pooled vs non pool
Grading and Feedback on Exam
- Grading in progress, some evaluations pending.
- Open notes exam nature allows for longer, more applied questions.
- Opportunity to improve answers and resubmit for partial credit after grading.
Upcoming Class Considerations
- Feedback on exam length noted; future exams may occur during lab periods.
- Teams for final projects must be formed by the end of the week, choice of independent or group work for master's and PhD students.
Statistical Inference Overview
- Focus on inference for two population means after initial discussions on one population mean.
- Normal distribution described by central tendency (mean) and spread (standard deviation).
- Plan to transition from two means to one and two population variances and ANOVA (analysis of variance) for multiple means.
Data Analysis Steps
- Collect sample data (e.g., skull circumferences for men and women).
- Perform graphical and numerical analysis: normality checks, descriptive statistics.
- Compare sample means to infer population differences.
Statistical Procedures
- Central Limit Theorem applied to sample means.
- Difference of sample means calculated, assuming independence.
- Procedures include confidence intervals and hypothesis testing focusing on p-values.
Null Hypothesis Testing
- Typically: H0: μ1 = μ2; H1: μ1 ≠ μ2.
- Test statistic calculated to assess null hypothesis based on observed data.
- Apply significance levels and confidence intervals to make informed decisions about population means.
Non-Pooled vs. Pooled T-Test
- Non-pooled assumptions: independent samples, normality of distributions, large sample size if normality violates.
- Pooled t-tests used when population variances are assumed equal.
- Degrees of Freedom (df) computed differently for pooled t-tests vs. non-pooled t-tests.
Summary of Hypothesis Testing Process
- Set up null and alternative hypotheses.
- Calculate test statistics (either z or t based on conditions).
- Determine p-value or critical values for decision making.
- Interpret results based on test outcomes (reject or fail to reject H0).
Conclusion
- Analyzed situations where independence may be violated and how to address these in hypothesis testing.
- Emphasized importance of understanding degrees of freedom in t-distributions for accurate statistical interpretation.