Análisis de la relación entre dos variables cualitativas. Chi cuadrado: significación Módulo 4

Analysis of Contingency Tables

In this section, the focus is on exploring the causes of significance in contingency tables assuming statistically significant results from the chi-squared test.

Understanding Contingency Tables

• The example involves a contingency table with information on four treatments and their corresponding responses.

• Data from a statistics book by Martín Andrés and Luna del Castillo is used, assuming 560 subjects for analysis.

Hypothesis Testing and Significance

• Explanation of observed frequencies in the table and theoretical hypothesis of independence between response and treatment.

• Introduction to experimental value calculation, degrees of freedom, and asymptotic significance.

Identifying Causes of Significance

This part delves into determining the causes behind statistical significance in data analysis.

Interpreting Results

• Importance of p-values in hypothesis testing; rejecting null hypothesis based on p < 0.05.

• Calculating experimental value by comparing observed and expected frequencies under independence assumption.

Contribution Analysis

• Exploring contributions to experimental value using chi-squared model; identifying significant rows or columns.

• Detailed calculation process for contribution values per cell; assessing impact on overall experimental statistic.

Treatment Contributions to Experimental Value

Analyzing how each treatment group contributes to the overall experimental statistic in contingency table analysis.

Treatment Impact Assessment

• Breakdown of contributions from each treatment group towards the experimental statistic.

Exploring Treatment Effects

In this section, the speaker discusses the impact of different treatments on experimental values and responses, emphasizing the significance of Treatment 2.

Analyzing Treatment Contributions

• The speaker highlights Treatment 2 as the most impactful in terms of contributions to experimental values, while noting that Response Worsen provides significant information.

• Emphasizes further investigation into the significance caused by Treatment 2 and suggests starting with analyzing a row related to this treatment.

Understanding Treatment Impacts

• Considers Treatments 1, 3, and 4 as similar in their effects on responses, with only Treatment 2 showing distinct behavior.

• Removes data related to Treatment 2 to assess the homogeneity of Treatments 1, 3, and 4, leading to accepting the null hypothesis for these treatments.

Statistical Analysis

• Conducts hypothesis testing to determine if Treatments 1, 3, and 4 exhibit homogeneous behavior based on observed frequencies.

• Compares experimental data with critical values to accept or reject null hypotheses regarding treatment independence.

Comparing Treatments: Significance Testing

This section delves into statistical comparisons between treatments using critical values and experimental data.

Calculating Degrees of Freedom

• Explains degrees of freedom calculation based on table dimensions for conducting significance tests.

• Compares experimental results against critical values to make decisions regarding treatment differences.

Assessing Significance

• Combines Treatments 1, 3, and 4 into a single group for analysis while isolating Treatment 2 for comparison against other treatments.

Declarar Diferente el Tratamiento

In this section, the speaker discusses the differentiation in treatment effects and explores how certain treatments behave differently from others.

Treatment Effects Analysis

• Differentiation in treatment 4 is emphasized, while it has been distinct for two cases.

• Focus on checking if responses of equal and better behave similarly to explore differences with worse responses.

• Examination of combining treatments 1, 3, and 4 to analyze responses that remain the same or improve.

Contraste de Independencia

This part delves into conducting a contrast of independence to determine the relationship between response and treatment.

Independence Contrast Process

• Calculation of experimental value compared to critical value for independence contrast.

• Acceptance of null hypothesis due to low experimental value, indicating independence of equal or better response from treatment.

Relación entre Respuesta y Tratamiento

The discussion shifts towards exploring the relationship between response type and treatment administered.

Relationship Analysis

• Rejection of null hypothesis based on high experimental value, signifying a connection between response type and treatment.

• Reference to a detailed scheme outlining the significance search process in bio-statistics literature.

Significación Estadística y Conclusiones

This segment focuses on statistical significance determination and drawing conclusions from the study results.

Statistical Significance Assessment

• Identification of statistical significance through P-value analysis below 0.05 threshold.