Ch 3.2 Stats Refresher Continued

Question 1

Raw Score

Answer 1

• Straightforward unmodified account of performance
• Most basic level of info provided by a test
• Raw score means little to nothing without effectively organizing data

Question 2

Distribution

Answer 2

Test scores arrayed for recording or study

• Grouped frequency distribution
• Simple frequency distribution

Question 3

Measures of Central Tendency

Answer 3

• Mean

Most appropriate for interval/ratio data when distributions are believed to be normal

• Mean
• Mode

Mainly used with nominal data

Question 4

Measures of Variability

Answer 4

Indication of how scores in a distribution are scattered or dispersed

• Range

Quick but gross score and can be altered by extreme scores

• Standard Deviation (SD)

Degree of dispersion around the mean

Question 5

Normal Curve

Answer 5

Bell-shaped, smooth curve is highest at the center

Important characteristics

• Mean = median = mode
• Divided into standard deviation units
• 50% will be above the mean 50% will be below the mean

Question 6

Skewness

Answer 6

The nature & extent to which symmetry is absent from a distribution

• Positive Skew

– Few scores at high-end distribution

• Negative skew

– Few scores at low-end of distribution

Question 7

Standard Scores

Answer 7

Raw score converted to a scale to compare scores

Question 8

Why convert scores?

Answer 8

• Easier interpretation
• Communicates relative standing vs. other test takers

Question 9

Standard Score Types (Z)

Answer 9

• Number of SDs a raw score is from the mean
• Mean = 0 SD = 1
• Approximate range = -3 to +3
• Z = (your score - mean score)/SD

Question 10

Standard Score Types (T)

Answer 10

• Mean = 50 SD =10
• Approximate range = 0 to 100
• T = 10z + 50
• Advantage of no negative scores

Question 11

Correlations

Answer 11

Degree and direction of correspondence between two things

• -1.0 to +1.0
• 0 = NO correlation (now LOW correlation)
• Not an index of causal relations

Question 12

Types of correlations

Answer 12

• Pearson r
• Spearman’s rho

–Sample size small (<30 pairs of measurements)

– When both are in ordinal - or rank - order form

Question 13

Restriction of Range

Answer 13

When the range is restricted correlation becomes smaller

Question 14

Regression

Answer 14

• Analyzing relationships among variables to understand how one variable may predict another
• Simple: predicting a DV from one IV (predicting 1st-year grad GPA from GRE score)
• Multiple: predicting a DV from more than one IV (predicting 1st-year grad GPA from GRE score & undergrad GPA)

Question 15

Meta-Analysis

Answer 15

• Use data from several studies to best estimate the relationship between 2 variables
• Statistically combining information across studies

– Gives more weight to studies with a larger sample size