Overview

You Can Do This.

Welcome to Methodology Pro. Research is often presented as an insurmountable wall of mathematics. It is not. It is a logic puzzle.

You don't need to be a statistician to be a researcher. You just need a system. This tool is your reference system. By the end of this session, you will know exactly what to look for, how to run it, and how to write it up. Take a deep breath. You got this.

The Dashboard

Pocket Data

We use this tiny dataset (N=3) so you can verify the logic instantly.

Participant 001
Placebo
Dose0 mg
Stress9.0
Participant 002
Drug A
Dose10 mg
Stress6.0
Participant 003
Drug A
Dose50 mg
Stress3.0
Research Logic

Finding "The Gap"

Research isn't just about collecting data; it's about solving a missing piece of a puzzle. A "Gap" is the space between what we know and what we need to know. Select a domain and a logic type to see how a hypothesis is constructed.

1. Domain
2. Logic Type
3. The Gap Statement
"While CBT is known to reduce general anxiety, its efficacy specifically for Generalized Anxiety Disorder (GAD) remains understudied."
Why this is a Gap: We know the intervention works broadly, but we lack specificity for this particular diagnosis. This justifies the new study.
Prerequisites

The Hierarchy of Variables

Variables are the building blocks of research. You must identify them correctly to choose the right test. We follow the NOIR framework.

Categorical (Qualitative)

Level 1

Nominal

Names / Labels. No mathematical value. No order.
Ex: Eye Color, Group (Placebo/Drug).

Level 2

Ordinal

Rank Order. A strict order exists, but the distance between points is unknown.
Ex: Race finish (1st, 2nd, 3rd), Likert Scales (Agree/Disagree).

Continuous (Quantitative)

Level 3

Interval

Equal Distances. The gap between 1 and 2 is the same as 3 and 4. No "True Zero."
Ex: Temperature (Celsius/Fahrenheit), IQ Scores.

Level 4

Ratio

True Zero. Zero means "absence of the variable." Allows for multiplication/division.
Ex: Height, Time, Reaction Time, Weight.

Experimental Roles

Independent (IV)

The Cause. The variable you manipulate or group by.
"Does Drug A affect Stress?"

Dependent (DV)

The Effect. The outcome variable you measure.
"Does Drug A affect Stress?"

Group Differences

Independent Samples T-Test

1. The Logic
2. R Syntax
3. SPSS GUI
4. Output

The Narrative: Signal to Noise

A T-Test compares the Signal (the difference between the two group means) against the Noise (the variance/spread within the groups). If the Signal is much louder than the Noise, we conclude the groups are truly different.

Assumption Protocol (Interactive)
✓
Normality Check
Is data bell-curved?
✓
Homogeneity of Variance
Are group spreads equal?

R Syntax

t.test(Stress ~ Group, data = pocket_data, var.equal = TRUE)

SPSS Steps

Analyze > Compare Means > Independent-Samples T Test.
Test Variable: Stress. Grouping Variable: Group.

Significance
p < .05
Cohen's d
1.2 (Large)
Group Differences

One-Way ANOVA

1. The Logic
2. R Syntax
3. SPSS GUI
4. Output

The Narrative: Variance Partitioning

ANOVA argues that the variance between the 3 groups is greater than the variance within them. If we ran 3 separate T-Tests, our error rate would inflate to ~15%. ANOVA keeps it at 5%.

Assumption Protocol (Interactive)
✓
Independence
Groups are separate people?
✓
Homogeneity (Levene's)
p > .05 required.
model <- aov(Stress ~ Dose_Group, data=data)
TukeyHSD(model)

Analyze > Compare Means > One-Way ANOVA. (Post Hoc: Tukey).

F-Ratio
5.66
Post-Hoc
50mg > 0mg
Relationships

Correlation & Regression

Correlation
Regression
1. The Logic
2. R Syntax
3. SPSS GUI
4. Output

The Narrative: Association

We ask: "Do these two variables move together?" It is non-directional. X is associated with Y, but X might not cause Y.

Assumption Protocol
✓
Linearity
Is relationship a straight line?
cor.test(data$Dose, data$Stress)

Analyze > Correlate > Bivariate.

Pearson's r
-.82
Meaning
Strong Negative
Mechanism

Mediation Analysis

1. The Logic
2. R Syntax
3. SPSS GUI
4. Output

The Narrative: The Process

We argue that the effect of Drug (X) on Stress (Y) is not direct, but travels through Sleep (M).

Operationalization: We test the "Indirect Effect" (path a * path b). Normal theory assumes this value is normally distributed, but it is usually skewed. Therefore, we MUST use Bootstrapping.

Assumption Protocol
✓
Bootstrapping
5000 Samples Enabled?

R Syntax (lavaan)

# Define Model
model <- '
Sleep ~ a*Drug
Stress ~ b*Sleep + c*Drug
indirect := a*b
'
# Run SEM with Bootstrapping
fit <- sem(model, data=data, se="bootstrap")

SPSS Steps (PROCESS Macro)

1. Analyze > Regression > PROCESS v4

2. Y: Stress, X: Drug, M: Sleep.

3. Options > Check "Bootstrap inference for indirect effects".

Indirect Effect
a * b
Confidence Interval
Does not cross 0
Publication Engine

Report Generator

Select your test to build the specific write-up components for a professional manuscript.

Select a test above to begin constructing your paper...
01. Abstract Construction
A. The Context (Aim)
...
B. The Method
...
C. The Findings
...

Final Abstract (APA Format):

Loading...

02. Results Write-Up (APA)

Copy/Paste Ready:

Loading...

03. Discussion Logic
Loading...