Week 7 Introduction to Statistical Inference
1. Introduction to Statistical Inference
Course: Biostatistics and Epidemiology (BIOE 211)
Institution: Our Lady of Fatima University College of Medical Laboratory Science
2. Understanding Statistical Inference
Definition: Process of generalizing conclusions about a target population based on results from a sample.
Method: Utilizes sample statistics to infer unknown parameters of the population.
3. Key Components of Statistical Inference
3.1 Parameters and Statistics
Parameters: Measures computed from the entire population; classified as constants.
Statistics: Measures computed from samples; considered variable.
3.2 Sampling Variation
Variability in statistics due to random selection of samples.
Recognizes that statistics can differ from one sample to another.
4. Statistical Notation Summary
Measure | Parameter | Statistic |
|---|---|---|
Mean | π» | π₯Μ |
Variance | πΒ² | sΒ² |
Standard Deviation | π | s |
Proportion | P | p |
5. Methods of Statistical Inference
5.1 Estimation
Definition: Process of utilizing sample statistics to estimate population parameters.
Steps:
Collect data from sample respondents.
Calculate summary measures.
Use statistics to estimate parameter values.
5.2 Hypothesis Testing
Definition: Evaluating the validity of a hypothesis regarding a population based on sample data.
Steps:
Collect sample data.
Calculate relevant statistics.
Apply statistical tests.
Make a decision regarding the hypothesis.
6. Properties of the Sampling Distribution of π₯Μ
Mean of the sampling distribution π₯Μ (π xΜ ) equals the population mean (π).
Standard deviation of π₯Μ (π xΜ ) equals population SD (π) divided by the square root of n.
The sampling distribution of π₯Μ approximates a normal distribution:
Normally distributed samples lead to normally distributed π₯Μ distribution.
Non-normally distributed populations approximate normality with a sufficiently large n (Central Limit Theorem).
7. Estimation of the Population Mean
7.1 Point Estimate
Definition: A single number representing the parameter estimate.
Application: Best estimate of population π is the mean of the sample; requires a random sample.
Drawback: Subject to sampling error.
7.2 Interval Estimate
Definition: Estimate of the parameter within a specified range of values.
Procedure: Add and subtract an amount from π₯Μ to create an interval that likely contains π.
Considerations: Select desired confidence level (e.g., 90%, 95%, 99%).
8. Interval Estimate Calculation
Using Standard Deviation:
Formula: π₯Μ Β± Z(π/βn)
Confidence Levels:
99%: Z = 2.575
95%: Z = 1.96
90%: Z = 1.645
9. Example of Interval Estimate
Objective: Estimate the mean weight of school-aged children.
Sample Data: 70, 74, 75, 78, 74, 64, 70, 78, 81, 73, 82, 75, 71, 79, 73, 79, 85, 79, 71, 65, 70, 69, 76, 77, 66.
Mean (π₯Μ ): 74.16 lbs; Standard deviation: 5.37 lbs.
95% CI: [72.05, 76.27] lbs.
10. Another Example of Interval Estimate
Scenario: Health services utilization affected by distance.
Sample Information: Mean distance travelled = 7 km; Standard deviation = 3.2 km.
Constructing 95% CI: (Details of calculations omitted).
11. Student's t Distribution Table
Displays critical values for various degrees of freedom and significance levels for one-tailed and two-tailed tests.
12. Conclusion
Reinforces the foundational concepts of statistical inference, the significance of estimations, and hypothesis testing in biostatistics.