Ch. 10 - The t Test For Two Independent Samples

Page 1: Recap - Z-Test

  • Calculate Z-Test

    • Components needed:

      • Mean (M)

      • Population Mean (µ)

      • Population Standard Deviation (σ) or Sample Size (n1)

Page 2: Sample Variance

  • Sample Variance (s²)

    • Represents the mean squared distance from the mean.

    • Obtained by dividing the sum of squares by (n - 1).

  • Estimated Standard Error of the Mean (sM)

    • Used when σ is unknown.

    • Computed from the sample variance or standard deviation.

    • Provides an estimate of the difference between a sample mean (M) and the population mean (µ).

Page 3: Recap - Single Sample T-Test

  • Single Sample T-Test

    • Formula: t = (M - μ) / SM

      • M = Sample mean

      • μ = Population mean (hypothesized from H0)

      • SM = Estimated standard error (computed from the sample data)

Page 4: Recap - Single Sample T-Test Calculation

  • Calculate T-Test Components

    • We need:

      • Sample Mean (M)

      • Population Mean (µ)

      • Sum of Squares (SS) or Standard Deviation (s)

      • Sample Size (n)

Page 5: Recap - Variables

  • Independent Variable (IV)

    • Manipulated by the researcher.

    • The variable that is tested or manipulated.

  • Dependent Variable (DV)

    • Observed to assess the effect of the treatment.

    • The outcome variable; referred to as the ‘Dependent Measure’.

Page 6: The T-Test for Two Independent Samples

  • Introduction to Chapter 10.

Page 7: Research Designs for Two Sets of Data

  • Research studies often require the comparison of two or more sets of data.

  • Two General Research Designs:

    1. Separate groups of participants.

    2. Same group of participants.

  • Independent-Measures Design

    • Utilizes completely separate groups of participants for each treatment condition.

Page 8: Two Basic Designs

  • Between Groups (Independent Measures)

    • IV manipulated between different groups of subjects.

    • Different subjects experience a single condition.

    • Best design when carryover effects are expected.

  • Within Subjects (Repeated Measures)

    • IV manipulated within a single group of subjects.

    • Each participant is exposed to all levels of the IV; use when carryover effects are unlikely.

Page 9: Nomenclature for Independent Samples

  • Independent-Measures Terms

    • Between-Subjects Design

    • Between-Groups Design

    • Independent (sample) t-test

    • Values/Levels of an IV are Groups or Treatments.

    • IV can also be referred to as “Factor.”

Page 10: Levels of an IV

  • Differences in experiments:

    • Between Groups Experiment: Groups or Treatments.

    • Within Subjects Experiment: Conditions.

Page 11: Basic Between-Groups Design

  • Single IV manipulated at 2 levels

    • e.g., Placebo vs 5 mg of drug, Men vs Women.

Page 12: Testing the Basic Between-Groups Design

  • Use independent t-test to test the difference between means.

  • Group A and Group B scores are independent of each other.

Page 13: Independent-Measures Research Study

  • Involves two separate samples to obtain information about two unknown populations or treatment conditions.

Page 14: Goal of Independent-Measures Study

  • Evaluate the mean difference between two populations or treatment conditions.

    • Means represented as µ1 for the first population and µ2 for the second.

    • Null hypothesis states no difference: µ1 - µ2 = 0.

    • Alternative hypothesis states mean difference exists: µ1 - µ2 ≠ 0.

Page 15: Null Hypothesis for Independent-Measures Test

  • Null Hypothesis (H0): µ1 - µ2 = 0 or µ1 = µ2.

  • Alternative Hypothesis (H1): The means are different.

  • Non-directional hypothesis (two-tailed).

Page 16: One-Tailed Predictions for Population 1 and 2

  • Predictions for Population 1:

    • H0: μ1 ≤ μ2

    • H1: μ1 > μ2

  • Predictions for Population 2:

    • H0: µ1 ≥ µ2

    • H1: μ1 < μ2

Page 17: Independent-Measures T-Statistic

  • Evaluates hypothesis about differences between two population means using sample means difference (M1 − M2).

Page 18: Standard Error in T-Statistic

  • Measures error expected using sample means to represent population means.

  • Represented by symbols sM1-M2.

Page 19: Estimated Standard Error of M1 − M2

  • Interpreted as the average distance between a sample statistic and the population parameter.

Page 20: Formula Development for s(M1-M2)

  • Total error in using two sample means to approximate two population means calculated by adding errors from each sample.

Page 21: Error Measurement in Independent-Measures

  • Error from each sample contributes to total error in using sample means.

Page 22: Total Error Calculation for Sample Means

  • Calculate the total error for two sample means.

Page 23: Sample Variance with One Sample

  • Sample variance calculated and combined into pooled variance corrects bias in the standard error.

Page 24: Independent T-Test – Pooled Variance

  • Pooled Variance combines the sample variances to correct bias in standard error.

  • Allows larger sample to contribute more.

Page 25: Independent T-Test - Estimated Standard Error of the Mean

  • Relevant formula updates with pooled variance considerations.

Page 26: Final Formula for Independent-Measures T Statistic

  • Complete formula incorporating components and degrees of freedom.

Page 27: Degrees of Freedom Calculation for T Statistic

  • Degrees of freedom (df) calculated as: df = (n1 – 1) + (n2 – 1) = n1 + n2 - 2.

Page 28: Degrees of Freedom for Entire Data Set

  • Necessary for determining critical t-value.

Page 29: Basic T Statistic Calculation

  • Formula example: t = (M1 - M2) / SM.

Page 30: Elements of a T Statistic

  • Each component defined:

    • Hypothesized parameter, estimated error, standard variance.

Page 31: Independent-Measures T Statistic Overview

  • Uses two samples to evaluate mean difference between two treatment conditions.

Page 32: Result Write Up for Significant Results

  • Guide for writing results when significant:

    • "The analysis showed a significant effect of IV on DV ..."

Page 33: Result Write Up for Non-Significant Results

  • Guide for writing results when not significant:

    • "Even though mean measure was ... this difference was not significant."

Page 34: Hypothesis Tests Steps with Independent-Measures T Statistic

  • Steps involve stating hypotheses, computing degrees of freedom, obtaining data, and making decisions.

Page 35: The Critical Region in Hypothesis Tests

  • Overview of critical region and decision criteria for rejecting null hypothesis.

Page 36: Example Study: Sleepytime

  • Sample size analysis with control and experimental groups detailed.

Page 37: Step 0: Identification of DV and IV

  • Clarification of dependent and independent variables.

Page 38: Step 1: Hypothesis and Alpha Level

  • Null hypothesis defines equal means; alternative defines different means.

Page 39: Step 2: Locate T Critical Value

  • Critical value based on total score number and degrees of freedom.

Page 40: Step 3: Calculate Test Statistics

  • Preliminary calculations shown with example data.

Page 41: Formulas to Pool Variances

  • Summarizes formulas for combining sample variances.

Page 42: Step 4: Make a Decision

  • Decision-making process based on comparison of t-obtained vs t-critical.

Page 43: T Distribution Table Overview

  • Value entries for t corresponding to proportions in tails.

Page 44: Reporting the Result - Example Statement

  • Guide for structuring result interpretation statements in research.

Page 45: Assumptions of Independent-Measures T Formula

  • Key assumptions stated; independent observations and population normality required.

Page 46: Homogeneity of Variance Assumption

  • Importance of variance equality when sample sizes differ.

Page 47: Hartley’s F-Max Test Overview

  • Test for homogeneity of variance through sample variances comparison.

Page 48: Computing Hartley’s F-Max Test Steps

  • Steps to compute sample variance and F-max.

Page 49: Critical Value Comparison in Hartley's Test

  • Explain how to locate critical values in F-max table.

Page 50: F-Max Critical Values Table

  • Display of critical values by sample sizes and alpha levels.

Page 51: Hartley’s F-Max Test Decision Criteria

  • Outcome implications based on F-max values.

Page 52: Handling Violations of Homogeneity Assumption

  • Post-hoc analysis adjustments when homogeneity is violated.

Page 53: Effect Size Overview

  • Defines effect size and relevance of measuring treatment effects.

Page 54: Effect Size Decision-Making Criteria

  • Scenarios for determining when to reject the null hypothesis based on t statistics.

Page 55: Measuring Effect Size with Cohen’s d

  • Calculation of mean differences standardized through Cohen's d.

Page 56: Percentage of Variance Accounted for (r2)

  • Discusses effect size in terms of variance explained by treatment.

Page 57: Percentage of Variance Calculation Format

  • Formal method to express variance accounted for by treatment effect.

Page 58: Confidence Intervals for Mean Differences

  • Step for estimating the unknown population difference based on sample data.

Page 59: Estimated t Value Implications

  • Use of t distribution for estimating population mean difference through confidence intervals.

Page 60: Confidence Interval Interpretation

  • How to determine population mean score range based on sample analysis.

Page 61: Impact of Variability and Sample Size on Tests

  • Discussion on how both factors impact hypothesis test outcomes.

Page 62: Standard Error Influences on T Statistic

  • Exploration of relationships between variance, sample size, and outcomes.

Page 63: Independent-Measures T-Test Key Features

  • Review of key components needed for independent measures t-test construction.