GENERAL MEASUREMENT CONCEPTS

1
GENERAL MEASUREMENT CONCEPTS
2
ELEMENTS OF SCORE ANALYSIS

• TYPES OF SCORES
• COMMON UNIT OF MEASURE
• THREE STEPS OF ANALYZING A SET OF SCORES
• CALCULATORS AND COMPUTERS

3
WHY ANALYSIS TEST SCORES?

• FOR A LARGE GROUP A SIMPLE LIST OF SCORES HAS NO
  MEANING
• OVERALL GROUP PERFORMANCE AND/OR IMPROVEMENT CAN
  ONLY BE INTERPRETED BY CONDENSING THE INFORMATION
  AND APPLYING DESCRIPTIVE STATISTICS TO IT

4
WHY ANALYSIS TEST SCORES?

• ONCE THE OVERALL GROUP PERFORMANCE IS
  DETERMINED, INDIVIDUAL PERFORMANCE CAN THEN BE
  EVALUATED IN RELATION IN RELATION TO THE OVERALL
  GROUP PERFORMANCE

5
TYPES OF SCORES

• CONTINUOUS SCORES
• DISCRETE SCORES

6
CONTINUOUS SCORES

• HAVE POTENTIALLY AN INFINITE NUMBER OF VALUES AS
  THEY CAN BE MEASURED WITH VARYING DEGREES OF
  ACCURACY
• BETWEEN ANY TWO VALUES EXIST COUNTLESS VALUES
  THAT CAN BE EXPRESSRED AS FRACTIONS
• EXAMPLES TIME, WEIGHT, LENGTH

7
CONTINUOUS SCORES

• MOST CONTINUOUS SCORES ARE ROUNDED OFF TO THE
  NEAREST UNIT OF MEASUREMENT WHEN RECORDED
• 5 OR MORE IS ROUNDED UP AND 4 OR LESS IS ROUNDED
  DOWN
• EXAMPLE ROUNDING TO A WHOLE NUMBER 1.6 2
  1.4 1

8
DISCRETE SCORES

• LIMETED TO A SPECIFIC NUMBER OF VALUES
• ARE NOT EXPRESSED IN FRACTIONS
• EXAMPLES - FREE THROWS MADE IN BASKETBALL OR
  SUCCESSFUL THROWS AT A TARGET

9
OTHER CLASSIFICATION OF SCORES

• RATIO
• INTERVAL
• ORDINAL
• NOMINAL

10
RATIO SCORES

• COMMON UNIT OF MEASUREMENT BETWEEN EACH SCORE
• TRUE ZERO POINT
• STATEMENTS ABOUT EQUALITY OF RATIOS CAN BE MADE
  1 SCORE CAN BE REFERRED TO AS 2 TIMES GREATER
  THAN ANOTHER SCORE
• EXAMPLES INCLUDE DISTANCE, TIME, WEIGHT, AND
  LENGTH

11
INTERVAL SCORES

• COMMON UNIT OF MEASUREMENT BETWEEN EACH SCORE
• NO TRUE ZERO (A SCORE OF ZERO ON A KNOWLEDGE TEST
  DOES NOT INDICATE TOTAL LACK OF KNOWLEDGE) AND
  HENCE STATEMENTS ABOUT EQUALITY OF RATIOS CAN NOT
  BE MADE

12

• MOST PHYSICAL PEFORMANCE SCORES ARE EITHER
  RATIO OR INTERVAL SCORES

13
ORDINAL SCORES

• DO NOT HAVE A COMMON UNIT OF MEASUREMENT BETWEEN
  SCORES AND DO NOT HAVE A TRUE ZERO
• THERE IS AN ORDER IN THE SCORES THAT MAKES IT
  POSSIBLE TO DESCRIBE OR LIST ONE SCORE HIGHER
  THAN THE OTHER SCORE
• EXAMPLE CLASS RANK ON A PUSH-UP TEST

14
NOMINAL SCORES

• CANNOT BE RANKED ORDER
• MUTUALLY EXCLUSIVE SCORES
• CANNOT SAY THAT ONE SCORE IS BETTER THAN ANOTHER
  SCORE
• EXAMPLES SPORT PREFERENCE, GENDER, AND ETHNICITY

15
COMMON UNITS OF MEASURE

• SCORES MUST BE RECORDED IN A SINGLE UNIT OF
  MEASUREMENT SUCH AS CENTIMETERS RATHER THAN
  METERS AND CENTIMETERS
• USUALLY RECORDED IN THE SMALLER ONE SUCH AS
  CENTIMETERS
• RECORDING SCORES IN THE SMALLER UNIT OF MEASURE
  AS THEY ARE COLLECTED IS LESS TIME-CONSUMING THAN
  CONVERTING THEM INTO THAT FORM LATER

16
SELECTING A CRITERION SCORE
17
MEAN SCORE VERSUS BEST SCORE

• CRITERION SCORE IS THE MEASURE USED TO INDICATE A
  PERSONS ABILITY
• GENERALLY A BETTER INDICATOR WHEN DEVELOPED FROM
  MORE THAN 1 TRIAL (I.E., MEAN SCORE FROM MORE
  THAN 1 TRIAL)
• RELIABILITY INCREASES AS THE NUMBER OF MEASURES
  INCREASES

18
MEAN SCORE VERSUS BEST SCORE

• GENERALLY THE MOST RELIABILE CRITERION SCORE AND
  THE BEST INDICATOR OF TRUE ABILITY (TYPICAL
  PERFORMANCE) MAY BE THE MEAN SCORE OF MULTIPLE
  TRIALS AS THE SUM OF THE ERRORS OF MEASURMENT IS
  ZERO
• NOTE OBSERVED SCORE EQUALS TRUE SCORE PLUS
  ERROR SCORE

19
MEAN SCORE VERSUS BEST SCORE

• FOR MULTIPLE-TRIAL PHYSICAL PERFORMANCE TESTS
  WHERE MAXIMAL ABILITY RATHER THAN TYPICAL ABILITY
  IS DESIRED, AND THE EQUIPMENT AND SCORING
  PROCEDURE ARE HIGHLY ACCURATE, THE BEST SCORE IS
  THE BEST INDICATOR OF A PERSONS TRUE ABILITY AS
  A PERSON CANNOT SCORE BETTER THAN HER OR HIS
  PHYSIOLOGICAL CAPACITY
• TYPICAL OF MANY PHYSIOLOGICAL MEASUREMENTS TAKEN
  IN A LABORATORY SETTING)

20
MEAN SCORE VERSUS BEST SCORE

• IF THE BEST SCORE IS USED AS THE CRITERION SCORE
• - IT SHOULD BE BASED ON MULTIPLE TRIALS
• - IT IS NOT THE SAME TRIAL FOR EACH PERSON

21
MEAN SCORE VERSUS BEST SCORE

• DEPEDNING ON THE SITUATION AND WHAT THE TEST
  ADMINISTRATOR BELIEVES IS MOST APPROPRIATE, AN
  INDIVIDUALS CRITERION SCORE FROM MULTIPLE TRIAL
  TESTS MAY BE DETERMINED IN ANY OF THE FOLLOWING
  WAYS
• - MEAN OF ALL TRIAL SCORES
• - BEST SCORE OF ALL TRIAL SCORES
• - MEAN OF THE SELECTED TRIAL SCORES BASED ON
  THE TRIALS ON WHICH THE GROUP SCORED BEST
• - MEAN OF THE SELECTED TRIAL SCORES BASED ON
  THE TRIALS ON WHICH THE INDIVIDUAL SCORED
  BEST

22
THREE STEPS OF ANALYZING A SET OF TEST SCORES

• SELECT THE APPROPRIATE ANALYSIS TECHNIQUE
• COMPLETE CALCULATIONS WITH THE ANALSYS TECHNIQUE
• INTERPRET THE RESULTS OF THE ANALYSIS TECHNIQUE

23
CALCULATORS AND COMPUTERS

• SCORE ANALYSIS SHOULD BE ACCURATE AND QUICK
• CALCULATORS WORK WELL WITH SMALL DATA SETS BUT
  WITH LARGER DATA SETS THE USE OF CALCULATORS
  BECOME MORE TIME-CONSUMING AND MORE MISTAKES TEND
  TO BE MADE
• COMPUTERS WITH DATA ANALYSIS SOFTWARE SUCH AS
  SPSS ARE VERY FAST AND ACCURATE

24
QUESTIONS OR COMMENTS??