Week 5: Quantitative Research Methods

Page 1

Factors in a survey:

Taste: diversity/flexibility of the questions. Phone is bad for this, face to face is good for this.

Respondent: values anonymity, social desirability, there may be a low exhibit rate ie. the amount of the population that does the thing you're looking for.

Situation: cost/speed

In a witness simulation it is possible to calibrate variables

lab experiments: parameteres and extraneous variables are controlled

field experiment: hard to control.

Internal validity: causation v correlation

mediating v moderating: intermediate steps that link iv and dv v changing iv and dv relationship

Causality

eliminate other possible causes

x must occur before or at the same time as y.

Choosing constructs:

-identify a construct that may need to be measured

-do a lit review on that construct

Types of scales: likert/Thurstore/Gattson