Chevy versus Ford NASCAR Race Effect Size – A Meta-Analysis Data Description • All 256 NASCAR Races for 1993-2000 Seasons • Race Finishes Among all Ford and Chevy Drivers (Ranks) – Ford: 5208 Drivers (20.3 per race) – Chevrolet: 3642 Drivers (14.2 per race) • For each race, Compute Wilcoxon Rank-Sum Statistic (Large-sample Normal Approximation) • Effect Size = Z/SQRT(NFord + NChevy) Wilcoxon Rank-Sum Test (Large-Sample) Number of Ford and Chevy Cars in race i : N i N Ford,i N Chevy ,i N i ( N i 1) T T Ford,i Chevy ,i 2 N Ford,i N i 1 Expected Rank Sum for Ford Under No Brand effects : Ford,i 2 N Ford,i N Chevy ,i N i 1 Standard Deviation : Ford,i 12 TFord,i Ford,i Z - Statistic for testin g Brand Effects : Z i Rank Sums for Ford/Chevy in race i : TFord,i TChevy ,i Ford,i Effect Size : d i Zi Z i N Ford,i N Chevy ,i Ni V d i 1 Ni Note : Negative values mean Ford better, Positive means Chevy better Chevy Effect by Race 0.8000 0.6000 0.4000 Chevy Effect 0.2000 0.0000 -0.2000 -0.4000 -0.6000 -0.8000 1 9 17 25 33 41 49 57 65 73 81 89 97 105 113 121 129 137 145 153 161 169 177 185 193 201 209 217 225 233 241 249 Race Number Evidence that Chevrolet tends to do better than Ford Histogram of Effect Sizes 40.0000 35.0000 30.0000 Frequency 25.0000 20.0000 15.0000 10.0000 5.0000 0.0000 -0.2000 -0.1500 -0.1000 -0.0500 0.0000 0.0500 0.1000 0.1500 0.2000 0.2500 0.3000 0.3500 Effect Size Effect Sizes Appear to be approximately Normal 0.4000 0.4500 0.5000 Combining Effect Sizes Across Races • Weighted Average of Race-Specific Effect Sizes • Weight Factor 1/V(di) = Ni = (NFord,i+NChevy,i) wi 1 V d i Ni 1 V d N 256 j j 1 256 j 1 d w wi d i N i i 1 i 1 256 256 2 256 j 1 256 di Ni i 1 701.49 .0793 8850 2 256 1 256 2 2 1 V d w N i N i V d i N i N i N i 1 / 8850 i 1 i 1 i 1 i 1 N i i 1 1 SE d w .0106 8850 95% CI for Population Mean : .0793 1.96(.0106) .0793 .0208 (.0585,.1001) 256 256 Test for Homogeneity of Effect Sizes i True Effect Size for Chevy for race i H 0 : 1 ... 256 H A : Not all i are equal 256 Test statistic : Q i 1 P - value : P 2 255 d d 2 w i ^ 2 d i 242.5 242.5 .7037 No evidence of Heterogene ity of Effect sizes Average Effect by Year (Mean +/- 1SD) 0.4000 0.3000 Average Effect +/-1SD 0.2000 0.1000 0.0000 -0.1000 -0.2000 1992 1993 1994 1995 1996 Year 1997 1998 1999 2000 Testing for Year Effects Notation : p 8 years (classes) mi Number of Races in year i ij True effect for j th race in year i i Weighted Average for year i Weighted Average for all races d ij Observed effect for j th race in year i mi d i Weighted Average for year i N ij j 1 mi N j 1 Weighted Average for all races N ij i 1 j 1 p d 1 mi ij 1 d ij p mi N i 1 j 1 ij d ij Testing for Year Effects Three Models regarding ij : H 0 : ij * i, j p mi QT d mi QB 2 d mi N ij dij d p ^ 2 j 1 dfT mi 1 i 1 2 p mi p N ij di d N i d i d 2 i 1 j 1 i 1 2 df B p 1 ij 2 ij p 2 ij d d d N d d mi H 2 : ij unrestricted i 1 j 1 di d i 1 j 1 QWi d ^ 2 i 1 j 1 p ij H1 : ij i* i, j i ^ 2 mi j 1 ij ij di 2 p QW QWi i 1 p dfW mi 1 i 1 ij QT QB QW Source Between Years Within Years Total dfT df B dfW df 7 248 255 Q-Statistic 15.27 227.19 242.46 P-value 0.0327 0.8243 0.7037 Null H0 H1 H0 Alternative H1 H2 H2 Chevy Effect Size vs Track Length 0.8 0.6 Effect Size 0.4 0.2 0 -0.2 -0.4 -0.6 0 0.5 1 1.5 Track Length 2 2.5 3 Chevy Effect Size vs Laps 0.8 0.6 Effect Size 0.4 0.2 0 -0.2 -0.4 -0.6 0 100 200 300 Laps 400 500 600 Testing for Year and Race/Track Effects • Regression Model Relating Effect Size to: – – – – Season (8 Dummy Variables (No Intercept)) Track Length Number of Laps Race Length (Track Length x # of Laps) • Weighted Least Squares with weighti = Ni Regression Coefficients/t-tests Variable 1993 1994 1995 1996 1997 1998 1999 2000 TrkLen Laps RaceLen beta 0.0281 -0.0548 -0.0368 0.0090 -0.0916 -0.0992 -0.0389 -0.0250 0.0004 0.0397 -0.0001 se(beta) 0.1239 0.1235 0.1221 0.1207 0.1214 0.1202 0.1196 0.1159 0.0003 0.0520 0.0002 t-stat 0.2270 -0.4440 -0.3014 0.0745 -0.7542 -0.8256 -0.3254 -0.2157 1.3402 0.7644 -0.8560 P-val 0.8206 0.6574 0.7634 0.9407 0.4514 0.4098 0.7452 0.8294 0.1814 0.4454 0.3929 Controlling for all other predictors, none appear significant C2 – Tests for Sub-Models and Overall H 10 : All Factors have No Effects 1 ... 11 0 ^ ^ 1 ^ Test Statistic : Q1 ' 75.5 df 11 P .0000 H 02 : No Track Length/Lap s effects 9 10 11 0 ^ ^ 1 ^ Test Statistic : Q2 ' 4.61 df 3 P .2023 H 03 : No Year effects 1 ... 8 0 ^ ^ 1 ^ Test Statistic : Q3 ' 14.77 df 8 P .0637 Sources • Hedges, L.V. and I. Olkin (1985). Statistical Methods for Meta-Analysis, Academic Press, Orlando, FL. • Winner, L. (2006). “NASCAR Winston Cup Race Results for 1975-2003,” Journal of Statistical Education, Volume 14, #3 www.amstat.org/publications/jse/v14n3/datasets.winner.html