Professional Documents
Culture Documents
1. Calculate observed and expected matrices Expected Count: City Suburb Rural Republicans 93.82 72.97 58.20 Democrats 97.16 75.57 60.27 Independents 25.02 19.46 15.52 216 168 134 Observed Count: City Suburb Rural Republicans 126 61 38 Democrats 71 93 69 Independents 19 14 27 216 168 134 Expected Percent: City Republicans 43.44% Democrats 44.98% Independents 11.58% 100% Observed Percent: City Republicans 58.3% Democrats 32.9% Independents 8.8% 100%
2. Subtract observed and expected matrices Observed - Expected: City Suburb Rural Republicans 32.18 -11.97 -20.20 Democrats -26.16 17.43 8.73 Independents -6.02 -5.46 11.48
3. Square the differences City Suburb Rural Republicans 1035.40 143.35 408.23 Democrats 684.26 303.89 76.14 Independents 36.23 29.81 131.76
4. Standardize: Divide the squared values by the expected values 11.04 7.04 1.45 5. Sum the matrix calculated chi-squared value: 43.81 1.96 4.02 1.53 7.01 1.26 8.49
6. Find the critical value from the chi-square table using 0.5 significance and degrees of freedom defin d.f.: (3-1)*(3-1) = 2*2 = 4
9.49
7. Compare the chi squared value and the citical value to make conclusion
Since chi-squared value (43.81) is greater than critical value (9.49), the matrices are NOT the same (i.e. there IS a relationship between voting and reg So we reject the null hypothesis of no association. Null hypothesis: Voting preference and region are NOT associated Alternative hypothesis: Voting preference and region ARE associated
Chi-squared formulas:
Suburb Rural 43.44% 43.44% 44.98% 44.98% 11.58% 11.58% 100% 100% Suburb Rural 36.3% 28.4% 55.4% 51.5% 8.3% 20.1% 100% 100%
ALTERNATIVE METHOD (use observed count matrix): City Repub. Democ. Indep. 126 71 19 216 0.33 0.10 0.03 1.08 Suburb 61 93 14 168 0.10 0.22 0.02 0.08 Rural 38 69 27 134 0.05 0.15 0.09 518 43.81 225 233 60 518
Are trips and income related? Average Household Income Low Average High 1-5 258 138 80 Annual 40 38 24 Out-of-village 6-10 Trips more than 10 43 31 132 341 207 236 341 207 236 341 207 236 0.41 0.05 0.03 0.19 0.07 0.02 0.06 0.02 0.36 1.21 0.21 161.44
othesis and conclude that there IS a relationship between trips and income.
Rank (1=low) Is there a linear relationship between ranks of energy consumption and income? Per Per capita capita energy income Atlantic 1 1 Quebec 5 3 Ontario 4 4 Prairie 2 2 British C. 3 5
ption and income? there a linear relationship between ranks of moisture and temperature? Is
Moisture Temp. 47 35 28 27 44 66 60 75 73 45
21 18 9 25 27 38 33 48 49 32
mperature?
Rank (1=low) Per capita energy Atlantic Quebec Ontario Prairie British C. 1 5 4 2 3 Per capita income 1 3 4 2 5
d 0 2 0 0 -2
d^2 0 4 0 0 4 8
Rank of Temperature
3 2 1 1 2 2 3 3 4 4 5 5 6
12 10 8 6 4 2 2
r=1-((6*8)/(5*(5^2-1)))
0.6
Non-Parametric Conclusion: mild positive correlation between per capita energy and per capita income
Alternative hypothesis: there is a linear relationship Two-tailed test: reject null hypothesis if r>0.648 r<-0.648 degrees of freedom = # of pairs
Moisture Temp. 47 35 28 27 44 66 60 75 73 45
21 18 9 25 27 38 33 48 49 32
Moisture Rank 6 3 2 1 4 8 7 10 9 5
Temp. Rank 3 2 1 4 5 8 7 9 10 6
d 3 1 1 -3 -1 0 0 1 -1 -1
d^2 9 1 1 9 1 0 0 1 1 1 24
Rank of Temperature
12 10 8 6 4 2 2 4 6 8 10 12
Rank of Moisture
r=1-((6*24)/(10*(10^2-1)))
0.85
Non-Parametric Conclusion: strong positive correlation between per capita energy and per capita income
Parametric Conclusion: 0.855>0.648 so reject null hypothesis and conclude that there is a linear relationship between ranks of moisture and temperature df 10 critical value (from table) 0.65
and temperature
% of working % of persons population in aged 25+ managerial with college rank or tech jobs degrees rank jobs degrees Alaska 22.7 15.7 3 2 Arkansas 15.4 14.5 1 1 California 28.6 23.4 6 6 Georgia 34.1 31.1 8 11 Maine 32.3 26.5 7 7 Michigan 24.7 17.4 4 4 New Jersey 40.5 30.4 11 10 North Carolina 22.3 17.3 2 3 Oregon 35.5 29.8 9 8 Washington 27.7 22.9 5 5 New York 39.1 30.3 10 9
d 1 0 0 -3 0 0 1 -1 1 0 1
d^2 1 0 0 9 0 0 1 1 1 0 1 14
6*(d^2) r=1n(n^2-1) 0.94 >0.623 CONCLUSION: Reject the null hypothesis. There is a strong linear relationship between rank jobs and rank degrees.
nd rank degrees.
x^2 2209 1225 784 729 1936 4356 3600 5625 5329 2025 27818
y^2 441 324 81 625 729 1444 1089 2304 2401 1024 10462
xy 987 630 252 675 1188 2508 1980 3600 3577 1440 16837
1837 53.08
0.91 38.24
Non-Parametric conclusion: r=0.905 so there is a strong positive correlation between moisture and temperature
Parametric conclusion: 0.905>0.632 so reject null hypothesis and conclude that there is a linear relationship between moisture and temp df 8 critical value (from table) 0.63
Temperature
60 50 40 30 20 10 20 30 40 50 Moisture 60
d temperature
50
60
70
80
oisture
% of working % of persons population in aged 25+ managerial with college or tech jobs degrees x^2 y^2 xy Alaska 22.7 15.7 515.29 246.49 356.39 Arkansas 15.4 14.5 237.16 210.25 223.3 California 28.6 23.4 817.96 547.56 669.24 Georgia 34.1 31.1 1162.81 967.21 1060.51 Maine 32.3 26.5 1043.29 702.25 855.95 Michigan 24.7 17.4 610.09 302.76 429.78 New Jersey 40.5 30.4 1640.25 924.16 1231.2 North Carolina 22.3 17.3 497.29 299.29 385.79 Oregon 35.5 29.8 1260.25 888.04 1057.9 Washington 27.7 22.9 767.29 524.41 634.33 New York 39.1 30.3 1528.81 918.09 1184.73 TOTAL 322.9 259.3 10080.49 6530.51 8089.12
Non-Parametric conclusion: r=0.952 so there is a strong correlation between % workers in management/tech and %
Parametric conclusion: 0.952>0.602 so reject null hypothesis and conclude that there is a linear relationship between % workers in management/tech an df 9 critical value (from table) 0.6
20.45
% adults with college degrees
F
35 30 25 20 15 10 10 15
rkers in management/tech and % persons age 25+ with college degrees degrees of freedom = n-2, where n= # of pairs