¿CÓMO ELEGIR ANÁLISIS ESTADÍSTICOS COMPARATIVOS? t Student, Mann Whitney, Wilcoxon y Chi Cuadrado.
Comparative Analyses in Inferential Statistics
Overview of Comparative Analyses
- The video introduces comparative analyses as a key component of inferential statistics, emphasizing their division into related and independent samples.
- Independent samples are defined as groups with no prior relationship among subjects, such as comparing sick versus not sick participants or different age groups.
- Related samples involve evaluating the same group at different times, like measuring cognitive abilities at ages 5 and 10.
Statistical Tests for Independent Samples
- For comparing two-level categorical variables with a quantitative dependent variable in independent samples, the Student's t-test is applicable if parametric requirements are met; otherwise, the Whitney U test should be used.
- These tests can address research questions regarding differences between genders on development scales or income disparities between professions.
Statistical Tests for Related Samples
- When dealing with related samples to compare measurement results over time, the Student's t-test for related samples is appropriate if parametric conditions are satisfied; otherwise, the Wilcoxon W test is recommended.
- This analysis can answer questions about IQ differences in children at different ages or theft rate changes over consecutive years.
Chi-Square Analysis
- For qualitative independent variables across two groups, chi-square analysis can determine relationships and independence of variables.
- It addresses questions like voting preferences based on gender or employment opportunities based on language proficiency.
Longitudinal Studies and Categorical Variables
- In cases where there are two measures of a qualitative variable from related samples, the McNemar test identifies changes over time in longitudinal studies.
- This analysis helps assess whether participants develop anxiety disorders post-intervention by categorizing responses into yes/no outcomes.
Advanced Analysis Techniques
- For theoretical analyses involving more than three related measures (e.g., pretest, posttest), Cochran's Q analysis is utilized to evaluate changes within the same group over time.
Conclusion: Application in SPSS