S06 Statistics 305 Some Useful Populations Defined in Mathematical Theory (Supplemental Material) Populations defined in mathematical theory do not exist, they are useful figments of our imagination. Random samples from such theoretical populations can be taken using numerical algorithms available in programs like JMP. The populations described below are all theoretical and are grouped into families having one or more indexing parameters to identify family members. The elements of every population are numbers. The set of unique elements in some of the populations is a finite or countably infinite set of numbers. Other populations are such that the unique elements form a continuum. Each population has an associated function, called the Cumulative Distribution Function (CDF), which can be evaluated to find the proportion of population elements which are less than or equal to any given number. A second function, called the Density Function, is also associated with each population. Its graph shows how population proportions change across the unique real number population elements. When we study Probability in Chapter 5, these will be called Probability functions, and proportions will be called probabilities. Don’t worry about probability now. The general term Discrete Population is used for any population whose unique elements form a finite or countably infinite set. Continuous Populations have unique elements that form a continuum. Every theoretical population, discrete or continuous, will be assumed to have an infinite number of elements. 1. Binomial Family of Populations B(n, p). (These are Discrete Populations.) Indexing parameters: n (an integer > 0) and p (a real number 0 < p < 1). Unique population elements: {0, 1, 2, …, n}. Density Function: n! ⎧ x n− x if x = 0, 1, 2, ..., n ⎪⎪ x ! (n − x) ! p (1 − p ) f ( x) = ⎨ ⎪ ⎪⎩ 0 otherwise Cumulative Distribution Function for any x = 0, 1, 2, …, n F ( x) = x ∑ n f ( y) (Note that y =0 ∑ f ( y) = 1 ) y =0 y integer y integer 1 S06 Example: A Binomial B(3, 0.1) is any population whose unique elements are 0, 1, 2 3! 3! and 3 and these occur in respective proportions (0.1)0(0.9)3, (0.1) (0.9)2, 0 !3! 1! 2 ! 3! 3! (0.1)2 (0.9) and (0.1)3 which are density function values. Evaluating these 2 !1 ! 3!0! expressions we have 0.729, 0.243, 0.027, and 0.001. Thus almost 73% of the elements in a Binomial B (3, 0.1) population are zero. The Cumulative Distribution Function value at 2 is F(2) = 0.729 + 0.243 + 0.027 = 0.999. Thus 99.9% of the elements in a Binomial (3, 0.1) population do not exceed 2 (i.e. are not 3). The graph of the Density Function is a bar chart having 4 bars. (See Figure 1 for plots of B (5, 0.2), B (10, 0.2) … etc). 2. Geometric Family of Populations G(p). (These are Discrete Populations.) Indexing parameter: p (any real number 0 < p < 1). Unique population elements: {all the positive integers} Density function: ⎧ p(1 − p) x −1 for x = 1, 2, ... ⎪ f ( x) = ⎨ ⎪ 0 otherwise ⎩ Cumulative Distribution Function for an integer x > 0 F ( x) = x ∑ ∞ p (1 − p ) y −1 (Note: y =1 ∑ f ( y) = 1 ) y =1 y integer See Figure 2 for a graph of the Density Function. 3. Normal Family of Populations N(μ, σ 2). (These are Continuous Populations.) Indexing parameters: μ (any finite real number) σ 2 (any finite positive real number). Unique population elements: all real numbers (-∞, ∞). Density function: φ ( x) = 1 σ 2Π − e 1 ⎛ x−μ ⎞ ⎜ ⎟ 2⎝ σ ⎠ 2 −∞ < x < ∞ Cumulative Distribution Function: x Φ( x) = ∫ φ ( y )dy for any x, − ∞ < x < ∞ −∞ 2 S06 ∞ ∫ ( Note: φ ( y )dy = 1 ) −∞ The graph of the Density Function is a bell-shaped curve, as shown in Figure 3. The Standard Normal Population is that one indexed by μ = 0 and σ 2 = 1, i.e. N(0, 1). 4. Exponential Family of Populations E (α). (These are Continuous Populations.) Indexing parameter: α (any real number > 0) Unique population elements: (all real numbers > 0) Density function: ⎧ 1 − ⎪ ⎪α e f ( x) = ⎨ ⎪ ⎪⎩ 0 x α for x > 0 otherwise Cumulative Distribution Function: ⎧0 ⎪ ⎪ F ( x) = ⎨ x − ⎪ f ( y )dy = 1 − e ∫ ⎪ ⎩0 if x ≤ 0 x α if x > 0 Figure 4 shows graphs of Exponential Density Functions. The Standard Exponential Population is E(1). 5. Uniform Family of Populations. U(a, b). (These are Continuous Populations.) Indexing parameters: interval endpoints (a, b). (a and b are numbers with a < b.) Unique population elements: all real numbers in the interval (a, b). Density function: ⎧ 1 ⎪⎪ f ( x) = ⎨ b − a ⎪ ⎩⎪ 0 a< x<b otherwise Cumulative Distribution Function: x F ( x) = ∫ f ( y )dy = −∞ x−a , b−a The Standard Uniform Population is U(0, 1). 3 a< x<b S06 4