Week 5 - Research Methods - Experimental Design

Three Defining Features of a Psychological Experiment

• tests a casual hypothesis

• manipulates the IV

• controls other potential influences on the DV

Independent Variable

= the variable manipulated by the researcher

Dependent Variable

= the outcome measured to see weather the IV has an effect

Extraneous Variable

= any variable other than the IV that could influence the DV

Confounding Variable

= an extraneous variable that systematically varies with the IV, and could explain differences in the DV

Experimenter Expectancy Effects

= when a researcher’s expectations unintentionally influence ppt behaviour and potentially bias the results

Subject Expectancy Effect

= when ppt’s knowledge or assumptions about the hypothesis influence their behaviour or responses

Standardisation

= keeping procedures, instructions and conditions consistent to reduce extraneous influences

Single-blind study

= ppts are unaware of their condition - expectancy effects can be reduced

Double-blind study

= both ppts and experimenters are unaware - it protects both experimenter and subject expectancy effects

Random Allocation

= assigning ppts to experimental conditions by chance

• reduces systematic differences between groups, making confounding by ppt characteristics less likely

Natural Group Designs

= it compares pre-existing groups

• limitation = naturally occurring groups may differ in many other characteristics

Matched Groups Designs

= ppts are matched on potential confounders

• limitation = important variables may be missed + matching many characteristics is difficult

Between-subject design

= different ppts complete different conditions

• strength = no carry-over between conditions

• limitation = ppt differences can confound results

Within-subjects design

= the same ppts complete multiple conditions

• strength = controls stable individual differences - more data per ppt

• limitation = order effects can influence results

Order Effects - 6 types

= occur when ppt’s responses are influenced by the sequence in which conditions are experienced

6 types:

• fatigue/boredom - performance declines as ppts become tired

• practice - performance improves through repetition

• habituation - responses decrease after repeated exposure

• sensitisation - responses increased after repeated exposure

• adaptation - ppts physically adjust to conditions

• contrast - earlier stimuli influence judgements of later stimuli

Counterbalancing

= involves varying the order in which experimental conditions are presented to reduce confounding in within-subject designs

Incomplete within-subject design

= each ppt completes every condition only once, but the order varies between pts

• order effects are balanced between ppts

Complete within-subjects design

= each ppt completes every condition multiple times, with the order varied within that ppt

• order effects can be balanced within ppts

How many possible orders are there for three experimental conditions?

= 3

• 3×2×1 = 6 orders

Latin Square Designs

= a counterbalancing method that uses selected orders instead of every possible order

• limitation = doesn’t necessarily balance every possible preceding-condition effect

Block Randomisation

= contains one occurrence of every experimental condition

• eg - three conditions

• block 1 - R > G > Y

• block 2 - Y > R > G

• block 3 - G > Y > R

ABBA Designs

= presents a condition sequence followed by its reverse to help balance effects of presentation order

Differential Transfer

= occurs when experiencing one condition affects performance in a later condition differently from other conditions

• limitation of counterbalancing because it can create carryover effects that counterbalancing may not eliminate

How to critically evaluate an experiment

Five questions to ask:

• confounders = could another variable explain the results?

• standardisation = were procedures and instructions consistent?

• blinding = could ppt or experimenter expectations influence behaviour?

• random allocation = were ppts assigned to groups by chance?

• counterbalancing = if ppts completed multiple conditions, were order effects controlled?