Introduction to Descriptive Statistics 2/25/03 Population vs. Sample Notation Population Greeks , , Vs Sample Romans s, b Types of Variables ~Nominal (Quantitative) Ordinal Nominal (Qualitative) Interval or ratio Describing data Moment Center Spread Mean Variance (standard deviation) Skew Skewness Non-mean based measure Mode, median Range, Interquartile range -- Peaked Kurtosis -- Mean n x i i 1 n X Variance, Standard Deviation ( xi ) 2 , n i 1 2 n ( xi ) n i 1 n 2 Variance, S.D. of a Sample ( xi ) 2 s , n 1 i 1 2 n ( xi ) s n 1 i 1 n 2 Coefficient of variation c.v. 100 Skewness Symmetrical distribution • IQ • SAT Frequency Value Skewness Asymmetrical distribution • GPA of MIT students Frequency Value Skewness (Asymmetrical distribution) • Income • Contribution to candidates • Populations of countries • “Residual vote” rates Frequency Value Skewness n x x 3 i i 1 n (n 1) n2 3/ 2 x x i 1 3/ 2 i (mean mod e) / s 3 (mean median) / s Skewness Frequency Value Kurtosis k>3 Frequency k=3 k<3 Value A few words about the normal curve • Skewness = 0 • Kurtosis = 3 Frequency Value 1 ( x ) / 2 2 f ( x) e 2 More words about the normal Frequency curve 34% 34% 47.7% 47.7% x Value SEG example The instructor and/or section leader: Mean s.d. Skew Kurt Gives well-prepared, relevant presentations 6.0 0.69 -1.7 8.5 Explains clearly and answers questions well 5.9 0.68 -1.0 4.8 Uses visual aids well 5.6 0.85 -1.8 8.9 Uses information technology effectively 5.5 0.91 -1.1 5.0 Speaks well 6.1 0.69 -1.5 6.8 Encourages questions & class participation 6.1 0.66 -0.88 3.7 Stimulates interest in the subject 5.9 0.76 -1.1 4.7 Is available outside of class for questions 5.9 0.68 -1.3 6.3 Overall rating of teaching 5.9 0.67 -1.2 5.5 Graph Graph some SEG variables The instructor and/or section leader: s.d. Skew Kurt 5.6 0.85 -1.8 8.9 Graph .6 Fraction Uses visual aids well Mean 0 1 6.1 0.66 -0.88 3.7 .6 Fraction Encourages questions & class participation 7 (mean) q3 0 1 7 (mean) q6 Binary data X prob( X ) 1 proportion of time x 1 s x (1 x ) s x x (1 x ) 2 x Commands in STAT for getting univariate statistics • • • • • summarize summarize, detail graph, bin() normal graph, box tabulate [NB: compare to table] Explore Q9: Overall teaching evaluation subject 3.371 3.982 3.14 14.02D 21W.803 21M.480 17.906 2.51 q9 6.4375 6.73333 6.46154 5.66667 5.66667 5.69231 5.28571 5.88235 n 16 15 13 3 12 13 14 17 Graph Q9 . graph q9 Fraction .505495 0 2.33333 7 (mean) q9 Divide into 7 “bins” and have them span 1, 1..2, 2..3, … 6..7 . graph q9,bin(7) xscale(0,7) Fraction .57326 0 0 7 (mean) q9 Add ticks at each integer score . graph q9,bin(7) xscale(0,7) xlabel(0,1,2,3,4,5,6,7) Fraction .57326 0 0 1 2 3 4 (mean) q9 5 6 7 Add a finer grain to the bars . graph q9,bin(14) xscale(0,7) xlabel(0,1,2,3,4,5,6,7) Fraction .318681 0 0 1 2 3 4 (mean) q9 5 6 7 Even finer grain • . graph q9,bin(28) xscale(0,7) xlabel(0,1,2,3,4,5,6,7) Fraction .181319 0 0 1 2 3 4 (mean) q9 5 6 7 Superimpose the normal curve (with the same mean and s.d. as the empirical distribution) . graph q9,bin(28) xscale(0,7) xlabel(0,1,2,3,4,5,6,7) norm Fraction .181319 0 0 1 2 3 4 (mean) q9 5 6 7 Do the previous graph with only larger classes (n > 20) . graph q9 if n>20,bin(28) xscale(0,7) xlabel(0,1,2,3,4,5,6,7) Fraction .202532 0 0 1 2 3 4 (mean) q9 5 6 7 Draw the previous graph with a box plot . graph q9 if n>20,box ylabel (mean) q9 7 6 5 4 3 Draw the box plots for small (0..20), medium (21..50), and large (50+) classes . gen size = 0 if n <=20 (237 missing values generated) . replace size=1 if n > 20 & n <=100 (196 real changes made) . replace size = 2 if n > 100 (41 real changes made) (mean) q9 8 6 . sort size . graph q9 ,box ylabel by(size) . graph q9 ,box ylabel by(size) 4 2 0 1 2 A note about histograms with unnatural categories From the Current Population Survey (2000), Voter and Registration Survey How long (have you/has name) lived at this address? -9 -3 -2 -1 1 2 3 4 5 6 No Response Refused Don't know Not in universe Less than 1 month 1-6 months 7-11 months 1-2 years 3-4 years 5 years or longer Simple graph Fraction .557134 0 1 6 PES8 Solution, Step 1 Map artificial category onto “natural” midpoint -9 -3 -2 -1 1 2 3 4 5 6 No Response missing Refused missing Don't know missing Not in universe missing Less than 1 month 1/24 = 0.042 1-6 months 3.5/12 = 0.29 7-11 months 9/12 = 0.75 1-2 years 1.5 3-4 years 3.5 5 years or longer 10 (arbitrary) Graph of recoded data Fraction .557134 0 0 1 2 3 4 5 longevity 6 7 8 9 10 Density plot of data 0 0 1 2 3 4 5 6 longevity 7 8 9 10 15