1 STAT 6020 Introduction to Biostatistics Fall 2005 Dr. G. H. Rowell Class 2
2 Review Questions Chapter 1 Statistical Inference Chapter 2 Data Types: Numerical/Categorical Chapter 3 What is the difference in a bar chart & a histogram? Describe a useful transformation & how it works.
3 Ch4: Theoretical Distributions, An Overview Probability Samples/Population Distributions Continuous Normal, Lognormal, Uniform Discrete Binomial, Poisson
4 Ch 4: Probability We teach an entire course on this – STAT 6160 Not a main focus of this course Understand Basic Axioms Randomness Independence Probability Distributions Functions
5 Ch 4: Probability - Basics S = Sample space E = an event in the Sample Space P(E) = Probability that event E occurs 0<= P(E) <=1 P(S) = 1 If E1, E2, E3, … are mutually exclusive events, then probability of the union of events = sum of the individual events P(E1 U E2 U E3 U …) = P(E1) + P(E2) + P(E3) + … for a finite or an infinitely countable number of events
6 Ch 4: Probability - Independence Independent Events Events A & B are independent if and only if P(A given that you know everything about B) = P(A) OR P(A and B) = P(A) * P(B) Over simplifying: A & B are independent if knowing the outcome of A tells us nothing about B
7 Ch 4: Sample & Populations Population Sample Goal of Statistics
8 Ch 4: Probability Distributions Decision: Continuous or Discrete ? If Continuous, what is the shape of the relative frequency of the outcomes? Flat – Uniform Bellshaped – Normal Positively Skewed – Lognormal
9 Ch 4: Probability Distributions Decision: Continuous or Discrete ? If Continuous, what is the shape of the relative frequency of the outcomes? Flat – Uniform Bellshaped – Normal Positively Skewed – Lognormal
10 Ch 4: Probability Distributions If Discrete, what experiment is the variable modeling Counts number of successes – might be binomial Counts number of trials to the first success – might be geometric Counts independent, random, and RARE events – might be Poisson
11 Ch 4: Normal Distribution Mound-shaped and symmetrical Mean and standard deviation used to describe the distribution “Empirical Rule”
12 Standard Normal Normal with mean zero and standard deviation 1 Notation: N(0, 1) Z-score Formula Meaning Tools for finding probabilities Tables, software, applets
13 Statistical Software Online StatCrunch StatiCui VassarStats
14 Visualization What does “normal” look like? Histogram: See Figure 4.7, page 60. Normal Density Function Normal Cumulative Distribution
15 Ch 4: Example, Normal If the average daily energy intake of healthy women is normally distributed with a mean of 6754 kJ and a standard deviation of 1142 kJ than what is the probability that a randomly selected women is below the recommended intake level of 7725 kJ per day? Above 7725 kJ? Between 6000 and 7000 kJ?
16 Ch 4: Serum Albumin Example Data: 216 patients with primary biliary cirrhosis mean serum albumin level: g/l, st dev = 5.84 g/l See histogram, Fig 4.5 page 56, follows normal distribution Constructing Chart on Page
17 Ch 4: A Continuous Skewed Right Distribution: Lognormal Example: Serum Bilirubin, page 61
18 Ch 4: Continuous Distribution: Uniform Conditions for Uniform Visualization
19 Ch 4: Discrete Distributions Binomial Distribution Binomial Experiment: Binomial Random Variable: Binomial Distribution Function:
20 Ch. 4: Binomial Example
21 Ch. 4: Binomial Visualization Homework: Complete the Binomial Visualization Activity found at 1/Pages1/Home.htm Be sure to submit the “Pretest” and the “Lesson.” You may want to print the results as a back-up. This is a Hand-in Homework worth 10 points.
22 Ch 4: Discrete Distributions: Poisson Distribution Conditions for a Poisson Distribution: Poisson Visualization: /applets/PoiDensityApplet.html
23 Ch 4: Homework Exercises # 1 – 8 Check you answers in the Back of the book. Bring to class for next week – the mean and standard deviation for heights of Americans of your gender.