Annoucements. MT2 - Review. one variable. two variables

Transcription

1 Housekeeping Annoucements MT2 - Review Statistics 101 Dr. Çetinkaya-Rundel November 4, 2014 Peer evals for projects by Thursday - Qualtrics will come later this evening Additional MT review session this evening 7-8pm at Physics 130 OH before MT: Derek - Tuesday: 4-6pm Xinyi - Tuesday: 7-9pm Dr. C-R - Wednesday: 12:30-2:30pm Michael - Wednesday: 5-7pm SEC both days: 4-9pm Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Map of concepts Map of concepts inference HT CI one variable 2 levels 3+ levels H0: mu = mu_0 H0: = med_0 randomization H0: p = p_0 H0: p follows hypothesized distribution S-F fail -> randomization E >= 5 -> chi-sq GOF E < 5 -> randomization theoretical Z, T, F, chi-sq Z, T HT x: 2 levels H0: mu_1 = mu_2 simulation randomization bootstrap two variables y = x = x: 3+ levels x&y: 2 levels H0: med_1 = med_2 randomization H0: All mu_i are equal H0: p_1 = p_2 ANOVA, F H0: All med_i are equal randomization S-F fail -> randomization y = x = x/y: 3+ levels H0: x and y are independent E >= 5 -> chi-sq independence E < 5 -> randomization Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22

2 Map of concepts one variable 2 levels mu 3+ levels N/A p bootstrap S-F fail -> bootstrap Which of the following is true? Modeling ( response) CI two variables y = x = y = x = x: 2 levels x: 3+ levels x&y: 2 levels mu_1 - mu_2 med_1 - med_2 N/A N/A p_1 - p_2 bootstrap S-F fail -> bootstrap (a) If the sample size is large enough, conclusions can be generalized to the population. (b) If subjects are randomly assigned to treatments, conclusions can be generalized to the population. (c) Blocking in experiments serves a similar purpose as stratifying in observational studies. (d) Representative samples allow us to make causal conclusions. (e) Statistical inference requires normal distribution of the response variable. x/y: 3+ levels N/A Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Which of the following is the best visualization for evaluating the relationship between two variables? Modeling ( response) Two students in an introductory statistics class choose to conduct similar studies estimating the of smokers at their school. Student A collects data from 100 students, and student B collects data from 50 students. How will the standard errors used by the two students compare? Assume both are simple random samples. Modeling ( response) (a) side-by-side box plots (b) mosaic plot (c) pie chart (d) segmented frequency bar plot (e) relative frequency histogram (a) SE used by Student A < SE used as Student B. (b) SE used by Student A > SE used as Student B. (c) SE used by Student A = SE used as Student B. (d) SE used by Student A SE used as Student B. (e) Cannot tell without knowing the true of smokers at this school. Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22

3 Which of the following is the best method for evaluating the relationship between two variables? data Inference Which of the following is the best method for evaluating the relationship between a and a variable with many levels? data Inference Modeling ( response) Modeling ( response) (a) chi-square test of independence (b) chi-square test of goodness of fit (c) anova (d) linear regression (e) t-test (a) z-test (b) chi-square test of goodness of fit (c) anova (d) linear regression (e) t-test Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Data are collected at a bank on 6 tellers randomly sampled transactions. Do average transaction times vary by teller? Modeling ( response) Response variable:, Explanatory variable: ANOVA Summary statistics: n_1 = 14, _1 = , sd_1 = n_2 = 23, _2 = , sd_2 = n_3 = 15, _3 = 82.66, sd_3 = n_4 = 15, _4 = , sd_4 = n_5 = 44, _5 = , sd_5 = n_6 = 29, _6 = , sd_6 = H_0: All s are equal. H_A: At least one is different. Analysis of Variance Table Response: data Df Sum Sq Mean Sq F value Pr(>F) group Residuals Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22

4 Application exercise: ANOVA Data are collected on download times at three different times during the day. We want to evaluate whether average download times vary by time of day. Fill in the??s in the ANOVA output below. data Inference Modeling ( response) Response variable:, Explanatory variable: Summary statistics: n_early (7AM) = 16, _Early (7AM) = , sd_early (7AM) = n_eve (5 PM) = 16, _Eve (5 PM) = , sd_eve (5 PM) = n_late (12 AM) = 16, _Late (12 AM) = , sd_late (12 AM) = Analysis of Variance Table Response: data Df Sum Sq Mean Sq F value Pr(>F) group???????? 1.306e-11 Residuals?? ?? Total?? What is the result of the ANOVA? Early (7AM) Evening (5 PM) Late Night (12 AM) Since 1.306e-11 < 0.05, we reject the null hypothesis. The data provide convincing evidence that the average download time is different for at least one pair of times of day Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Application exercise: ANOVA (cont.) The next step is to evaluate the pairwise tests. There are 3 pairs of times of day 1 Early vs. Evening: left side of class (facing the board) 2 Evening vs. Late Night: center of class 3 Early vs. Late Night: right side of class Determine the appropriate significance level for these tests, and then complete the test assigned to your team. α = 0.05/3 = (1) Early vs. Evening T 45 = = = p val < 0.01 (2) Evening vs. Late Night T 45 = = = p val < 0.01 (3) Early vs. Late Night T 45 = = = 4.81 p val < 0.01 Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22

5 What percent of variability in download times is explained by time of day? Modeling ( response) n = 50 and ˆp = Hypotheses: H 0 : p = 0.82; H A : p We use a randomization test because the sample size isn t large enough for ˆp to be distributed nearly normally ( = 41 < 10; = 9 < 10). Which of the following is the correct set up for this hypothesis test? Red: success, blue: failure, ˆp sim = of reds in simulated samples. Modeling ( response) Response: data Df Sum Sq Mean Sq F value Pr(>F) group e-11 Residuals (a) = 0.67 (b) (c) (d) (a) Place 80 red and 20 blue chips in a bag. Sample, with replacement, 50 chips and calculate the of reds. Repeat this many times and calculate the of simulations where ˆp sim (b) Place 82 red and 18 blue chips in a bag. Sample, without replacement, 50 chips and calculate the of reds. Repeat this many times and calculate the of simulations where ˆp sim (c) Place 82 red and 18 blue chips in a bag. Sample, with replacement, 50 chips and calculate the of reds. Repeat this many times and calculate the of simulations where ˆp sim 0.80 or ˆp sim (d) Place 82 red and 18 blue chips in a bag. Sample, with replacement, 100 chips and calculate the of reds. Repeat this many times and calculate the of simulations where ˆp sim 0.80 or ˆp sim Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Randomization distribution What is / should be the center of the randomization distribution? What is the result of the hypothesis test? observed randomization statistic Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Summary of methods Inference for data: One : Parameter of interest: µ n 30 Z, T, n < 30 T One vs. one (with 2 levels): Parameter of interest: µ 1 µ 2 n 1 and n 2 30 Z, T, n 1 or n 2 < 30 T If samples are dependent (paired), first find differences between paired observations One vs. one (with 3+ levels) - : Parameter of interest: NA ANOVA HT only For all other parameters of interest: simulation Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22

6 Summary of methods Inference for data - binary outcome: Summary of methods Inference for data - 3+ outcomes: Binary outcome: One : Parameter of interest: p S/F condition met Z, if not simulation One vs. one, each with only 2 outcomes: Parameter of interest: p 1 p 2 S/F condition met Z, if not simulation S/F: use obs. S and F for CIs and exp. for HT 3+ outcomes: One, compared to hypothetical distribution: Parameter of interest: NA At least 5 exp. successes in each cell χ 2 GOF, if not simulation HT only One vs. one, either with 3+ outcomes: Parameter of interest: NA At least 5 exp. successes in each cell χ 2 Independence, if not simulation HT only Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22 Statistics 101 (Dr. Çetinkaya-Rundel) MT2 - Review November 4, / 22