T STUDENT & WILCOXON (Muestras relacionadas) | Explicación, interpretación y reporte en SPSS

Understanding Experimental Design in Weight Loss Studies

Overview of the Study Design

• Researchers propose a weight loss treatment using pills, involving two groups: a control group (obese individuals not receiving treatment) and an experimental group (obese individuals taking the pills) .

• Both groups undergo pre-treatment and post-intervention measurements to assess the effects of the pills combined with exercise sessions .

Key Comparisons in Research

• The first comparison assesses whether both groups are similar at the start; if they differ, any end-of-study differences cannot be attributed solely to the intervention. A p-value greater than 0.05 indicates no significant difference between groups initially .

• The second comparison occurs post-intervention, where researchers expect differences due to treatment. Here, a p-value less than 0.05 would lead to rejecting the null hypothesis, indicating effectiveness of the treatment .

Intragroup Comparison Importance

• To confirm that changes in weight for the experimental group result from treatment rather than other factors, an intragroup comparison is necessary—comparing pre-treatment weights against post-treatment weights within the same group .

• This analysis uses related samples statistics; if parametric assumptions hold, a Student's t-test is applied; otherwise, non-parametric tests like Wilcoxon W are used for analysis of related samples .

Statistical Analysis Techniques

Choosing Appropriate Statistical Tests

• For parametric data: use Student's t-test for related samples; for non-parametric data: apply Wilcoxon W test. It's crucial to differentiate these tests as their formulas and execution methods vary significantly .

Application Example with Children’s Executive Function Measures

• A study on children aged 7 to 12 applies a cognitive stimulation program aimed at improving executive functions. Post-intervention comparisons will determine if there are significant differences in performance measures .

Analyzing Results Using SPSS

• When analyzing data in SPSS:

• Input pre-treatment scores as variable one and post-treatment scores as variable two.

• Review means, standard deviations, and correlations between variables to assess significance .

Interpreting Statistical Outcomes

Reporting Findings from SPSS Output

• The output includes mean scores showing improvement from pre-treatment to post-treatment measurements. Strong correlation results indicate significant relationships between variables measured .

Significance Testing Results

• Reported t-values and corresponding p-values help determine statistical significance; if p < 0.05, reject null hypothesis indicating effective treatment impact on children's executive functions .

Non-parametric Analysis Considerations

Conducting Non-parametric Tests

• If data does not meet parametric criteria:

• Use Wilcoxon W test via SPSS by selecting appropriate variables.

• Ensure descriptive statistics options are set correctly before running analyses .

Final Interpretation of Results

• Analyze ranges based on internal formulae used by non-parametric tests versus means used in parametric tests.

Understanding Effect Sizes in Statistical Analysis

The Importance of Effect Sizes

• The difference statistic alone does not confirm that observed changes are due to treatment; other variables (e.g., age) may influence results.

• When comparing two samples, ensure the variable being compared is the one responsible for any differences. Effect sizes help clarify this relationship.

• Journals require effect size calculations in submissions; they validate that reported differences stem from the independent variable or treatment.

Calculating Effect Sizes with SPSS

• In recent SPSS versions, effect sizes can be calculated directly within the analysis path, providing immediate results in data output tables.

• Cohen's d is used for t-tests while Hedge's g applies to Wilcoxon tests. A moderate effect size indicates a meaningful impact of treatment between measurements.

Reporting Results According to APA Standards

• After analysis, report means and standard deviations for dependent variables. Include t-values, degrees of freedom, p-values, and effect sizes in your findings.

• Present results either as a table or within text; tables are preferred for multiple dependent variables while fewer comparisons can be integrated into paragraphs.

Example Reporting Format

• An example format includes stating significant differences found in verbal fluency scores with specific mean values and statistical significance indicators (e.g., t-value and p-value).

• Ensure clarity by including all relevant statistics such as means, standard deviations, t-values, p-values, and effect sizes without omitting leading zeros where necessary.

Non-parametric Analysis Reporting

Analysis of Descriptive Statistics and Frequencies

Overview of Statistical Analysis

• The analysis focuses on descriptive statistics and frequencies for study variables, specifically pre and post fluidity rings and interference.

• Frequency tables are omitted in this instance; instead, central tendency (median) and dispersion (range) are selected for reporting.

• It is recommended to present results in a table format when dealing with multiple pairs of repeated measures.

Reporting Results

• A significant finding was reported: verbal fluency showed statistically significant differences between pretest (median = 6, range = 10) and posttest scores (median = 11, range = 10), with Z = -5.26 and p < 0.001.

• When presenting results, both paragraph form and tables should include median, range values, Z value, p value, and effect size (gd h).

Formatting Guidelines

• In statistical reporting, omit the zero before the decimal point for p values less than one; however, effect sizes can exceed one.

• For bilateral significance values or p < 0.001 from SPSS output, use "<" followed by "0.001" in tables as probabilities cannot equal zero.

Chi-Square Analysis for Categorical Variables

Introduction to Chi-Square Test

• The chi-square test is introduced as a method for analyzing categorical variables with independent measures.