PART A
QUESTION 1
Table 1 shows that the mean age of respondents who are purchase decision-makers is
39.21 years, with a standard deviation of 10.47 years.
The 95% confidence interval based on the mean age of the population of purchase
decision makers is calculated as below:
S N −n
Y ± 1.96
n N −1
Conclusion:
We are 95% confident that the mean age of respondents who are purchase decision-
makers is between 40.60 and 37.82 years old.
QUESTION 2
Table 2(a) shows that the mean of the income of the male decision-makers is
AUD$36,660 with a standard deviation of AUD$11,902.
Table 2(b) shows that the mean of the income of the female decision-makers is
AUD$37,585 with a standard deviation of AUD$12,295.
= 8.214
Conclusion:
We are 95% confident that the mean income of all purchase decision-makers in the
population is between AUD$355.32 and AUD$387.5.
QUESTION 3
Table 3 shows that the proportion of female respondents that had a positive response to
the attitudinal question is 0.33.
The 95% confidence interval for proportion of positive attitude Female respondents is
calculated as:
p (1 − p ) N −n
p ±1.96
n N −1
Conclusion:
We are 95% confident that the proportion of female respondents that have a positive
response to the attitude question is between 24.4% and 41.6%.
QUESTION 4
H0: Positive, neutral and negative responses are equally likely amongst the population of
part-time workers.
HA: At least one of the responses is not equally likely amongst the population of part-time
workers.
Conclusion:
Based on the sample of part-time workers, we reject the null hypothesis at the 0.05 level
of significance that positive, neutral and negative responses are equally likely amongst
the population of part-time workers. Hence, we conclude that at least one of the responses
is not equally likely amongst the population of part-time workers.
Question 5
H0: Attitudes are independent of promotion methods that the decision makers in the
population are exposed to.
HA: Attitudes are dependant on the promotion methods that the decision makers in the
population are exposed to.
• The null hypothesis should be rejected because the significance of the chi-square
test is 0.010, which happens to be smaller the level of significance for the test
which is 0.05
• The significance value of 0.010 implies that there is strong evidence to reject the
null hypothesis.
Conclusion:
Based on the sample of 200 purchase decision makers, at 0.05 level of significance, we
can come to the conclusion that attitudes are dependent on promotion methods that the
purchasers are being exposed to.
Question 6
H0: Distributions of purchases in the month before being exposed to the advertising
campaign are the same for both males are females in the population of purchase decision
makers.
HA: Distributions of purchase in the month before being exposed to the advertising
campaign are not the same for both males and females in the population of the purchase
decision makers.
Table 6: (a)
Gender N Mean Rank Sum of Ranks
Before Male 94 112.10 10537.00
Female 106 90.22 9563.00
Total 200
• The null hypothesis should be rejected because the significance of the wilcoxon
Rank Sum Test is 0.008, which happens to be smaller than the level of
significance for the test which is 0.05.
• The significance value of 0.08 indicates that there is sufficient evidence in order
to reject the null hypothesis.
Conclusion:
Based on the sample data of the purchase decision makers, at the level of significance of
0.05 we can conclude that the distributions of purchase in the month of being exposed to
the advertising campaign are not the same for males and females.
Question 7
H0: Purchase in the month before the advertising exposure is the same as the month after
for the population of purchase decision makers.
HA: Purchase in the month before the advertising exposure is not the same as the month
after for the population of purchase decision makers.
Table 7(a): Wilcoxon Signed Rank Test of Purchases before and after Advertising
Exposure
N Mean Rank Sum of Ranks
After - Before Negative Ranks 115(a) 76.47 8794.00
Positive Ranks 79(b) 128.11 10121.00
Ties 6(c)
Total 200
Table 7(b): Wilcoxon Signed Rank Test of Purchase before and after Advertising
Exposure
After - Before
Z -.848(a)
Asymp. Sig. (2-tailed) .396
Conclusion:
Based on the sample data of the purchase decision makers, at a level of significance of
0.05 we can come to the conclusion that for the purchase decision makers, purchase in
the month before the advertising exposure is not the same as in the month after.
Part B
Question 8
H0: Means sales are the same in all three exhibition locations
HA: Atleast one means sales is not the same in the three exhibition locations.
Table 8: 1-way ANOVA Test of means sales in all three exhibition Locations.
Sales ($100)
Sum of df Mean Square F Sig.
Squares
Between Groups 874810633. 2 437405316.849 51.928 .000
698
Within Groups 817065641. 97 8423357.127
303
Total 169187627 99
5.000
• The significance value of the 1-way ANOVA is 0.000 which is smaller than the
level of significance for the test which is 0.05.
• The significance value of 0.000 indicates that there is sufficient evidence to reject
the null hypothesis.
Conclusion:
Based on the results above, at the significance level of 0.005, we can come to the
conclusion that means sales in all three exhibition locations are not the same.
Question 9
Table 9:Tukey Confidence Interval For mean differences between Sales Locations
Dependent Variable: Sales ($100)
(I) (J) Mean Std. Sig. 95% Confidence Interval
Location Location Difference Error Lower Upper
(I-J)
New York London 6287.414(*) 713.048 .000 4590.20 7984.63
New York Tokyo 6625.727(*) 732.142 .000 4883.07 8368.39
London Tokyo 338.313 694.918 .878 -1315.75 1992.37
* The mean difference is significant at the .05 level.
• There is a significant difference in the mean sales in the pairs of display locations
New York-London and New York-Tokyo.
• The difference in mean sales between New York-London and New York-Tokyo
as can be seen from the level of significance which is 0.000 which is less than
0.05, the level of significance for this test.
• The difference in mean sales of the different pairs of display locations New York-
London and New York-Tokyo are significant as the level of significance is 0.000
which is less than the level of significance of 0.05.
• We are 95% confident that the average sales in London exceed sales in New York
by between $4590.20 and $7984.63. The mean sales in Tokyo exceed sales in
New York by between $4883.07 and $8368.39.
Conclusion:
Based on the sample data, at the level of significance of 0.05, we can conclude that the
means sales are not all the same in all three exhibition locations.
QUESTION 11
Table 11(a): Tukey Confidence Intervals for Mean Difference between Sales
Locations.
Location (I) Location (J) Mean Std. Error Sig. 95% Confidence Interval
Difference Upper Lower
(I-J) Bound Bound
• The pairs of display locations New York-London and New York-Tokyo have
significant difference in their mean sales.
• The difference in average sales of the different pairs of display locations New
York-London and New York-Tokyo are significant as the level of significance is
0.000 which is less than the level of significance of 0.05.
• We are 95% confident that the average sales in London exceed sales in New York
by between 4653.1 and 7921.7 units. The average sales in Tokyo exceed sales in
New York by between 4947.7 and 8303.8 units.
Conclusion:
Based on the observations, it can be concluded that there is a significant level of
difference between the mean sales in the pairs of display locations New York-London
and also Tokyo-New York.
Table 11(b): Tukey Confidence Intervals for Mean Differences between Artists.
(I) Artist (J) Artist Mean Std. Error Sig. 95% Confidence Interval
Difference Upper Lower
(I-J) Bound Bound
Picasso Monet 4354.05(*) 800.427 .000 2258.34 6449.76
Picasso Cezane 3623.83(*) 844.964 .000 1411.51 5836.15
Picasso Modiaglini 470.10 837.247 .943 -1722.01 2662.21
Cezane Monet 730.22 758.800 .771 -1256.51 2716.94
Modiaglini Monet 3883.95(*) 750.197 .000 1919.75 5848.15
Modiaglini Cezane 3153.73(*) 797.544 .001 1065.57 5241.90
Conclusion:
We can conclude that at the significance level of 0.05, at least one
promotion strategy has a different mean sale for the showroom.
(b) Effect of Cities:
• We can observe that the significance level for each of the blocks (cities) is
0.000 which is smaller than 0.05 therefore we will reject the 2nd
Hypotheses thus it shows that there is sufficient evidence to reject the null
hypotheses.
Conclusion:
We can conclude that at the significance level of 0.05, at least one of the
cities has a different mean sale upon looking at the 100 sample
observations.
Question 13
Table 13 (a): Tukey Confidence Intervals for Mean Sales Difference between
Treatments
Mean 95% Confidence Interval
Difference Std. Upper Lower
(I) Promotion (J) Promotion (I-J) Error Sig. Bound Bound
Leaflet Internet Websites -1302.46 841.017 .413 -3502.22 897.29
Distributions
Leaflet International Art
1072.20 848.299 .588 -1146.60 3291.00
Distributions Magazines
Leaflet Newspapers
-1247.97 821.886 .431 -3397.68 901.75
Distributions
Internet Websites International Art
2374.66(*) 792.056 .018 302.97 4446.36
Magazines
Internet Websites Newspapers
54.50 763.701 1.000 -1943.03 2052.02
International Art Newspapers -
771.713 .018 -4338.65 -301.68
Magazines 2320.17(*)
The above table shows the difference between treatments for mean sales, and the results
obtained indicates the following:
• The significance level for Leaflet distributions-Internet Websites is 0.413, the
significance level for Leaflet distributions-International Art Magazines is 0.588
and the significance level for Leaflet distributions-Newspapers is 0.431.
Conclusion:
The results show that the significance values for the promotional strategies are all
greater than 0.05 therefore the mean sales for the promotion strategies are all the
same.
Table 13(b): Tukey Confidence Intervals for Mean Sales Difference between Cities
Mean 95% Confidence Interval
Difference Upper Lower
(I) Location (J) Location (I-J) Std. Error Sig. Bound Bound
New York London 6287.41(*) 694.710 .000 4633.03 7941.80
New York Tokyo 6625.73(*) 713.313 .000 4927.04 8324.41
London Tokyo 338.31 677.047 .872 -1274.01 1950.63
The above table shows the difference between cities for mean sales, and the results
obtained indicates the following:
• New York-London and New York-Tokyo both have a significant level of 0.000
which shows that it is smaller than the significance level of 0.05 thus indicating
that the pair cities have different mean sales.
• London-Tokyo however has a significance level of 0.872 which is higher than the
significance level of 0.05 thus resulting that London-Tokyo have the same mean
sales.
Conclusion:
We can conclude that New York-London and New York-Tokyo have different
mean sales while London-Tokyo has the same mean sales.
Question 14
H0: The distribution of sales are the same in all three display locations
HA: The distribution of sales for at least one display location is different.
Table 14(a): Kruskall-Wallis test for all three display Locations
Location N Mean Rank
Sales(AUD$’00) New York 30 77.37
London 37 41.23
Tokyo 33 36.47
Total 100
Table 14(b):
Sales
Chi-Square 37.232
df 2
Asymp. Sig. .000
a Kruskal Wallis Test
b Grouping Variable: Location
Conclusion:
Question 15
Table 15 : Summary of sales data based on location
N Valid 100
Missing 0
Mean 7503.50
Std. Deviation 4133.964
Skewness 1.328
Std. Error of Skewness .241
Kurtosis .814
Std. Error of Kurtosis .478
Histogram
25
20
Frequency
15
10
Mean =7503.5
Std. Dev. =4133.964
N =100
The results
0 from the above data show the following:
0 5000 10000 15000 20000
• The skewness statistic is 1.328 which
Sales is larger than 0 thus indicating that the data
($100)
is positively skewed.
• The Kurtosis statistic is 0.814 which is larger than 0 thus indicating that the data
has a higher peak than a normal curve.
• The histogram appears to have a bell-shape
Conclusion:
• The sample size is 100 thus it is not appropriate making the 1-way
ANOVA test to appear risky as the data is not normally distributed.
• Because the data appears to be not normal, it would be advisable to use the
Kruskall-Wallis test rather than using 1-way ANOVA test.