Quantitative Methods for Researchers Paul Cairns
Your objectives Pretty general! – Refresher or introduction to experiments – Specific stats questiosn Your research areas? Quant for RS, Paul Cairns
My objectives Experimental Style – Exploratory or severe test Validity Testing as modelling – Distributions – Choice of tests Quant for RS, Paul Cairns
Context HCI examples Widely applicable principles Quant for RS, Paul Cairns
Why do we do experiments? Quant for RS, Paul Cairns
Philosophy of experiments Test theories Isolate phenomena Severely test Quant for RS, Paul Cairns
Some consequences Intrinsic value Big is not always better Narrow focus is essential Quant for RS, Paul Cairns
Experimental argument Belief:X causes Y – A reason for looking Try: change X and measure Y Analyse carefully Produce evidence Quant for RS, Paul Cairns
Statistical experiments Natural variation – People, environment, stochastic Systematic vs chance differences No certainty Quant for RS, Paul Cairns
Exploratory experiment Controlled environment Unknown focus – Unknown phenomenon? Reason for looking Weak evidence Quant for RS, Paul Cairns
Devising an experiment Research question (disposable) One sentence May use jargon Answer is “yes/no” but probably “maybe” Question suggests how to answer it Quant for RS, Paul Cairns
Devise a research question In groups of two or three, each have a go at a research question. Take turns to explain and be criticised. Be happy to be wrong/stupid. RQs are disposable. Quant for RS, Paul Cairns
Experimental Design Addresses question Design => Data => Interpretation Validity Quant for RS, Paul Cairns
Working example Does making things harder to read reduce transcription errors? Quant for RS, Paul Cairns
Variables Independent variable (IV, X) – Experimental conditions Dependent variable (DV, Y) – quantitative Confounding variables Quant for RS, Paul Cairns
Validity Construct Internal External Ecological Quant for RS, Paul Cairns
Construct validity What are you measuring? What are you manipulating? Quant for RS, Paul Cairns
Internal validity Is X causing Y? Is X the only thing causing Y? – Confounding variables Experimental drift Quant for RS, Paul Cairns
External validity Generalisability People Places Systems Interactions Quant for RS, Paul Cairns
Ecological validity Relevance to out there? Quant for RS, Paul Cairns
Critique the paper Challenges to validity Quant for RS, Paul Cairns
Components of a Method What you are trying to find out What the variables are Who/what you examined What you used to do it How you did it Quant for RS, Paul Cairns
Components of a Method Research question Design Participants/test sets Materials Procedure Quant for RS, Paul Cairns
Homework Does making game playing more sociable improve the player experience? Devise an experiment Quant for RS, Paul Cairns
Reading Hacking, Representing and Intervening Cairns, Encyclopedia of HCI, 2 nd edn, chap 34 Cairns, Cox, Research Methods for HCI: chaps 1, 6 Harris, Designing and reporting experiments in psychology, 3rd edn Quant for RS, Paul Cairns