PSYC2009 Study Guide - Quiz Guide: Statistical Inference, Autosuggestion, Statistical Parameter
Inferential Statistics
Aims:
• Infer some characteristic of a population/phenomenon on the basis of a sample evidence
• Most studies in human sciences involve generalisations and inferences beyond the people
actually studied.
• Kinds of generalisations
1. To the population from which the sample came
• Cultural differences condition people to have different cognition to those who
grew up in a different culture, therefore, often, results are not really transferrable
or generalisable
2. Across time
3. To other populations than the one sampled
• Usually only the first kind is covered by inferential statistics
• Motives
o Unrealistic to think that we can get information form absolutely everyone in a
population, so a sample needs to be taken that is representative of the entire population
Procedure:
• Parameter estimation: where we get a sample estimate of the statistic were interested in and
use this information to estimate the population parameter
• Sampling frame: population from which your sample was taken. The population you can
statistically generalise the population parameter to
• Selection procedure: how we select our sample
Experimental Design and Statistical Inference
• Experimental: controls influence of at least one variable
• Motivation behind experimental control: ability to make casual inferences
o To know when the outcomes of a study are due to only one variable and no others
• Extraneous variables: influence confused (confounded) with the variable were interested in,
namely, which treatment people received
o Randomised assignment: most widely used and respected solution
o Patients randomly assignment old and new treatment groups will result in groups that
differ on all variables by change alone unless the new and old treatments really differ
o Blinding means that the experimenter or participant doesn’t know which condition they
have been assigned to
• Blinding the participant ensures against auto-suggestion form the participant
regarding which condition they are placed in
• Blinding the experimenter ensures against unconsciously influencing the
participant or other outcome aspects
• Between-subjects
o These designs expose each unit to only 1 condition randomised assignment pertains to
this kind of design
• Repeated measures or within-subjects
o Expose all units to all conditions
o Example: effect of a drug on task performance
• Alcohol vs. no alcohol conditions: might compare the same people in both
conditions
• Internal validity handled by between-subjects experiments but not within-subjects
experiments:
o History: anything that may change between one measurement occasion and another
that is not under the experimenters control
o Maturation: any age-related process that can affect the dependant variable
find more resources at oneclass.com
find more resources at oneclass.com
