Lecture note on statistical analysis methods: for MBA students
Highlight 4 main ideas of the following statistics
analysis methods in the context of doing MBA dissertation projects and provide
2 application examples (that utilize questionnaire survey data) on each of
these methods: 1. correlation analysis 2. Anova (single factor) 3. Chi-square
test
Below is an
MBA-dissertation-friendly guide to the three methods, focusing on what each
method is for, how to think about it, and how to use questionnaire survey data
in practice. I’ve kept the examples close to common dissertation topics such as
consumer behavior, employee attitudes, and service quality.utwente+2
1) Correlation analysis
Correlation
analysis examines whether two quantitative variables move together, and in what
direction. It is useful when your dissertation asks whether higher levels of
one survey score are associated with higher or lower levels of another score.utwente
4 main ideas
- It measures the strength
of association between two variables.
- It shows the direction
of the relationship: positive, negative, or near zero.
- It does not prove
causation, only co-movement.
- It works best when both
variables are measured on a scale such as a Likert composite score, total
score, or index created from questionnaire items.utwente
2 questionnaire-based application examples
- Customer
satisfaction and repurchase intention. Use
survey items to create a satisfaction score and a repurchase intention
score, then test whether more satisfied respondents also report stronger
intention to buy again.
- Work
engagement and job performance self-rating. Use questionnaire scales for engagement and self-rated
performance, then test whether respondents with higher engagement scores
also report higher performance scores.
2) One-way ANOVA
One-way ANOVA
tests whether the mean of a quantitative variable differs across three or more
groups defined by a single categorical factor. In MBA dissertations, it is
commonly used when you want to compare average survey scores across segments
such as age groups, income groups, job levels, or customer types.utwente+1
4 main ideas
- It compares group means
rather than individual responses.
- It uses one categorical
independent variable with two or more levels.
- It is appropriate when the
dependent variable is quantitative.
- If the overall test is
significant, you usually follow up with post-hoc comparisons to see
which groups differ.utwente+1
2 questionnaire-based application examples
- Service
quality perceptions across age groups. Survey
respondents rate service quality on a 1–5 scale, then compare mean ratings
across age groups such as under 30, 30–45, and over 45.
- Purchase
intention across income groups. Use a Likert-based purchase
intention score and test whether the average intention differs among low-,
middle-, and high-income respondents.
3) Chi-square test
The chi-square
test is used with categorical data to examine whether two variables are related
or whether observed responses differ from expected patterns. For questionnaire
surveys, it is especially useful when responses are recorded as categories such
as yes/no, male/female, brand choice, or grouped levels like
satisfied/neutral/dissatisfied.libraryguides.mcgill+2
4 main ideas
- It works with categorical
variables, not scale means.
- It compares observed
frequencies with expected frequencies.
- It can test independence
between two survey variables.
- It is common for
cross-tabulation analysis in business research questionnaires.libraryguides.mcgill+2
2 questionnaire-based application examples
- Preferred
payment method by age group. Survey respondents choose
payment method categories such as cash, card, or mobile wallet, and you
test whether preference is associated with age group.
- Brand
awareness by education level. Respondents indicate
whether they know a brand or not, and you test whether awareness is
independent of education level.
Quick way to choose
- Use correlation when
both variables are numerical scores and you want association.
- Use one-way ANOVA
when one categorical factor explains differences in a numerical survey
score.
- Use chi-square when
both variables are categorical or grouped into categories.
