The t Test for Two Independent Samples txtbook

Chapter 10 | The t Test for Two Independent Samples

Preview

• Description of a soccer shoot-out as a high-tension event in sports.

• Greenlees, Eynon, and Thelwell (2013) conducted a study on penalty kick performance in college soccer.

• Experiment involved 10 penalty kicks taken by players against goalkeepers in different jersey colors (red, green, blue, yellow).

• Key findings:

  • Players scored significantly fewer goals against a goalkeeper in a red jersey (average M = 5.40) compared to a green jersey (average M = 7.50).

  • Possible explanations for this difference:

  1. A real effect of jersey color on success.

  2. Sampling error leading to differences in means.

• Introduction of the independent-measures t test for hypothesis testing with two samples, addressing mean differences between two treatment conditions.

10-1 Introduction to the Independent-Measures Design

Learning Objective 1

• Define the concepts of independent-measures and repeated-measures designs with examples.

• Importance of comparing two (or more) sets of sample data in research.

• Examples in psychology showing comparisons (e.g., generational differences, teaching methods, therapy techniques).

• Independent-Measures Design:

  • Involves separate and independent samples for each treatment condition.

• Repeated-Measures Design:

  • Involves the same individuals being measured under different conditions.

  • E.g., measuring depression before and after therapy in the same patients.

Sample Structure

• Visualization of independent-measures research design involving two separate samples representing two unrecognized populations or treatments.

• Example setup:

  • Population A vs Population B

• Analysis of statistical techniques for independent measures design is discussed.

Examples of Independent vs Repeated Measures

• Independent Measures: Comparison of vocabulary sizes based on socioeconomic status.

• Repeated Measures: A clinical study comparing patient depression scores pre- and post-therapy.

10-2 The Hypotheses and the Independent-Measures t Statistic

Learning Objectives

1. Describe the hypotheses for an independent-measures t test.

2. Definitions and differences between independent-measures t statistics regarding single-sample t.

3. Calculate pooled variance and standard error for sample mean differences.

4. Complete t statistic with degrees of freedom.

5. Use subscripts to denote sample data in calculations (e.g., n1, n2 for sample sizes).

Hypotheses Statement

• Null Hypothesis (H0): States there is no difference (H0: m1 - m2 = 0).

• Alternative Hypothesis (H1): Indicates mean difference exists (H1: m1 - m2 ≠ 0).

T Statistic Formulation

• The t statistic for independent measures compares two sample means:
  $t = \frac{(M<em>1 - M</em>2) - (m<em>1 - m</em>2)}{s<em>{(M</em>1 - M_2)}}$.

• Understanding estimated standard error for t statistic.

Calculating Standard Error and Pooled Variance

• Pooled variance estimation is essential when the population variances are assumed equal:

  1. $s<em>p^2 = \frac{SS</em>1 + SS<em>2}{df</em>1 + df_2}$

  2. Calculate weighted variance from sample sizes, assisting in minimizing biases.

• Emphasis on avoiding biases by utilizing the pooled variance.

Conclusion of Section

• Essential understanding of hypotheses underpinning independent-measures t tests, including constructing variance calculations.

10-3 Hypothesis Tests with the Independent-Measures t Statistic

Learning Objectives

1. Evaluation of treatment significance through hypothesis testing with independent measures.

2. Conducting one-tailed tests under directional hypotheses effectively.

3. Understanding assumptions related to independent-measures tests (e.g., homogeneity of variance).

• Example study comparing cognitive performance under varying light conditions showcasing effects.

  • Critical region finding and analysis of effect outcomes are discussed.

Calculations of t Statistic

1. Overall structure setup outlined, focusing on summary calculations leading to hypothesis testing by degrees of freedom.

2. Student example outlining data collection guidelines and outcome conclusions regarding cognitive differences based on light conditions.

10-4 Effect Size and Confidence Intervals for the Independent-Measures t

Learning Objectives

1. Measuring effect size using Cohen’s d and r2 based on independent measures.

2. Confidence interval calculations for exploring mean differences.

3. Reporting variables cohesively with significance and effect measures in scientific reporting standards.

Cohen's d Representation

• Cohen’s d methodical approach emphasizes the absolute magnitude of the effect beyond mere statistical significance:
  $d = \frac{M<em>1 - M</em>2}{s_{p}}$.

• Cases summarizing outcomes reflect substantial treatment effects among different scenarios.

Summary

• The independent-measures t test offers valuable statistical methodologies enhancing research efficacy, specifically in understanding relational dynamics among separate samples.

• Throughout the analysis, core terms such as homogeneity, pooled variance, and the critical values provide templates for understanding and executing hypothesis testing in independent frameworks.