Ch 2

Introduction to Statistics in Psychometrics

  • Descriptive statistics: Utilized to describe and summarize what has occurred in a data set. Allows for evaluation and analysis of observations.

  • Inferential statistics: Involve making inferences about a larger population based on observations from a smaller sample.

Scales of Measurement

  • Purpose: Different scales of measurement are crucial for determining the appropriate statistical tests and operations to apply.

Properties of Scales

  1. Magnitude: Refers to the presence of an order or ranking within the scale where one quantity can be considered “more” or “less” than another.

  2. Equal Intervals: This property implies that the difference between values is consistent across the range of values.

  3. Absolute Zero: Represents a point where the property being measured is absent; distinguishes actual measurement from arbitrary relationships.

Types of Scales of Measurement

Nominal Scale

  • Definition: Simple classification without any order or ranking.

  • Examples: Race/ethnicity, Sex/gender (e.g., 1 = male; 2 = female).

  • Properties: Magnitude = No; Equal Intervals = No; Absolute 0 = No.


Ordinal Scale

  • Definition: Measures magnitude and provides a rank order.

  • Characteristics: While it indicates order (e.g., preferences), the intervals between ranks are not necessarily equal.

  • Properties: Magnitude = Yes; Equal Intervals = No; Absolute 0 = No.


Interval Scale

  • Definition: Contains properties of both magnitude and equal intervals but lacks an absolute zero, making true ratios impossible.

  • Properties: Magnitude = Yes; Equal Intervals = Yes; Absolute 0 = No.

  • #### Ratio Scale

    • Definition: Incorporates all properties: magnitude, equal intervals, and an absolute zero.

    • Properties: Magnitude = Yes; Equal Intervals = Yes; Absolute 0 = Yes.

Permissible Operations Based on Scale

  • Nominal Data: Can organize data but allows no mathematical operations (categorical).

  • Ordinal Data: Arithmetic operations can be applied, but conclusions may be difficult due to missing intervals (categorical).

  • Interval Data: All arithmetic operations can be utilized, but ratio conclusions cannot be drawn due to the absence of absolute zero (quantitative).

  • Ratio Data: Full range of operations permitted (quantitative).

Knowledge Check 1

  • To measure differences in aggression based on biological sex, the appropriate scale of measurement would be: Nominal.

Frequency Distributions

  • Definition: A frequency distribution displays how often each score occurs within a dataset, typically displayed on ascending X and Y axes.

  • Note: Not all frequency distributions are symmetrical.

Example of Frequency Distribution

  • Example dataset for rainfall (inches) over years is provided, illustrating frequency distribution.

  • Shows how to interpret data across several years of observed rainfall:

    • Mean (Average) Rainfall = 10.11 inches.

Percentile Ranks

  • Definition: Percentile ranks measure how many items fall below a given point in a dataset.

  • Example Application: Determines how many participants in a race were outperformed by a given individual.

Percentile Calculation Formula

  • The formula for calculating a percentile rank is as follows: Pr=BN×100Pr = \frac{B}{N} \times 100

    • Where:

    • PrPr = Percentile rank

    • XiXi = Score of interest

    • BB = Number of scores below XiXi

    • NN = Total number of scores.

Knowledge Check 2

  • If the number of scores below a person's height is 60 out of 250 total scores, calculate the percentile rank.

Percentiles

  • Definition: Percentiles refer to specific points within a distribution that divide the data into hundredths.

  • Example: 95th percentile for IQ is equivalent to a score of 125, where the percentile rank for that score is also 95.

Describing Distributions

  • Mean: The arithmetic average of a data set.

    • Formula: Xˉ=∑XN\bar{X} = \frac{\sum X}{N}

    • Where ∑X\sum X = sum of all scores and NN = total score count.

  • Standard Deviation: Measures variance in a dataset.

    • Definition: The square root of the variance provides insights into data variability, enhancing the meaning gleaned from the mean.

Standard Deviation Calculation

  • Standard Deviation (Population):
    σ=∑(X−μ)2N\sigma = \sqrt{\frac{\sum (X - \mu)^{2}}{N}}

  • Standard Deviation (Sample):
    s=∑(X−Xˉ)2N−1s = \sqrt{\frac{\sum (X - \bar{X})^{2}}{N - 1}}

  • Illustrates how to calculate values based on age within the PSYCH 323 cohort.

Psychological Testing in Everyday Life

  • Population vs. Sample

    • Population: All elements with a specific, common definition.

    • Sample: A subset of the population, intended to represent it fairly.

  • Descriptive Characteristics: Parameters for populations, statistics for samples.

Symbols Used in Statistics

  • Population Mean: μ\mu

  • Sample Mean: Xˉ\bar{X}

  • Population Standard Deviation: σ\sigma

  • Sample Standard Deviation: ss

Z Scores

  • Definition: A Z score standardizes scores by measuring how many standard deviations a score is from the mean.

    • Formula for Z score:
      Z=(X−μ)σZ = \frac{(X - \mu)}{\sigma}

  • Interpretation:

    • Positive Z score: Score is above the mean.

    • Negative Z score: Score is below the mean.

Standard Normal Distribution

  • Definition: A symmetrical binomial probability distribution using Z scores on the X-axis. Illustrates the distribution of scores within the population based on standard deviations.

  • Recognize that typically 68% of data falls within one standard deviation of the mean, 95% within two, and 99.7% within three standard deviations (68–95–99.7 rule).

McCall's T

  • A system of data transformation to provide intuitive meaning to raw scores.

  • Assigns percentile equivalents: mean set at 50, with a standard deviation of 10.

  • T score formula: T=10Z+50T = 10Z + 50

Quartiles and Deciles

  • Quartiles: Divide a data set into four equal parts (e.g., 50th percentile is median).

  • Deciles: Divide data into ten equal parts; Stanine system converts scores to a range of 1-9 (Standard Nines).

Norms

  • Definition: Reflect the performance of defined groups on tests, represented via various statistics (e.g., Z scores, quartiles).

  • Provide contexts for individual performance against a standardization sample.

Age-Related Norms and Tracking

  • Definition: Norms that differ based on demographic factors, such as age—important for assessing growth measures (e.g., height and weight).

  • Tracking: Refers to maintaining a consistent level relative to peers over time.

Criterion-Referenced Tests

  • Contrast with norm-referenced tests, focusing instead on comparing individual skills to established criteria. Important for creating individualized plans.