KHA350 Lecture Notes - Lecture 1: Descriptive Statistics, Statistical Inference, Blind Experiment

34 views5 pages
Research Methods: Week One
Experimental design:
- aim to establish causal link between one variable that you manipulate, and
some measure of participant response
- In charge of independent variable: dependent: outcome measures
oMust work out how many IV’s and how many levels there are within
these
- What is the effect of some variable, on some aspect of human behaviour or
performance
- Independent variable affects the dependent variable
- Extraneous variables:
oManipulation of the IV
oOutcome: variable of the dependent measure
oExtraneous secondary variable
oHow do we know if the changes in the outcome are due to IV or
secondary variable
oMust try to minimize the impact the effect of these secondary variables
oThis is where good design is important
oWays of controlling for the impact of these variables:
Elimination:
Completely remove the presence of extraneous
Exclude participants that may have some effects
Constancy:
Cant exclude things from experiment
Instead keep them constant
Making secondary variable into an IV:
If interested in testing whether there is an impact of
other variables, you can build it into experiment as
another IV
Eg. Drug treatment for depression: drug vs. placebo,
plus two levels of IV’S (men and women)
Randomization:
If numerous or unknown influences: smears the error
across groups
Statistical adjustment:
Pre test- post test: use change scores
Analysis of covariance
oHypothetical experiment:
Eg. Impact of alcohol on driving ability
What are the possible secondary variables that might impact on
driving performance:
Tolerance to alcohol
Body weight
Sensory factors
Medication use
How to keep these controlled:
Social drinkers included
find more resources at oneclass.com
find more resources at oneclass.com
Unlock document

This preview shows pages 1-2 of the document.
Unlock all 5 pages and 3 million more documents.

Already have an account? Log in
Pre test for vision
Pre test driving ability, post test: look at the change in
people’s performance
Elimination:
oOnly have none drinkers, normal vision
Constancy:
oStandardized time of day and weather
- Source of error:
oApart from extraneous variables
oDesign and measurement problems:
Invalidity
Or was the change to the DV actually caused by the IV
Was it something else?
Proactive history:
Subject variables: usually try to minimize through
elimination, constancy etc. AKA secondary variables
Retroactive history:
Things that may happen during the course of the
experiment that may influence
Repeated testing:
Learning or sensitization from pre-test
Improving and becoming more familiar
Learning strategies also included
Reduce the impact of bias is important
Statistical regression towards the mean:
Test unreliability: extreme scores will move closer to
the mean if you re-test them
Controlled groups help surmount this
Bonus variation wont continue to grow
Control: gives sense of the degree of change that
happens just by the pure retest component
Can partial this out in contrast to the change in your
independent variable
Maturation:
Changes in participants
Particularly if over a long time period
Eg. Teaching intervention for children over a six month
period
oAttributing it to the teaching methods is
problematic as they are also going through
cognitive development rapidly
Clinical examples:
oCBT over 12 weeks
oMeasure depression at base line compared to
week 12
oSometimes people do other things outside of the
CBT that impacts their depression levels
find more resources at oneclass.com
find more resources at oneclass.com
Unlock document

This preview shows pages 1-2 of the document.
Unlock all 5 pages and 3 million more documents.

Already have an account? Log in

Document Summary

Experimental design: aim to establish causal link between one variable that you manipulate, and some measure of participant response. In charge of independent variable: dependent: outcome measures: must work out how many iv"s and how many levels there are within these. What is the effect of some variable, on some aspect of human behaviour or performance. Exclude participants that may have some effects. If interested in testing whether there is an impact of other variables, you can build it into experiment as another iv. Drug treatment for depression: drug vs. placebo, plus two levels of iv"s (men and women) If numerous or unknown influences: smears the error across groups. Pre test- post test: use change scores. What are the possible secondary variables that might impact on driving performance: Pre test driving ability, post test: look at the change in people"s performance. Elimination: only have none drinkers, normal vision. Constancy: standardized time of day and weather.

Get access

Grade+
$40 USD/m
Billed monthly
Grade+
Homework Help
Study Guides
Textbook Solutions
Class Notes
Textbook Notes
Booster Class
10 Verified Answers
Class+
$30 USD/m
Billed monthly
Class+
Homework Help
Study Guides
Textbook Solutions
Class Notes
Textbook Notes
Booster Class
7 Verified Answers

Related Documents

Related Questions