.png)
Provisional Dissertation Title: Principal Component Analysis of Valid Performance Indicators in Non-Elite Volleyball
Those who read my Facebook status updates will know I am currently busy writing my dissertation. I put it off for a year but it is now time to get back to it and get my Distinction grade. I am excited about it because it is a question I came up with and it’s an question that I really want answered. So what is it all about? Read on if you want to know…
Literature Review of Valid Performance Indicators in Volleyball
I wrote a literature review of valid performance indicators in volleyball for my Masters degree. A (PI) is a combination of raw performance data, known as action variables, that describes performance. I have written about it before. For example, the combination of action variables for a kill percentage is calculated as: total kills/ total attack attempts. It describes the attack performance but specifically the percentage of attack attempts that results in a kill. A PI is whether the PI is associated to a successful outcome. A successful outcome can be defined as the difference between winning and losing teams or the difference between first in the league and last in the league. Research studies into valid PIs aims to establish this link.
So what did the literature review of valid performance indicators studies reveal? I have written about it before. In essence, the literature shows that attack PI, in one form or another, has the strongest association to a successful outcome. Serve and block PIs has some support but there is weak support for reception PIs and no support for dig PIs as decisive for success. No study has looked into setting. Below is a summary of studies that looked into it:
Limitations
One limitation is it doesn’t reflect a common volleyball coaching philosophy that attack is dependent upon the first ball touch of reception and defence. That is, if a team can’t defend, they can’t set and if they can’t set, they can’t mount an effective attack. This would mean that if attack PIs are critical to success then defence PIs should be decisive too. (PCA) may be a method that incorporates defence as critical to success. Principal component analysis is a statistical analysis technique that reduces a large number of variables to a smaller number of independent factors.
A second limitation is these studies used data at the elite level such as the Olympic Games. It be argued that at this level, every team can receive and defend effectively so every team can mount an effective attack. Thus the difference between teams at the elite level is how effective they can attack. This could explain the results from the research studies. At non-elite level however, non-elite teams would not able to execute all six technical skills compared to the elite teams. This may mean that performance indicators at elite level may be different at non-elite level, which might include defence.
Dissertation Rationale
1. No research study have used PCA to identify key performance indicators. It may incorporate defence as critical to success.
2. No research study have used data at the non-elite level. Valid PIs may be different to those at the elite level.
Principal Component Analysis of Valid Performance Indicators in Non-Elite Volleyball
This is the provisional dissertation title that I am writing. I will be using PCA to identify valid performance indicators in non-elite volleyball. Non-elite volleyball data will be taken from England’s Super 8s league. I already done an initial study using data that I collected for the Super 8s Statistics. Below is a table showing what I found out.
It shows the Points Efficiency Ratio (points scored/ points lost) as the best indicator of performance, followed by attack and setting but surprisingly limited support for serving. Of course, this is not enough for a Masters dissertation as there are some glaring weaknesses, e.g. the data may not be accurate or reliable (even though I collected it) and if this is the case, how can we trust the results? No statistical tests were performed on the data so how do we know these results didn’t come about by chance? Some PIs may not be the best, e.g. if a team has played a number of 5 set matches, their per set averages will go down because Set 5 is only played up to 15 points and not the normal 25. All these weaknesses will be addressed in my dissertation.
Progress
So far I have written a first draft of my introduction and literature review. It’s far from perfect but I am on a schedule to complete it. I am now on my data collection phase. It will take a while because I basically need to put in 39 hours of data collection per week over three weeks. This is on to of my job at Wagamama (20+ hours), looking for performance analyst jobs (on Mondays) and social activities such as being with my newly born niece and playing the sport of volleyball itself!
Watch this space to see what I found out…
Comments