short skis (70-90cm) were fitted to the short-ski-participants' boots. They were taught using the principles set out in Puchtler 9 and Lawler 8. BSam and Squirrel continued to instruct the control group using ATS. The measurements for the first afternoon were made by two observers and their results are shown in the following table:
First Day Results:
on Average per Person:
Vertical Meters skied
Number of Direction changes
Falls per 10 direction changes
Table 1 Success on first afternoon
The short-ski participants were given longer skis each day. On the fifth and final day the skis' lengths were 140-160cm. (predominantly 150cm).
The skis used for the short-ski group were:
70cm-120cm Sawed-off skis out of the trash heap.
130cm Kästle Firn Extreme and Hagan Tour Extreme
135cm-140cm Various Learning skis bought used from American Skirental shops-
150cm-160cm Youth skis with Adult Bindings from the AFRC Ski Rental
On day two the control group was reshuffled. The better skiers went to Squirrel's group and the weaker skiers went to Bsam's group. One short-ski participant quit the ski course as did several from the control group. On the third day Squirrel's group had 6 and Bsam's group 8 as opposed to 11 in the short-ski group. This finer differentiation of abilities as well as the smaller group size were an advantage factor for the control group.
From the original groups 9 out of 12 of the short-ski group and 6 out of 17 of the control group were present for the entire Post-Test on the fifth and final day. The Post-Test was carried out as described above. The finishing places in the three events were summed to yield the Post-Test Results (column labeled Sum of Post-Ranks in Table 2). The graphical representation of those results is shown in the right half of illustration 2. The Pre-Test Order reflects the ranks of the steps taken over the rope in the Pre-Test described above. The Post-Test order represents the ranks of the sums of the results of all three events described above. Notice that the short-ski group was slightly behind in the Pre-Test, but was clearly superior after the two methods of instruction had been used in the Post-Test. In the slalom race the first two women's places and the first two men's places were from the short-ski group. The number-of-turns test was won by a latecomer (advanced beginner) in the control group who had not been present for the Pre-Test. The next 6 places were from the short-ski group. In the free run race the first 6 places were from the short-ski group. The complete results are listed in Table 2.
Table 2 Complete Results
GRAPHICAL REPRESENTATION OF THE RESULTS
Illustration 2 Rank order before and after the test period
STATISTICAL ANALYSIS OF THE RESULTS
Hypothesis tests test whether two treatments produce different results. Based on the results of the sample or test, the significance level indicates the probability that a hypothesis is true. (as opposed to random events causing any difference in results.) (See Conover 6 p. 76.)The testing procedure must decide whether the results of the two treatments are different enough to be able to assume that the 2 treatments always produce different results from one another.
The null hypothesis H0 states the equality of the two treatments which is to be disproved. The alternative hypothesis H1 states that there is a significant diffence between the two treatments. The experimental layout here is the randomized two sample model. For this model there are two distribution free significance tests. Each of these tests is valid with reference to its two hypotheses.
MANN-WHITNEY TEST OF STATISTICAL SIGNIFICANCE
H0 : The middle short-ski participant improves his overall ranking less than or equal to the middle control group participant during the test period .
H1 :The middle short-ski participant improves his overall ranking more than the middle control group participant during the test period .
Note that the term "improved more" is used rather than "skis better". The evaluation "skis better" would include the effect of the skiers inate ability as well as the effect fo the teaching method.
The ranks of the three Post-Test Events are added together for each participant. These sums are replaced by their respective ranks (column labeled End Rank in Table 2). From each rank subtract the corresponding rank from the Pre-Test yielding the rank difference for each participant (column labeled Rank Dif in Table 2). The largest negative difference shows the most improvement as a result of instruction.
From this comparison we can conclude that the short-ski method is superior with a statistical significance of approximately 0.998. This means, that the probability of this result or a stronger result under conditions of the null Hypothesis (no difference in the effectiveness of the two teaching methods) is approximately 0.002 (2 chances in a thousand).
BINOMIAL TEST OF STATISTICAL SIGNIFICANCE
We could reword the hypotheses as follows:
H0 : The Probability is less than or equal to ½, that a short-ski participant improves his rank within the long ski group.
H1:The Probability is greater than ½, that a short-ski participant mproves his rank within the long ski group.
In this case we can use the binomial test to test the significance of the data.
Since each one of the 9 short-ski participants improved his ranking among the control group, the binomial test indicates that this result has a statistical significance of 0.998. This means that this result would occur in only two out of a thousand cases if the underlying probability were less than or equal to ½. The underlying probability is the probability that a short-ski participant would outimprove the members of the control group).
ANALYSIS OF THE DROP-OUT RATE DIFFERENCE WITH A CONFIDENCE INTERVAL
The high signifikanz level of the results of the hypothesis tests applies to those who held out until the end of the course and the final test. 65% of the control group participants (long ski) was incapable or too embarrased to take part to the Final test or quit the skiing course entirely. In comparison only 25% of the short ski participants fell into this category. To what degree can this relationship (40 percentage points difference) be applied to the parent population? One can calculate an area around the observed difference, in which one would expect to find the difference in further equivalent experiments with, say a 0.95 probability. This area is described as a 95% confidence interval. ( see Conover 6 S . 99. ):
The number of the participants of the respective group which would be missing at the final test has a binomial distribution (see Conover 6 S. 81). The parameter p of this distribution control groupe . Assuming that these two proportions are independent from one another, the difference has a Bivariat binomial distribution.
The 95% Konfidenzintervall lies in the area of 7% to 70%. Ths means: We expect that the portion of the conntrol group (long ski) which is absent at the final test to lie between 7% and 70% more than the corresponding portion at the short ski group in 95% of further equivalent experiments
Statistically the results show that it is highly unlikely that such a strong difference in the measurements could have resulted from chance. With these facts it can safely be asserted that the short-ski method is still superior to the standard teaching method when teaching adult beginners to ski. Specifically beginners learn to ski more controlled as well as faster than they would with the standard teaching method. In addition the short ski method produces less ski school dropouts than the long ski method.
1 Ralf Sturm; Vergleich zweier Skilehrmethoden; Diplomarbeit, Friedreich-Alexander-Universität Erlangen-Nürnberg, Institut für Sportwissenschaft, Professor E. Hecht; 1992
2 Karl Pfeiffer; Learning to Ski by the Graduated Length Method ;Ski Magazine (USA); November 1966
3 Art Furrer; Graduation Day for Graduated Length; Ski Magazine (USA); January 1968
4 John Fry; Teaching Systems Compared; Ski Magazine (USA); November 1966
5 Friedl List; Mini Ski Rettung der Hoffnungslosen? ; Ski Magazine(BRD); November 1967
6. W.J. Conover 'Practical Nonparameteric Statistics' John Wiley & Sons, Inc. 1971.
7 Erhart Gattermann et. al.;DSLV Skilehrplan 1; BLV; 1981
8 Erhart Gattermann et. al.; DSLV Skilehrplan Band 1; BLV; 1994
9 Ken Lawler; Teaching Adult Beginners to Ski; Selbstdruck; 1996
10 Heinz Maegerlein, Friedl List, Martin Puchtler; Neuer Schwung auf kurzem Ski; BLV; 1967
11 Max Lundberg; Alpine Skiing; PSIA; 1993
12 Horst Üeberhorst, Walter Kuchler; Vom Kurzski zum Normalski; Lehrhilfen für die Leibeserziehung; 1/69
13 Harald Kiedaisch; Alles redet über den Kurzski; Lehrhilfen für die Leibeserziehung; 1/69