Week 08 – Research Questions, Variables, and Study Design

1. What Is Research?

Research begins with curiosity — a question about how the world works.
In this course we work with existing datasets, but behind each dataset there is (or should be) a research question.

This week connects your earlier work on data types and levels of measurement (Lecture 01) with:

• research questions and hypotheses
  
• independent and dependent variables
  
• study design
  
• how variables are represented in data sets

By the end of this week you should be able to:

• turn an informal idea into a research question and a testable hypothesis
  
• identify independent and dependent variables
  
• classify variables by type and level of measurement
  
• understand how design and variables shape statistical analysis

---

2. The Research Cycle

We can think of research as a simple cycle:

A simplified research cycle

1. Observation
   Something you notice about the world.
   Example: Some students study a lot but do not always achieve high marks.

2. Research question
   A focused, answerable question.
   Example: Is there an association between hours of study and overall mark?

3. Hypothesis
   A testable prediction.
   Example: Students who study more hours per week will achieve higher marks.

4. Identify variables
   Decide what you will measure.

   • IV: Study_hours
     
   • DV: Overall_mark

5. Collect data
   Gather values for these variables in a sample.

6. Analyse data
   Summarise variables and test relationships.

7. Interpret and conclude
   Relate the results back to the research question.

This week focuses on Step 4 because everything else depends on identifying variables correctly.

---

2.1 From Constructs to Variables (link to Lecture 01)

In Lecture 01 we noted that before data collection, researchers must decide what they intend to measure. They begin with abstract ideas (constructs), such as stress, learning, or academic success. Constructs cannot be observed directly — we only see them through the measures we choose.

The logic is:

• Constructs are theoretical ideas.
  
• Measures operationalise these ideas.
  
• Variables are the recorded data.
  
• All variables contain some measurement error.

In this course we mainly work at the level of variables, but it is important to remember where they come from.

---

3. Variables in Research

A variable is any characteristic that varies between individuals.

In Lecture1_data.csv, variables include:

• Gender, Degree, Exercise
  
• Study_hours, Sleep_hours
  
• Stress_level, Satisfaction_Likert_item
  
• Overall_mark, IQ

Variables can be classified in two ways:

1. By role — how they function in the research design
   
2. By type — the kind of information they contain

---

3.1 Variables by Role: Independent and Dependent

Role	What it does	Example from the dataset
Independent variable (IV)	Used to explain/predict something; sometimes manipulated	Study_hours, Gender, Exercise
Dependent variable (DV)	The outcome being measured	Overall_mark, Sleep_hours, Stress_level

A simple way to remember:

IV → DV
“Does X influence / predict / relate to Y?”

Examples using the lecture dataset:

• Does study time predict overall marks?
  • IV: Study_hours
    
  • DV: Overall_mark
• Do students who exercisesleep more?
  • IV: Exercise
    
  • DV: Sleep_hours
• Is stress level related to hours of study?
  • IV: Study_hours
    
  • DV: Stress_level

---

3.2 Manipulated vs Measured Variables

• Manipulated variables

  • Assigned by the researcher (e.g. an experiment).

  • Not present in this dataset.

• Measured variables
  • Naturally occurring characteristics recorded by the researcher.
  • All variables in Lecture1_data.csv are measured.

---

4. Variables in the Lecture Dataset

library(readr)
library(dplyr)

lecture <- read_csv("Lecture1_data.csv")
glimpse(lecture)

Rows: 200
Columns: 14
$ ID                        <chr> "S001", "S002", "S003", "S004", "S005", "S00…
$ Degree                    <chr> "Architecture", "Linguistics", "English", "P…
$ Year                      <dbl> 2, 1, 2, 1, 2, 2, 2, 3, 2, 2, 1, 3, 1, 1, 2,…
$ Gender                    <chr> "Female", "Male", "Female", "Male", "Non-bin…
$ Study_hours               <dbl> 6.8, 3.1, 8.2, 5.5, 4.0, 7.4, 2.7, 9.0, 5.9,…
$ Sleep_hours               <dbl> 7.4, 5.8, 7.6, 6.7, 5.9, 7.8, 5.0, 8.2, 6.6,…
$ Stress_level              <chr> "Moderate", "High", "Low", "Moderate", "High…
$ Satisfaction_Likert_item  <chr> "Satisfied", "Dissatisfied", "Very Satisfied…
$ Satisfaction_Likert_value <dbl> 4, 2, 5, 4, 2, 5, 1, 5, 4, 4, 2, 5, 3, 2, 5,…
$ Coffee_per_day            <dbl> 1, 3, 1, 2, 2, 1, 4, 0, 2, 1, 3, 0, 2, 3, 1,…
$ Social_media_hr           <dbl> 2.4, 4.7, 2.0, 3.0, 4.1, 2.1, 5.2, 2.5, 3.2,…
$ Exercise                  <chr> "Yes", "No", "Yes", "No", "No", "Yes", "No",…
$ Overall_mark              <dbl> 68, 55, 78, 65, 58, 74, 49, 81, 66, 70, 56, …
$ IQ                        <dbl> 102, 108, 95, 110, 99, 106, 88, 120, 100, 10…

Below is a description of all variables in the Lecture1_data.csv dataset.

Variable Name	Description
ID	Participant ID (unique identifier)
Degree	Degree programme the student is enrolled in
Year	Year of study (1, 2, 3 …)
Gender	Self-reported gender
Study_hours	Hours spent studying per day (continuous)
Sleep_hours	Hours of sleep per night (continuous)
Stress_level	Reported stress level (Low, Moderate, High)
Satisfaction_Likert_item	Satisfaction rating (text: e.g. Satisfied)
Satisfaction_Likert_value	Numerical coding of satisfaction (1–5)
Coffee_per_day	Cups of coffee consumed per day (count)
Social_media_hr	Time spent on social media per day (hours)
Exercise	Whether the student exercises (Yes/No)
Overall_mark	Overall module mark (%)
IQ	Estimated IQ score

4.1 Variables by Type: Categorical and Numeric

We also classify variables by the kind of information they contain. This affects:

• how we summarise them (counts, means, medians, etc.)
  
• which plots we use (bar plots, histograms, boxplots, scatterplots)
  
• which statistical tests are appropriate

4.1.1 Categorical variables

They group observations into labels.

Example variables in the lecture dataset:

• Gender

• Exercise

• Degree

• Stress_level

• Satisfaction_Likert_item

• Satisfaction_Likert_value (ordinal coded)

Subtypes

• Nominal: no natural order (e.g. Gender, Degree)

• Ordinal: meaningful order (e.g. Stress_level, Satisfaction ratings)

We summarise categorical variables using:

• frequencies (counts)

• percentages / proportions

• bar plots

4.1.2 Numeric variables

Examples in the lecture dataset:

• Study_hours

• Sleep_hours

• Overall_mark

• IQ

• Coffee_per_day

• Social_media_hr

Subtypes:

• Discrete (counts): Coffee_per_day

• Continuous (measurements): Study_hours, IQ, Sleep_hours

We summarise numeric variables using:

• means, medians

• standard deviation, IQR

• histograms, boxplots, scatterplots

---

5. Levels of Measurement

In Lecture 01 you met four classic levels of measurement. Here we briefly reconnect them to research design, using variables from the Lecture1_data.csv dataset.

Level	Meaning	Examples (lecture dataset)
Nominal	Categories with no intrinsic order	Gender, Degree, Exercise
Ordinal	Ordered categories, unequal gaps	Stress_level, Satisfaction_Likert_value
Interval	Numeric, equal spacing, no true zero	IQ (treated as an approximate interval scale in psychology)
Ratio	Numeric, equal spacing, true zero	Study_hours, Sleep_hours, Overall_mark, Coffee_per_day, Social_media_hr

In practice we often talk simply about categorical vs numeric variables, but levels of measurement sit in the background and influence what is sensible:

• you would not normally compute a mean of nominal data
  
• you should be cautious interpreting means of ordinal scales
  
• most parametric tests (e.g. t-tests, correlation) assume numeric (interval/ratio) data

---

6. Exercise: Classifying Variables in the Lecture Dataset

Use the variables from Lecture1_data.csv to practise identifying:

1. whether each variable is categorical or numeric
   
2. the level of measurement
   
3. whether it could be an IV, a DV, or either

Consider the following variables:

• Gender
  
• Exercise
  
• Study_hours
  
• Overall_mark
  
• Stress_level
  
• Coffee_per_day
  
• Social_media_hr

Your Task

Complete the table below by filling in the Type, Level, and Possible Role(s) for each variable.

Variable	Type (Categorical/Numeric)	Level	Possible Role(s)
Gender
Exercise
Stress_level
Study_hours
Overall_mark
Coffee_per_day
Social_media_hr

Now write one possible research question using any two variables above.

Example format:

1. “Does ___ predict ___?”
   
2. “Do ___ and ___ differ in ___?”
   
3. “Is there an association between ___ and ___?”

---

Suggested Answers

Classification Table

Variable	Type	Level	Possible Role(s)
Gender	Categorical	Nominal	IV (grouping variable)
Exercise	Categorical	Nominal	IV or DV
Stress_level	Categorical	Ordinal	IV or DV
Study_hours	Numeric	Ratio	IV or DV
Overall_mark	Numeric	Ratio	DV (usually an outcome)
Coffee_per_day	Numeric	Discrete	IV
Social_media_hr	Numeric	Ratio	IV or DV

Example research questions

• Does study time predict overall marks?
  
• Do students who exercise sleep more?
  
• Are stress levels associated with time spent on social media?
  
• Do male and female students differ in their study hours?