Wednesday, January 26, 2011

Modern Measurement: Chapter 7: The Normal Curve and Distributional Statistics

Description and Characteristics
-          Background
o   Frequency distribution function
o   Don't start from zero and then graduating to higher amounts à psychological traits more accurately viewed as relative quantities, and measurement of them can be made more meaningful by starting in middle, and extending outwards in both directions.
-          Features of normal curve
-          Origins of std deviation and more history

Significance to our Lives
-          Normal curve affords us a way to anticipate likely events in aggregate
-          Caution about over mapping
-           
Forms For Expressing Data in Variables
-          Three ways data may be cited:
o   In discrete classes
o   As graphic
o   By stating a rule

-          Frequency data in discrete classes
o   Shown in a table
-          Graphics Displaying Frequency Distribution Data
-          Rules describing frequency distribution data

Technical Depiction of Normal Distribution
-          Characteristics of normal distribution

Standardising the Normal Curve
-          z score to too many distributions
-          solution of normal curve density function
-          rationale, computation and proof for z scores
-          cumulative probabilities

Common Measurement Indexes Used With the Normal Curve of Distribution

Percentiles and Percentile Ranks
-          explanation of percentiles
-          cautions , strengths and weaknesses of quartiles
o   ordinal nature of percentile scale, few statistical manipulations are possible
o   distortion in scale when percentiles are calculated for tests used with small groups / test scale limited in range large number of possible scores not used
o   small distribution is skewed – causes further distortions in mapping process  [GW: this is treated as an advantage when dealing with data with outliers]
o   Percentile ranks as non-equal interval scale







               



No comments: