Case Study Questions: Unit 3

Case 1
A retail chain aims to improve customer satisfaction and purchasing behaviour. To guide its
decisions, a researcher designs a survey and collects responses from 200 customers across
three outlets. The dataset includes variables such as age, gender, satisfaction score (rated 1 to
10), and purchase decision (Yes/No). Using SPSS, the researcher begins by generating a
frequency table for satisfaction scores to understand general trends. Next, a cross-tabulation
is created to observe if gender affects purchase behaviour. To statistically validate the pattern,
the researcher performs a Chi-square test to test the association between gender and purchase
decision. The test produces a p-value of 0.03. The management wants to know whether this
result is meaningful enough to adjust their marketing strategy based on gender-specific
preferences. Based on following observations from attached data table answer following
questions:​
Data Table:

Purchase Decision:
Gender Purchase Decision: No Total
Yes
Male 48 22 70
Female 66 64 130
Total 114 86 200

Based on the case scenario and data table, explain how a frequency table helps summarize
customer satisfaction scores. What patterns would a research analyst look for when
interpreting such data?
Answer:​
A frequency table provides a clear summary of how often each satisfaction score appears. It
helps the analyst identify the most frequent scores (mode), detect whether scores are
concentrated around certain values, and assess the spread of responses. By examining these
patterns, the analyst can judge if satisfaction is generally high or low, whether responses are
skewed, and where interventions may be needed. The table also aids in segmenting customers
based on satisfaction levels for targeted improvements.
Using the data table below, identify the dependent and independent variables. Describe what
type of graph would best show the relationship between gender and purchase decision. Sketch
it and interpret whether gender influences purchasing behaviour.
Answer:​
Independent variable: Gender
Dependent variable: Purchase Decision
A grouped bar chart is suitable. Based on the table, more males said "Yes" while females are
split more evenly. This pattern suggests a possible gender influence, which analysts would
explore further with significance testing in a real-world setting.
Case 2
An education researcher investigates whether a new teaching method impacts student
performance. A sample of 30 students attends a workshop using the new method, while
another 30-follow traditional teaching. Test scores from both groups are collected. The
researcher uses SPSS to check data distribution, then runs an Independent Samples T-Test to
compare means. Levene’s Test indicates equal variances, and the p-value from the T-Test is
0.001. The researcher also calculates Cohen’s d to measure the effect size and obtains a value
of 0.75. School administrators ask whether this difference is both statistically and practically
meaningful before revising their teaching policies. Provide Interpretations based on following
table

Group Mean Test Score Standard Deviation N

In the context of the data below, explain how a research analyst would assess whether a new
teaching method impacts student performance. What basic indicators should be evaluated?​
Answer:​
The analyst would compare group means, examine variation using standard deviation, and
look for a meaningful gap in scores. These indicators suggest whether performance differs
across groups. The mean gap gives direction; variation informs reliability.
Identify the dependent and independent variables. Choose and sketch an appropriate graph
for visualizing the performance difference. Interpret whether the new method had an impact.​
Answer:​
Independent variable: Teaching Method
Dependent variable: Test Score
A bar chart comparing mean scores shows the new method group scored higher. With a
10-point mean difference, this likely indicates a positive impact, assuming variation is not
extreme.
Case Study 3
Scenario
A health organization surveys 150 individuals to assess exercise frequency and health
condition (healthy or not). The data includes age, gender, and number of exercise sessions per
week. The analyst uses SPSS to summarize exercise frequency with descriptive statistics and
plots a histogram to check the data distribution. A boxplot is also created to detect outliers.
The organization wants to understand if most people exercise enough and whether the data
distribution supports targeted health interventions (see table below)

Exercise Sessions per Week Frequency

0 10

1 18

2 35

3 40

4 25

5 12

6 6

7 4

Explain what kind of insights a research analyst can gain from analysing the frequency of
exercise sessions per week. How would this inform health interventions?​
Answer:​
An analyst would look for common activity levels, distribution shape, and presence of
inactivity or outliers. A large cluster at moderate levels may suggest acceptable behaviour,
while a spike at zero indicates the need for targeted awareness.
Based on the data, identify the variable of interest. Sketch a histogram showing the
distribution and explain the key trend. What might this suggest about community exercise
patterns?​
Answer:​
Variable: Sessions per Week
The histogram would peak at 2–3 sessions, then taper. This suggests many exercises
moderately but few exercise daily. Awareness campaigns might target increasing consistency.
Basically you need to draw a histogram and interpret it.
Case Study 4
Scenario
A bank wants to know if customer satisfaction differs based on service channels—online,
branch, or mobile app. A total of 90 customers are surveyed, 30 from each channel.
Satisfaction is rated from 1 to 10. The analyst uses One-Way ANOVA in SPSS to compare
mean satisfaction scores across groups. The output gives an F-ratio and p-value. A post hoc
test identifies which specific groups differ. Management seeks evidence before reallocating
resources among service channels.
 Service Channel Mean Satisfaction Std. Deviation N
Score
Online 8.2 1.1 30
Branch 6.7 1.3 30
Mobile App 7.5 1.2 30

As a research analyst, how would you evaluate whether differences in customer satisfaction
across service channels are statistically meaningful? Refer to the table and explain which test
is appropriate and why.
Answer:
To compare satisfaction scores across three service channels, an analyst would use a
One-Way ANOVA. This test is suitable when comparing the means of more than two
independent groups. ANOVA helps determine if observed differences in average satisfaction
scores are statistically significant or due to chance. From the table, online has the highest
mean, followed by Mobile App and then Branch. However, without testing statistical
significance, these differences may be misleading. For example, calculating the F-statistic
using the data below can help assess whether the variation in satisfaction scores is statistically
significant. Using the ANOVA formula:
F = Variance between group means / Variance within groups
Given the means, standard deviations, and equal sample sizes (N = 30), students can estimate
the sum of squares and degrees of freedom to compute the F-value. They can then compare
the calculated F-value to the critical value from the F-distribution table for df1 = 2 and df2 =
87 to assess statistical significance.
df1 = 2: This is the degree of freedom between groups. Since there are three service channels
(Online, Branch, Mobile App), the degrees of freedom is k − 1 = 3 − 1 = 2.
df2 = 87: This is the degree of freedom within groups. Each group has 30 samples, so the
total sample size is 90. Degrees of freedom is N − k = 90 − 3 = 87.
By calculating the F-statistic and comparing it to the critical value for F (2, 87) at a chosen
significance level (e.g., 0.05), determine whether to reject the null hypothesis that all group
means are equal. This process helps determine if the observed differences in satisfaction
scores across service channels are statistically meaningful.
What should management infer if ANOVA results show p = 0.04 and the post hoc test shows
online > branch?
Answer:​
A p-value of 0.04 indicates a statistically significant difference in satisfaction between at least
two channels. The post hoc result shows that online service scores higher than branch service.
This suggests customers are more satisfied online. Management can consider reallocating
resources or improving in-branch services. However, decisions should also consider effect
size and other factors such as cost and accessibility before implementation.
Case Study 5
Scenario
A university evaluates the effectiveness of a training program by comparing test scores
before and after the program for 40 students. The researcher uses SPSS to perform a Paired
Samples T-Test. The results show a mean difference of +8 points, with a p-value of 0.002.
The university is interested in understanding if the program had a meaningful impact on
student performance and whether this difference is statistically reliable for decision-making.
How is the Paired Samples T-Test used in SPSS for this evaluation?
Answer:​
In SPSS, the researcher selects Analyze > Compare Means > Paired-Samples T Test. The two
variables—pre-test and post-test scores—are entered as a pair. The test checks if the mean
difference between the two sets of scores is statistically significant. A p-value is calculated. If
it is below 0.05, the difference is unlikely due to chance. This method is suitable for
comparing the same group before and after an intervention.
What does a p-value of 0.002 suggest about the program’s impact on test scores?
Answer:​
A p-value of 0.002 indicates a statistically significant difference in test scores before and
after the program. The null hypothesis is rejected, meaning the score increase is unlikely due
to random variation. The program appears to have had a real effect on performance. Since the
result is statistically strong, the university may consider retaining or expanding the training,
especially if the change aligns with educational objectives and resource availability.
Case Study 6
Scenario
A marketing analyst investigates whether product preference is associated with age group.
Data from 200 respondents includes age group (18–25, 26–35, 36–50) and product choice
(Product A, B, or C). The analyst runs a cross-tabulation and Chi-square test in SPSS. The
Chi-square output shows p = 0.048. Management wants to know if this relationship is
meaningful and how to use the result in planning targeted marketing campaigns.
What does a cross-tabulation in SPSS show in this scenario?
Answer:​
Cross-tabulation in SPSS shows the frequency of product choices across different age groups.
It helps identify patterns, such as which age group prefers which product. Each cell includes
counts and percentages, making it easier to compare groups. This method visualizes
associations between two categorical variables. The table supports hypothesis testing and
allows marketers to observe differences before drawing conclusions from the Chi-square test
results.
How should the analyst interpret a Chi-square test with p = 0.048 in this case?
Answer:​
A p-value of 0.048 is just below the 0.05 threshold, indicating a statistically significant
association between age group and product preference. The null hypothesis is rejected,
suggesting that product choice is not random across age groups. The analyst should examine
which specific groups show stronger preferences. This insight helps tailor marketing
strategies to different age segments, such as promoting Product A more heavily among
younger consumers if that trend appears in the data.
Case Study 7
Scenario
A city council evaluates noise complaints across three residential zones: central, suburban,
and industrial. Complaint counts from each zone are collected for a 3-month period. The data
is analysed using SPSS through descriptive statistics and boxplots. The council wants to
know if the variability in complaints across zones is significant and whether outliers are
influencing perceptions about noise levels.
How does a boxplot help in analysing noise complaint data in SPSS?
Answer:​
A boxplot helps identify the distribution of complaints across zones. It shows the median,
interquartile range, and any outliers. The length of the box and whiskers reveals variability.
Outliers are marked separately, highlighting unusually high complaint counts. This visual tool
helps the council assess whether complaints are concentrated or vary widely and if certain
zones require immediate attention due to extreme cases or consistent high frequencies.
What does high variability in boxplots suggest about complaint patterns?
Answer:​
High variability suggests inconsistent complaint frequencies across the area. One zone may
have frequent complaints while others have fewer. This indicates that noise levels or
reporting behavior are uneven. It signals the need for zone-specific interventions rather than a
uniform policy. If outliers are present, the council should investigate whether isolated events
or systematic issues caused the spikes before forming conclusions.
Case Study 8
Scenario
A software company evaluates user satisfaction across two software versions—Standard and
Premium. 100 users are randomly selected, 50 from each version. Satisfaction scores (1–10)
are compared using an Independent Samples T-Test in SPSS. Levene’s Test confirms equal
variances, and the p-value is 0.06. The company also calculates Cohen’s d = 0.15. The team
needs to decide if the Premium version provides enough added value to justify a higher price.
How is an Independent Samples T-Test conducted in SPSS for this scenario?
Answer:​
In SPSS, go to Analyze > Compare Means > Independent-Samples T Test. Select satisfaction
as the test variable and software version as the grouping variable. Levene’s Test checks
variance equality. The T-Test compares group means. If p < 0.05, there’s a significant
difference. If p > 0.05, as in this case, the difference is not statistically significant. The output
also includes group means and standard deviations for further interpretation.
What do p = 0.06 and Cohen’s d = 0.15 imply about software version satisfaction?
Answer:​
A p-value of 0.06 exceeds the 0.05 threshold, suggesting no statistically significant difference
in satisfaction between versions. Cohen’s d = 0.15 indicates a small practical effect. These
results imply that the Premium version does not provide a strong enough improvement in
satisfaction to justify a higher cost. The company should reconsider pricing or add more
valuable features to make Premium meaningfully different from the Standard version.
Case Study 9
Scenario
A manufacturing company wants to compare defect rates between two production
shifts—day and night. Data is collected over one month, recording the number of defective
units per shift. The analyst uses SPSS to calculate the mean and standard deviation for each
shift, then performs an Independent Samples T-Test. The p-value is 0.045, and standard
deviation for the night shift is higher. Management needs to assess whether the shift
influences defect rates and if variability is a concern.
How does SPSS help compare defect rates between shifts?
Answer:​
SPSS calculates the mean and standard deviation for each shift. The Independent Samples
T-Test compares the mean defect rates to test if the difference is statistically significant. A
p-value below 0.05 indicates a meaningful difference between shifts. SPSS also checks
variance using Levene’s Test and provides descriptive statistics, helping the analyst determine
whether shift timing affects product quality.
What do a p-value of 0.045 and higher variability in night shift defects indicate?
Answer:​
A p-value of 0.045 indicates a statistically significant difference in defect rates between
shifts. The higher standard deviation for the night shift shows inconsistent performance. This
suggests quality issues may be more variable at night. Management should investigate
operational factors—such as staffing or supervision—and consider targeted improvements
during night shifts to ensure consistent product quality.
Case Study 10
Scenario
A public health researcher studies the relationship between smoking status (smoker,
non-smoker) and respiratory illness (yes, no). Data from 300 individuals is collected and
analyzed in SPSS using a Chi-square test. The p-value is 0.021, and Cramér’s V is 0.32. The
agency wants to use the findings to support policy development and targeted awareness
programs.
What does the Chi-square test measure in this case?
Answer:​
The Chi-square test assesses whether there is a statistically significant association between
smoking status and respiratory illness. It compares observed and expected frequencies across
categories. A p-value below 0.05 suggests the relationship is not due to chance. SPSS
generates a cross-tabulation with percentages and a Chi-square value, helping the researcher
determine if smoking status relates to the likelihood of respiratory issues.
How should the agency interpret p = 0.021 and Cramér’s V = 0.32?
Answer:​
A p-value of 0.021 indicates a significant association between smoking and respiratory
illness. Cramér’s V = 0.32 reflects a moderate effect size, meaning the relationship has
practical relevance. These results support the need for targeted interventions. The agency can
use this evidence to justify focused health campaigns and smoking cessation programs,
especially in communities with high smoking prevalence.

Case Study 11
Scenario
A logistics firm tracks delivery times across three regions: North, South, and Central. The
analyst collects average delivery time (in hours) from 40 deliveries per region. A One-Way
ANOVA is performed using SPSS to determine whether the average times differ significantly.
The ANOVA result gives p = 0.037. A post hoc Tukey test shows that North has significantly
higher delivery time than Central. The operations team wants to assess where delays are
occurring and why.
How is One-Way ANOVA used in SPSS to assess delivery time differences?
Answer:​
In SPSS, the analyst selects Analyze > Compare Means > One-Way ANOVA. Delivery time
is the dependent variable; region is the factor. The test compares mean delivery times across
the three groups. A p-value less than 0.05 suggests a statistically significant difference
between at least two regions. Post hoc tests such as Tukey are used to find where those
specific differences lie.
What does a p-value of 0.037 and Tukey test result suggest to the operations team?
Answer:​
The p-value of 0.037 shows significant differences in delivery times across regions. The
Tukey test indicates that delays are higher in the North compared to Central. This directs the
operations team to investigate region-specific factors, such as traffic, staffing, or routing
issues in the North. Addressing these local inefficiencies could reduce overall delivery delays
and improve service consistency across all regions.
Case Study 12
Scenario
A non-profit runs a survey to measure public opinion on three community programs:
Education, Health, and Employment. Respondents rate their support on a 1–10 scale. The
organization uses SPSS to check if gender influences program preference using
cross-tabulation and a Chi-square test. The p-value is 0.58. They expected some variation but
want to understand how to interpret the results for reporting purposes.
What does the Chi-square test reveal in the context of this survey?
Answer:​
The Chi-square test checks if there is a significant association between gender and program
preference. In SPSS, a cross-tabulation shows the count and percentage of preferences by
gender. A p-value greater than 0.05, such as 0.58, suggests no significant relationship exists.
This implies that gender does not influence how participants rate the community programs,
and observed differences are likely due to random variation.
How should the non-profit interpret a p-value of 0.58 when reviewing gender differences?
Answer:​
A p-value of 0.58 means there is no statistically significant association between gender and
program preference. The non-profit should avoid gender-based assumptions when reporting
or designing initiatives. Instead, it should look at overall support trends or consider other
variables (e.g., age, education). The finding supports a neutral approach, treating program
support as broadly consistent across male and female respondents.