Psych 3010 Final

Quasi-experimental design

No random assignment; natural groups

Regression to mean

Extreme scores move toward average over time

Spontaneous remission

Improvement happens without treatment

Selection bias

Groups differ before study starts

Maturation

Natural change over time

Pretest sensitization

Changes later responses

Non-equivalent groups

Not comparable at start

Local history

Event affects only one group

Contemporary history

Event affects all groups

Reversal (time series)

Effect disappears when treatment removed

Control interrupted time series

Control group helps rule out history effects

Switching replication (Switch & Replicate)

Groups switch treatment roles

Advantages of switching replication

Both groups get treatment; stronger evidence

Longitudinal challenges

Time, cost, dropout, practice effects

Generational effects

Differences from time period born in

Age effects

Differences from aging

Cross-sequential design

Combines cross-sectional + longitudinal

2x3 design

2 IVs; one has 2 levels (conditions), one has 3

Say you have a study about studying and memory.

Independent Variable 1: Study method

• Level 1: reading

• Level 2: flashcards

That’s a “2-level” variable (because it has 2 conditions).

Independent Variable 2: Time of day

• Level 1: morning

• Level 2: afternoon

• Level 3: night

That’s a “3-level” variable.

= 6 different combinations

Simple effects

looking at the effect of one variable under one specific condition of another variable.
Study method (reading vs flashcards)

• Time of day (morning vs night)

A simple effect question would be:

• “Which study method works better in the morning?”
  or

• “Which study method works better at night?”

So instead of looking at all the data together, you focus on one situation at a time.

Interaction

Effect of IV depends on another IV

Mixed factorial

combines Different groups of people (within subjects) AND Repeated testing of the same people (repeated measures)

P-value

“If there is really NO effect, how weird are these results?”

• Small p-value = results are surprising if there’s truly no effect

• Large p-value = results are not very surprising

P-value myth

A p-value does NOT tell you whether the null hypothesis is true or false.
It only tells you how unusual the data would be if the null hypothesis were true.

Alpha

your false alarm tolerance level before you start the study

If p is smaller than alpha = reject the null hypothesis

If p is larger than alpha = fail to reject the null hypothesis

p vs alpha

Alpha is the bar. P-value is how high your data jumps. If it clears the bar, you reject the null.

Type 1 error

False positive

Type 2 error

missed effect

Power

ability of a study to find a real effect

Power factors

Sample size, effect size, quality, design

Low power

Common in psych research. small samples or weak designs

Null decision

At the end of a study, you either:

• Reject the null hypothesis
  or

• Fail to reject the null hypothesis

You NEVER say:

• “The null hypothesis is proven true”

Why?
Because not finding evidence for an effect does not prove the effect do

Meta-analysis

Combines results across studies

Why meta-analysis

Summarize trends, resolve conflicts

k

Number of studies in meta-analysis

N

total sample size

r

Average correlation between two variables across studies.

rho

what the relationship likely really is

IV vs DV

IV- Cause, Manipulated

DV- Effected, measured

Null Hypothesis H₀

Assumption that there is no real effect, difference, or relationship between variables