8. Wrapping up Stats Block 1

Johnny van Doorn

University of Amsterdam

2025-09-17

In this lecture we discuss:

• Correlation
  • In JASP
• Linear model
  • In JASP
• Revisiting NHST
  • Alpha
  • Confidence intervals

Reading: Chapters 1-8.8, not 6

Loose ends

JASP

• Open non .jasp files (e.g., .sav / .csv) - important for exam
• Correlation
• Partial correlation
• Regression
  • Transforming Adverts (1,000£ vs. 100,000£)
  • Assumptions
  • Outliers
  • Export (not for exam)

Correlation & NHST

We will collect n = 10 observations

• Set alpha: for which \(t\)’s do we reject \(H_0\)?
  • P(\(t\) | \(H_0: r = 0\))?
  • P(\(t\) | \(H_1: r = 0.3\))?

n <- 10
correlation <- 0.3
t.r <- ( correlation*sqrt(n-2) ) / sqrt(1-correlation^2)
t.r

[1] 0.8894992

Correlation & NHST

We will collect n = 30 observations

• Set alpha
  • P(\(t\) | \(H_0: r = 0\))?
  • P(\(t\) | \(H_1: r = 0.3\))?

n <- 30
correlation <- 0.3
t.r <- ( correlation*sqrt(n-2) ) / sqrt(1-correlation^2)
t.r

[1] 1.664101

Confidence Intervals

• Based on the sampling distribution, centered on the observed statistic

Confidence Intervals

• On repeated sampling, \((100 - \alpha)%\) of the intervals contain population/true value
• If you conclude the present interval contains true value, you have \(\alpha\)% chance of being wrong (Misconception Mutt 2.1)

Closing

Next Week

• Exam!
  • Mix of using JASP/interpreting output/conceptual understanding
  • Field book available (like in this weeks WA), including glossary
  • WA will contain old exam questions

Recommended Exercises

• Exercise 7.6, Exercise 7.7
  • Note: pets.jasp can also be downloaded from the Data page
• Exercise 8.1

Contact

• ln.AvU@nrooDnav.B.J
• JohnnyDoorn

CC BY-NC-SA 4.0