Scales of Measurement →

Measurement →

→
→

Scale →

Continuous Scale
Ratio Scale

→
Discontinuous Scale
→

Discontinuous Continuous
Nominal Ordinal Interval Ratio
Error
Categorizes & labels variables ✓ ✓ ✓ ✓
Ranks categories in order ✓ ✓ ✓
Has equal intervals/differences ✓ ✓
Has a true or meaningful zero ✓

FOUR LEVELS OF MEASUREMENT

Describing Data
Nominal Scales
Distribution
→ Categorical Scale/Variable
→ Raw Scores

→ Standardized Scores

Ordinal Scales

FREQUENCY DISTRIBUTION
→
Frequency Distribution

→
Simple/Ungrouped FQ

→
Grouped FQ

→
Interval Scales
 Upper Limits MEASURES OF VARIABLITY
Lower Limits
Measure of Variability
Graph
→
Histogram

→
Abscissa (X-axis)

Ordinate (Y-axis)

Bar Graph Range

→ outlier
→
→
→
Frequency Polygon
Interquartile & Semi-quartile Ranges

MEASURES OF CENTRAL TENDENCY → Quartile

Measures of Central Tendency
→ Quarters

Mean
→ median/midpoint
→ quarter-points.

Median

→
Mode

→
Interquartile Range
→

Semi-interquartile Range
→
 → Perfectly symmetrical distribution →
Negative Skew

Average Deviation (AD) →

KURTOSIS

Kurtosis
→

∑|𝑥|
Formula: AD = 𝑛

Where: 𝑥

| |
Variance Platykurtic curve
Leptokurtic curve
Mesokurtic curve

∑ 𝑥2 High kurtosis distribution

Formula: 𝑠 2 = 𝑛

Where: 𝑥 Low kurtosis distribution

Standard Deviation (SD)
→

→ →

√∑ 𝑥 2
Formula: SD = 𝑛
√𝑠 2

Note: 𝜎 The Normal Curve

Normal Curve

→ →
asymptotically

SKEWNESS
→
Skewness
→
→
→

Positive Skew
Area Under the Normal Curve → zero plus or minus one scale.
→ μ = 0 and σ = ± 1
→ → positive
and negative
→
T Scores

𝑋̅− 𝜇
→ Formula: 𝑡 = 𝑠/√𝑛
𝑡 = (𝑧)10 + 50

→
→
→ fifty plus or minus ten scale
→ μ = 50 and σ = ± 10
Tail
→

Stanine

→ μ = 5 and σ = ± 2
→

Percentile Rank

Standard Scores →

Standard Score

Formula:
𝑝
=
100 ∙ (𝑛+1)
→ →
Where: 𝑝

z Scores

𝑋− 𝑋̅
Formula: 𝑧 = 𝑠

Where: 𝑋 𝑠
𝑋̅ 𝜇
IQ Scores Coefficient Correlation (r)

→ 100 plus or minus 15 scale

→ μ = 100 and σ = ± 15 →

→
▪
▪ Sign

Linear Transformation
▪ Magnitude

→
Positively (Directly) Correlated

→ Negatively (Inversely) Correlated

 
Nonlinear Transformation r=0

r=±1
→

→ perfectly positive correlation

→
perfectly negative correlation

→
CORRELATION ANALYSIS

normalized Pearson r / Pearson Correlation Coefficient / Pearson

→ Normalized Standard Score Scale Product-Moment Coefficient of Correlation

Correlation and Inference

∑(𝑋−𝑋̅)(𝑌−𝑌̅ )
Formula: 𝑟 =
Correlation ̅ )𝟐 ][∑(𝒀−𝒀
√[∑(𝑿− 𝑿 ̅ )𝟐 ]

∑ 𝑥𝑦
𝑟=
√(∑ 𝒙𝟐 )(∑ 𝒚𝟐 )

Where: 𝑥 𝑋 − 𝑋̅ 𝑦 𝑋 − 𝑌̅
→ 
 → → →

Coefficient of Determination or r2

→ r = .9 r2 = .92 x 100 = 81%

Outlier
→ 100(1 − r2)
19%,
→

Spearman Rho / Rank-Order Correlation Coefficient /

→
Rank-Difference Correlation Coefficient

Meta-Analysis

GRAPHIC REPRESENTATIONS OF CORRELATION Effect Size

Scatter Plot / Bivariate Distribution / Scatter Diagram /

Scattergram →

Some advantages to meta-analyses are:

→

Curvilinearity

nonlinear