Methodology Part 1 Research Design and Analysis SSP4000 Research Process Week 19 (SPA-specific) Methodology Part 1 Research Design and Analysis
Learning Outcomes By the end of this session you will be able to List the information needed to answer your research question Identify the output variables involved Identify the raw action variables needed Devise a method to answer your research question
Weeks 19, 20 and 21 Week Content 19 Methodological approach Information needs Study design 20 Measurement issues in performance analysis research Validity Objectivity Reliability 21 Interactive session on planning the research essay
Plenary task on methods 5% Complete the proforma by 6pm on Tuesday 15th March 2016 Detail of Study design Sample Data gathering procedures Data analysis Reliability and validity
Approaches Qualitative Quantitative
Information Needs Example – RED CARD STUDY Scope English FA Premier League Matches with a single red card before 60 minutes Matches where the team reduced to 10 men is drawing or winning by a single goal at the time of the dismissal Question (Level at time of dismissal) Does work-rate differ between outnumbered teams who avoid losing and those who lose? What about the opposition? Question (Leading by a goal at time of dismissal) Does work-rate differ between outnumbered teams who hold on to their lead and those who don’t?
Questions What do we mean by work-rate? How is it measured? Are there several variables? Where will data come from to allow us to make the comparisons we need to
Thinking ahead – results format
Results format If we have a well understood research question, it acts as a logical framework for data gathering and analysis The research question dictates the format of the answer
Work-rate variables Percentage of time spent performing high speed running (> 5.5 m.s-1) Number of high speed runs per minute We are mainly interested in period when the team was outnumbered Would be interesting to compare with period prior to the dismissal There is more than one way of doing this
Quantitative Data Analysis Data Processing For our example, the Prozone 3 system captures player movement data These are quality assured by Prozone staff visually verifying frames Data for individual matches then need to be brought together to determine results for “the average match” – quantitative analysis PROZONE Quantitative Data Analysis
Data Sources X, Y locations pitch dimensions 0,0 is centre circle Distance covered, velocity and acceleration derivatives Time of red card in each match We need an understanding of what Prozone3 provides and what we need to produce through data processing
Data gathering and analysis activities Consider Manual notation methods Human operated (tagging) computerised systems Fully automated systems Which of these involve the most / least data gathering effort? Which of these involve the most / least data analysis effort? Let the students consider these questions in groups – ask them what they think then go to next slide to discuss the answers
Data Gathering and Analysis Data analysis Manual OR Human operated computer system Something still needs to be observed and recorded by human operator Fully automated Very efficient Cost £$ Remember – manual verification in Prozone means it is NOT fully automated Manual May need to process chronological event lists to obtain totals / percentages Automated or human operated computer system Once data are stored in databases / spreadsheets, rapid processing, sorting and analysis is possible
Measurement Issues Consider Compare these in terms of Manual notation methods Human operated (tagging) computerised systems Fully automated systems Compare these in terms of Validity Objectivity Reliability Again – let the students discuss and then go to next slide for the answer
Measurement Issues Manual Notation Computerised (Human operated) Fully Automated Validity Depends on variables ! Depends on variables Objectivity Subjective judgement at point of data recording Subjective judgement at point of data entry Totally Reliability Depends on operator consistency Depends on operator consistency – could be complicated or simplified by the use of computer Very reliable - but it could be the same wrong answer recorded by 2 copies of the system ! For Validity – use prozone v human observer recognising work / rest as an example – or use the computerised scoring system in boxing.
Feasibility Questions for observational studies Do you have permission to use footage or record your own footage? Are such permissions realistic to obtain? Will footage show what you need to observe? Can the behaviours you are interested in really be distinguished? Are there enough matches satisfying criteria for inclusion Let the students consider these questions in groups – ask them what they think Consider the last point – how many matches does a player play at Wimbledon per year? How many matches are both teams in a soccer match level ahead and behind for at least 15 mins each?
Questions for experimental studies Will the participants of the level you are interested be willing to participate? Will they really interfere with their training programmes, to participate in a research study? Do they need to travel to lab / testing sites? How long will testing take – will facilities be available? Let the students consider these questions in groups – ask them what they think
Key thought Would you participate in a study like the one you have proposed? If so, why? What are the benefits? If not, why not? Think about what the organisational and logistic issues of the study are Honestly consider if it would really happen
Something you can think about Try this general approach for your own research topic What are the variables of interest? How will they be presented? What data will be collected and how? Is this feasible