Research Project Final Exam Summary

RESEARCH PROJECT FINAL EXAM 23RD MARCH
2021 – ROTTERDAM SCHOOL OF MANAGEMENT
TABLE OF CONTENTS
Lecture 1 ........................................................................................................................... 2
Lecture 2 – Experimentation .............................................................................................. 4
Workshop 1 – Experimentation ......................................................................................... 6
Lecture 3 – Regression Analysis ......................................................................................... 7
Workshop 2 – Regression Analysis (Introduction, background) .......................................... 9
Lecture 4 – Regression analysis: Inference ....................................................................... 11
Workshop 3 – Inference ................................................................................................... 13
Lecture 5: Regression Analysis: Assumptions and Model Assessment ............................... 14
Workshop 4 – Residual Analysis....................................................................................... 16
Lecture 6 – Regression analysis: Standardized coefficients and predictions ...................... 18
Workshop 5 – What is a good regression model?............................................................. 19
Supplementary ................................................................................................................ 21
Interpretation interaction ......................................................................................................... 21
Nominal predictors ................................................................................................................... 24
Lecture 7 – Standardized Coefficients in linear regression & prediction ............................ 27
Workshop 6 – prediction with regression ......................................................................... 28




1

,LECTURE 1
Concept/Construct
Concept is more abstract as a construct, but both are abstract and not directly measurable. When it is
operationalized, it becomes a variable. How to make it less abstract? Be more specific.

Cronbach’s Alpha
Cronbach’s Alpha (a) = internal consistency. It looks at the reliability between multiple variables. The
closer to 1, the higher your scale’s reliability (between 0 & 1). Normally used value a ≥ 0.70. R will give
you the input on which variable you need to remove to be above the threshold of 0.70. It also shows
whether reliability increases/decreases when 1 item is dropped. If you get a result of 1, it means that
the answers of all the questions are the same. If you need to remove an item, you also don’t take it
into account in calculating your means, etc.

If a ≥ 0.7, then you may compute the “composite” measure
If a < 0.7, then you may delete one or several items until a ≥ 0.7

Create a composite by simply calculating the mean. On this mean, a statistical test can be performed.

Reverse coding
Reverse-coded item à recode the values. E.g. 1, 2, 3, 4, 5 becomes 5, 4, 3, 2, 1
Recoding values in R à mapvalues

Always recalculate alpha again with all items (so with the recoded ones)

R tutorial
Multidimensional scale
One scale = 1 -5
Second scale = 1 - 7
Need to rescale one of them and then calculate a
It is not possible to merge scales where one is categorical, and one is interval

Nominal items are only possible to be merged (aggregated) if the answers are binary (2 answers) e.g.,
yes & no. the aggregated measure would become a ratio variable (how many yes and how many no)

If it’s not possible to merge, you are required to do 2 different tests

Hypothesis testing
A theoretical hypothesis without direction is not testable

When making interpretations about hypothesis testing, we need to take into account both of the
following general elements:

Direction of the effect
- Is the correlation coefficient positive or negative?
- Is the average mean in the group A higher than in the group B?
- Is the regression coefficient positive or negative?
Significance
- Is the p-value <0.05




2

, Theoretical hypothesis: IV positively influences DV (H1, H2, H3)
Statistical hypothesis: Null hypothesis (H0) and alternative hypothesis (Ha)

Theoretical interpretation:
If r>0 & p<0.05, the results confirm the hypothesis
If p>0.05, the results do not confirm the hypothesis

Statistical interpretation:
If r>0 & p<0.05, the researcher rejects the null hypothesis
If p>0.05, the researcher cannot reject the null hypothesis

Usually, researchers only mention whether the theoretical hypothesis was confirmed or not. They do
not say anything about the statistical hypothesis à they do not report it, nor do they say whether it is
rejected or not.

Research hypothesis = theoretical hypothesis

ANOVA
With an ANOVA test, a difference between two or more groups can be observed. But it is not enough
to conclude on statistical significance. Therefore, “General omnibus F test” à do at least two groups
have different average means? If F (p) value is <0.05, then you can conclude that at least 2 groups
show significantly different means. Which ones? Perform a Tukey multiple comparisons of means test.
Again look at p-value, if <0.05, there is a significant difference in means between 2 groups.




3