Operationalising variables and bringing concepts to life
14 November, 2022
The week ahead (week 6)
Operationalisation of variables requires a consideration of the reliability and validity of the method of operationalisation.
Operationalisation of variables also requires specification of the scale of measurement: nominal, ordinal, interval, or ratio.
Finally, operationalisation of variables can also specify details of the measurement procedure.
See Howitt and Cramer Chapter 3.3 (Box Research Example - Condon & Crano 1988)
This person would like(dislike) me
This person would like(dislike) working with me in an experiment
Two key concepts you’ll be thinking about and reading around…
Intelligence
important life outcomes
How are these important and very tricky concepts
Defined
Measured
And brought to life in the lab!
How have you defined your variables of interest?
How have you measured or categorised your variables of interest?
e.g. Social Media use
Gender
Frequency or low/high extraversion
•Develop research aims
•Specify research questions/hypotheses related to these aims
•Identify relevant constructs and concepts
•Translate constructs and concepts into variables (i.e., a logical set of characteristics/features)
•Translate variables into measurements (i.e., the quantification of characteristics/features)
A systematic examination of relationships between variables
‘Variables’ are ‘translated’ concepts, constructs or phenomena
Independent
Experimental - The variable I manipulate
Non-Experimental - Comparison groups
Dependent
Nominal/Categorical
Male/Female/…
Vegan, Vegetarian
Smoker/Non-Smoker
Ordinal
Numbers representing a rank position in a group
Not representative of an actual definite number/score/value - without information about the ‘gap’ between numbers
First, second, third
Tallest/Shortest
Interval
Numbers represent equal units giving information about the ‘gap’ between numbers
Temperature
Psychological Scales
Ratio
Interval measurements with an absolute zero, of equal units,
Weight
Length
Time/Reaction time*
A legend. Author of Learning Statistics with Jamovi/R
• My age is 33 years.
• I do not like anchovies.
• My chromosomal gender is male.
• My self-identified gender is male*
*see footnote 2, page 14 Learning Statistics with Jamovi
Dangerzone!
We often suggest a median split to dichotomise a continuous variable, e.g. for the purposes of creating a 2 level IV.
It’s a useful exercise in calculating a ‘computed variable’ in SPSS or Jamovi
It is NOT best-practice usually.
It is a key learning outcome that you are able to perform a standardised set of analysis, specifically, the 2x2 ANOVA with any necessary assumption checks and post-hocs + plots
You should think carefully about:
How you define your variables - this is probably a part of the introduction that students DON’T think about enough
How you measure or categorise your variables (IVs and DVs) - this is probably the single thing I look at first when peer-reviewing research!
How well your manipulation does what it claims to. Does you manipulation bring the thing it proposed to to life well?
Qualtrics (99% of you will use this!)- Full Support in Tomorrow’s labs (5mins)
Log in to an account with your Goldsmiths ID!!
10% of you will apply for the wrong type of account and be stuck for 2 weeks
Familiarise yourself with the Ethics Application process (10mins)
Consider the steps required to bring your study to life! (the rest)