In: Statistics and Probability
You have been asked by a men's clothing manufacturer, Nuke, to provide advice on whether there is evidence of differences in the average heights of male citizens in three different countries. You have taken random samples of the heights of males (in cms) in those three countries and your data set is attached in csv delimited format (as a .txt file) using the link on the right.
Input the data set into SPSS and perform the appropriate analysis to answer the question above (marked in bold). Follow SPSS instructions to set up the data in the correct format. (You will need to create a new quantitative variable - with four decimals - with all the heights and a new quantitative variable with values 1, 2 or 3 depending on whether the men are from country 1, 2 or 3. Both of those variables have to be defined as "numeric" in SPSS.)
Choose the correct answer to the following questions, based on your results:
a) Based on the outputs from SPSS, and looking at the appropriate table, can the variances be pooled based on this data set?
Yes, because 1.519 is smaller than 2.257
Yes, because 1.735 is smaller than 2.885
Yes, because 2.226 is smaller than 2.885
No, because all standard deviations for the data set are different
b) Possible null and alternative hypothesis for the ANOVA test could include:
i)Ho: X1=X2=X3, where 1,2 and 3
refer to country 1, country 2 and country 3.
Ha: not all sample means are the same.
ii)Ho: μ1=μ2=μ3, where 1,2 and
3 refer to country 1, country 2 and country 3.
Ha: not all sample means are the same.
iii)Ho: μ1=μ2=μ3, where 1,2 and
3 refer to country 1, country 2 and country 3.
Ha: not all population means are the same.
iv)Ho: μ1=μ2=μ3, where 1,2 and
3 refer to country 1, country 2 and country 3.
Ha: all population means are different.
The correct null and alternative hypothesis for the ANOVA test are: ?
c) Give the value of the test statistic to three decimal places:
d) Based on the ANOVA test, at the 1% significance level, we can reject/fail to reject/accept the null hypothesis.
e) Based on the ANOVA test, at the 1% significance level, we can conclude:
That the average heights of adult males in the three countries are all the same.
That the average heights of adult males in the three countries are not all the same.
That the average heights of adult males in the three countries are all different.
f) At the 1% significance level:
The average height of adult males in country 1 is not significantly different from the average height of adult males in country 2.
The average height of adult males in country 1 is not significantly different from the average height of adult males in country 3.
The average height of adult males in country 1 is significantly different from the average height of adult males in country 2 and in country 3.
Data:
Country1,Country2,Country3 173.9065,174.4416,173.3905 175.8508,172.408,171.8004 169.7117,176.3606,170.7854 174.3933,173.768,172.0378 173.5781,175.2501,171.3748 171.1385,174.2832,169.5553 172.4496,175.5553,172.5238 173.6139,174.1995,171.5606 178.4676,175.8349,171.3028 177.2542,176.209,171.8729 171.0752,177.6678,171.4921 177.6524,174.8088,169.6878 174.1881,171.8925,170.8565 173.0054,173.8406,170.8809 174.1721,177.1319,170.302 172.7926,173.4918,173.6232 172.9138,176.5414,169.7864 175.3345,175.2861,170.3743 175.2136,177.255,170.032 175.2258,172.1587,169.3387 174.1072,174.8035,170.8116 171.2888,173.2882,170.6889 174.1759,179.462,173.3951 175.5454,176.3378,170.7265 173.8333,177.1685,169.5037 174.652,173.5127,173.5052 174.1903,174.3971,172.952 172.6448,174.6913,170.7556 173.5408,176.7476,168.8408 171.9191,174.6832,170.4331 174.4326,176.1523,170.8661 171.3794,172.0223,171.5141 171.4967,174.5692,170.7083 171.8858,173.8646,171.7651 168.6836,172.7344,171.6878 175.2576,175.862,169.224 173.5878,175.523,169.6781 171.9676,175.1502,169.9883 175.1554,173.0995,170.3383 170.5327,176.7912,170.6191 172.9466,175.6253,171.1187 172.7378,174.6514,166.5562 173.5788,175.1343,170.4145 173.5693,174.707,172.9637 171.8027,172.4747,169.4999 173.0549,174.6715,172.5006 172.8527,173.853,171.6255 174.0416,173.6312,171.0565 174.7399,173.3654,171.3737 174.7639,174.2997,168.7524 171.8045,172.096,170.9732 173.216,176.5463,173.5059 171.2788,175.8801,171.2475 171.4297,175.07,171.1621 173.0897,175.0478,169.9987 175.3989,173.9028,171.0538 171.9455,176.628,171.4485 173.6571,174.9002,171.7396 172.7616,174.0282,170.5408 174.776,177.1271,170.7453 171.4664,174.7628,174.1355 173.1488,174.2165,167.7125 173.9288,174.6594,174.4442 174.7509,173.8281,171.6063 175.4163,173.4198,172.6001 173.2289,178.889,168.6038 170.8626,177.5832,170.2149 171.9865,175.5613,170.6829 171.5076,173.2143,171.7341 176.6257,173.8018,168.5947 172.1766,174.8352,171.8075 174.2221,176.2871,169.2807 172.8114,173.102,171.1993 174.4329,171.6052,172.0785 171.9527,172.9264,171.5906 170.9966,175.6003,172.724 170.9664,175.687,172.6091 173.8323,175.7775,170.1236 172.8339,174.9046,171.4856 172.8059,175.3755,169.6834 175.229,174.3858,169.1173 173.5374,176.393,172.4872 173.3967,173.0575,171.1001 175.4815,175.7825,171.0176 171.8933,173.8269,172.4667 174.1449,174.5977,171.9919 174.3526,175.9292,171.6253 172.7344,176.6586,172.9754 173.4235,173.4235,172.4947 171.3512,176.991,171.4596 171.3781,176.0902,170.0645 173.2573,174.9982,170.1227 174.1834,174.8072,172.8882 176.9782,174.7736,168.6823 172.0997,174.6453,171.0633 173.381,175.1346,168.1767 172.9763,175.1769,172.6307 170.2005,176.3391,172.3926
Let denote the mean heights of the adult males in the countries 1,2 and 3.
To test: Vs Not all means are equal
Using SPSS, Running a One way ANOVA,
We get the output:
a) Based on the output of Levene's test for homogeneity of variance, we test:
Variances are homogeneous and hence, can be pooled. Vs Ha: Variances are not homogeneous and hence, cannot be pooled
The levene's test statistic 1.521 is less than the critical value (at alpha = 0.01) 2.257.
a) The variances be pooled based on this data set because 1.519 is smaller than 2.257.
b) As mentioned above, the null and alternative hypothesis for the ANOVA test could include:
iii)Ho: μ1=μ2=μ3, where 1,2 and
3 refer to country 1, country 2 and country 3.
Ha: not all population means are the same.
c) From the output, the value of the test statistic is F = 148.797
d) Based on the ANOVA test, at the 1% significance level, since the p-value of the test 0.001 < 0.01, the test is found to be significant. We may reject the null hypothesis.
e) Based on the ANOVA test, at the 1% significance level, we can conclude:
That the average heights of adult males in the three countries are not all the same.
To find out which pair of means differ, we may run a post hoc test:
f) At the 1% significance level:
The average height of adult males in country 1 is significantly different from the average height of adult males in country 2 and in country 3.
In fact, all three means differ from each other. No pair of population means are equal based on this data.