Tables 15.A to 15.

D below summarize the results of data analyses of research conducted in a

a. Interpret the information contained in each of the tables in as much detail as possible
b. Summarize the results for the CEO of company
c. Make recommendations based on your interpretation of the results

TABLE 15.A
Means, standard deviations, minimum, and maximum
Variable Mean Std. Deviation Minimum Maximum
Sales (in 1000s of $) 75.1 8.6 45.2 97.3
No of salespersons 25 6 5 50
Population (in 100s) 5.1 0.8 2.78 7.12
Per capita income (in 1000s of $) 20.3 20.1 10.1 75.9
Advertising (in 1000s of $) 10.3 5.2 6.1 15.7

TABLE 15A provides data about the sales of the company. In reviewing the data provided by
the company, the independent and dependent variables will need to be labeled. The dependent
variable (DV) is the company sales. The independent variables (IVs) are number of sales
persons, population, income per capita, and advertising. While the sampling method was not
disclosed, it appears that the sample size of 150 may be too small. According to (Roscoe 2,
1975), as a rule of thumb, with a population of 500, the sample size should be over 260. In
analyzing the statistics provided for the DV sales, the range in company sales was from 97.3 to
45.2 which represent a wide spread. The mean of company sales was 75.1 which appear
reasonable in these circumstances when considering that the mean of the minimum and
maximum is 71.25 and the standard deviation (SD) is 8.6. (Sekaran and Bougie, 2009) stated
that the standard deviation offers an index of the spread of a distribution or the variability in the
data and it is used as measure of dispersion and is the square root of the variance. The mean and
standard deviation are the most descriptive statistics for interval and ratio scaled data. The SD,
in conjunction with the mean, is a very useful tool because of the following statistical rules, in a
normal distribution (Sekaran and Bougie, 2009):
1. Practically all observations fall within three standard deviations of the average or the mean;

2. More than 90% of the observations are within two standard deviations of the mean; and

3. More than half of the observations are within one standard deviations of the mean. This would
to the believe that there is a normal distribution and therefore each city (the unit of analysis) is
producing sales of between $66.5 and $83.7 thousand based on the standard deviation of 8.6. As
for the IVs, population appears to be normally distributed with a mean of 5.1 hundred thousand
individuals and a SD of 0.8. The IV of number of salespersons could also be a normal
distribution. Even though there is a high range, the mean of the minimum and maximum come
close to the overall mean. The SD is also close to the mean. However, the IVs of per capita
income and advertising are skewed. Income per capita has a very high range and the SD is not
close at all to the mean. In other words, there is a tremendous variance in the wealth of the cities
that must be taken into consideration in the proportion of sales by city. As for advertising, while
the range would appear reasonable, the SD varies a great deal indicating that this is a
disproportionate relationship. In summary, the data contained in Table 12A indicates the
following:

Sales could be affected by the placement of salespersons in cities, the company should focus on
cities with high per capita income, and the company should invest advertising funding into those
cities with the least sales. These changes will be beneficial to the company.

TABLE 15.B
Correlations among the variables
Sales Salesperson Population Income Ad.expenditure
s
Sales 1.0
No of salespersons 0.76 1.0
Population 0.62 0.06 1.0
Income 0.56 0.21 0.11 1.0
Ad.expenditure 0.68 0.16 0.36 0.23 1.0
All figures above 0.15 are significant at p = 0.05
All figures above 0.35 are significant at p ≤ 0.001

TABLE 15B contains the data of the correlation between the different variables in the study. The
correlation is derived by assessing the variations in one variable as another variable also varies
(Sekaran andBougie, 2009). Correlation measures the strength of the linear relationship between
the variables. The stronger the correlation the better the independent variable predicts the
dependent variable. Correlation values range from -1.0 to +1.0 indicating perfect positive
correlation at the +1.0 level to perfect negative correlation at -1.0, with zero indicating no
correlation (Investopedia, 2011). As we know, a significance ofp=0.05 is the generally accepted
conventional level of social research (Sekaran and Bougie, 2009). This indicates that 95 % out of
100, we can be sure that there is a true or significant correlation between the two variables, and
there is only 5% chance that the relationship does not exist (Sekaran and Bougie, 2009). In Table
12B, we know that all figures above 0.15 have a 95% confidence that the correlation between the
variables exists. Also, we know that all figures above 0.35 have a 99.9% for confidence that the
correlation between the variables exists. The independent variables are number of salespersons,
population, per capita income and advertising dollars. The dependent variable is amount of sales.
As expected, there is a strongly significant relationship between number of salespersons and
amount of sales. The correlation between number of salespersons and amount of sales is
significant at 0.76 out of 1.00 with a confidence interval of 99.9%. This relationship suggests
that as the number of salespersons increase, so does the amount of sales increase which indicated
a strong correlation between the variables. There is a medium correlation between population
and amount of sales that is significant at 0.62 out of 1.00, suggesting that sales are better with a
somewhat larger population as potential customers. The relationship between population and
amount of sales is also based on a confidence interval of99.9%. There is no significant
relationship between population and number of sales persons based on a correlation of .06 out of
1.00. The relationship between per capita income and amount of sales is moderately significant
at 0.56 out of 1.00 with a confidence interval of 99.9%. The correlation is low but still significant
between per capita income and number of salespersons at 0.21 out of 1.00 with a 95%
confidence interval. There is no significant relationship between per capita income and
population at 0.11 out of 1.00. The correlations for advertising dollars and other variables are
mixed. There is a somewhat strong and significant relationship between advertising dollars and
amount of sales at 0.68 out of 1.00 with a 99.9% confidence interval. This is not surprising and
suggests that advertising is a definite contributor to sales. We can say with 95% confidence, that
there is a slightly significant relationship between advertising dollars and number of salespersons
(0.16 out of 1.00). There is a mild but significant correlation between advertising dollars and
population at 5.36 out of 1.00 with a confidence interval of 95%. The statistics also reveal a low
but significant correlation between advertising dollars and per capital income at 0.23 out of 1.00
with a 95% confidence interval.
TABLE 15.C
Results of one-way ANOVA: sales by level of education
Source of variation Sums of squares df Mean square F Significant of F
Between groups 50.7 4 12.7 3.6 0.01
Within groups 501.8 145 3.5
Total 552.5 150

TABLE 15 C Source of variation between groups within groups Total Results of one-way
ANOVA: sales by level of education. Sums of squares 50.7 501.8 552.5 Degree of Freedom 4
145 150 Mean squares 12.7 3.5 F 3.6 Significant of F 0.01 Table 12C provided the results by
using a one-way ANOVA test to study sales by level of education. We would like to point out
that table 12C needs to be updated to resolve a calculation error on the total degrees of freedom
which should be 149 not 150. The ANOVA examines significant mean differences among more
than two groups on an interval DV. The DV in this study was sales and the IV was level of
education. The ANOVA formula (which is a ratio) compares the amount of variability between
groups (which is due to the grouping factor) to the amount of variability within groups (which is
due to chance). In other words, the within group difference is equal to the amount of variability
due to between group differences and any difference between groups will not be significant. The
F ratio is the ratio of variability between groups to variability within groups. To determine the
values it is necessary to compute the sum of squares for each source of variability between
groups, within groups, and the total. The between groups sum of squares equals the sum of
differences between the mean of all scores and the mean of each groups score, which is then
squared. The within group sum of squares is equal to the sum of differences between each
individual score in a group and the mean of each group, which is then squared. The total sum of
squares is equal to the sum of the between group and the 6 within group sum of squares (Mertler
and Vannatta, 2005). Since there are four degrees of freedom (df) between groups, the number
of groups would be five as follows:
4dfs= [5 groups (or levels of education) -1] x [3 categories of interest -1].
The F statistic can be thought of as a measure of how different the means are relative to the
variability within each sample (Levin & Stephan, 2004). Computing the F statistic requires
dividing each sum of squares by the degrees of freedom (which are in approximation to the
sample or group size), and then dividing the resulting mean sum of squares due to between group
differences by the mean sum of squares due to within group differences (Salkind, 2011). The F
statistic for table 12C was obtained by the dividing the between group of 12.7 by the within
group of 3.5 which provided the result of 3.6. The obtained value or F statistic of 3.6 is then
compared to the critical value (CV) which is the value needed for rejection of the null
hypothesis. The CV in this case is 3.32 as determined from the appropriate table. We may
therefore conclude that the null hypothesis may be rejected and that sales are indeed affected by
level of education. We can say this with 99% confidence since the significance of F is listed
as .01. TABLE 12 D Results of regression analysis. Model Summary Value 1 Multiple R
0.65924 R square 0.43459 Adjusted R square 0.35225 Standard error 0.41173 df (5.144) F 5.278
Sig 0.000 Variable Beta Training of salespersons 0.28 No. of Salespersons 0.34 Population 0.09
Per capita income 0.12 Advertisement 0.47 Value 2 Value 3 t 2.768 3.55 0.97 1.200 4.54 Sig. t
0.0092 0.00001 0.467 0.089 0.000017.
TABLE 15.D
Results of regression analysis
Multiple R 0.65924
R-square 0.43459
Adjusted R-square 0.35225
Standard error 0.41173
df (5.144)
F 5.278
Sig 0.000
Variable Beta t sig.t
Training of salesperson 0.28 2.768 0.0092
No. of salespersons 0.34 3.55 0.00001
Population 0.09 0.97 0.467
Per capita income 0.12 1.200 0.089
Advertisement 0.47 4.54 0.00001

TABLE 12D is the results of a regression analysis. The IVs of salespersons, population, per
capita income, and advertising were all significant to different degrees. When interpreting the
regression analysis, the multiple correlation or R of .65924 indicates that the combination of IVs
predicts the DV of sales and shows a good fit between the predicted and actual scores of the DV.
The squared multiple correlation (R squared) of .43459 tells us that the percentage of variance of
over 43% gives us a degree of "goodness of fit" that indicates that it is likely that there are more
IVs that could contribute to the research study. The adjusted squared multiple correlation (R
squared adjusted) of .35225 indicates that there was an overestimate of the Rand R squared
population. There is evidence of bias possibly caused by the small sample size. The ANOVA
generated an F test of 5.278 assuming a level of significance of 0.0. This is much lower than
would be expected in a linear relationship. The results should be in the range of 20.5 to 22.5. For
each individual IV, in order to see if that IV is significantly affecting the DV in the regression
model, the Beta test was run and at value generated. In running the Beta test, each IV had a
positive value, indicating that there is a positive change in the DV when the IV increases. There
are, however, different degrees of change among the IVs. For instance, population and per capita
income have low Beta test results. The t and p values are used to indicate the significance of the
beta weights applied. The results of the t tests showed that for a sample size of 150 and a t score
threshold of 2.768, the IVs of training of salespersons (t value of 2.768), number of salespersons
(t value of3.55), and advertisement (t value of 4.54) were above the t value threshold required to
be compliant with the probability of error which is approximately p < .01 and therefore showing
no significant difference from the 0. On the other hand, the t value results for population of 0.97
and per capita income of 1.2 were below the threshold for error. It should be noted that even
though the significant t value for population was higher at 0.467, the t value remained below the
required threshold and this difference was therefore significant. Executive Summary The
consulting team collected sales data across 50 cities and sampled 150 sales persons from the 500
member sales force. Sales was measured in 1000 s dollars, populations was of measured in
100,000s people, per capita income was measured in 1,000s dollars, advertising of was measured
in 1,000s dollars. Education level and training of salespersons was also of evaluated. The overall
results indicate that number of salespersons, education levels of sales persons, training of
salespersons, and advertising dollars are the biggest contributors to amount of sales. The
evidence for this is demonstrated across three different statistical tests: correlation, analysis of
variance, and regression analysis. Not surprisingly, correlation tests indicate there is some
evidence that a larger population provides more potential for larger sales but income does not
appear to play a major part based on the current analysis.

Recommendations

The research team recommends that the company invest in number of salespersons and training
of salespersons and also review and increase advertising dollars in those areas that are lacking. It
is also recommended that further analysis is done regarding education levels of salespersons to
better understand the educational threshold. Finally, it would be useful to create additional
regression models to find a better fit model and increase understanding of the factors that best
contribute and in what proportion to positively affect amount of sales.