Week 1 - recap

what data is categorical?

nominal data

what data is discrete/continuous

ordinal, interval, ratio

nominal data

numbers or names as labels, no numerical relationship between values e.g. gender, religion

ordinal data

organised by rank → values represent true numerical relationships, but intervals between values may not be equal → e.g. race position, likert scale ratings

interval data

true numerical relationships and intervals between values are equal → NO true 0 point e.g. temperature

ratio data

true numerical relationships, equal intervals and true 0 point e.g. height or distance

when would you use a mean?

discrete or continuous data which is normally distributed

when would you use a median

discrete or continuous data which is not normally distributed

when would you use a mode

categorical data

what test of difference would you use for 1 IV with 2 levels (between & within pps)

• between pps → independent t-test

• within pps → paired t-test

what test of difference would you use for 1 IV with >2 levels (between and within)

• between pps → 1-way ANOVA

• within pps → 1-way repeated measures ANOVA

what test of difference would you use for 2IVs, between, within and mixed design

• between pps → 2-way independent ANOVA

• within pps → 2-way repeated measures ANOVA

• mixed design → 2-way mixed ANOVA

True-experimental IVs

• IVs are actively manipulated

• random allocation is possible → can make claims about causality

Quasi-experimental IVs

IV reflects fixed characteristics

• random allocation is not possible (so must be cautious about implying causality)

• e.g. handedness (2 levels; left and right)

what does kurtosis represent - mesokurtic, platykurtuc, leptokurtic

kurtosis is the spread of standard deviation:

• mesokurtic → standard s.d.

• Platykurtic → large s.d., less concentrated, -ve kurtosis value

• Leptokurtic → small s.d., more concentrated, positive kurtosis value

what is sampling error?

degree to which sample statistics differ from underlying population parameters

what is a type 1 and type 2 error?

• type 1 = reject null when the null is true

• type 2 = fail to reject the null when null is false