Zusammenfassung der Ressource
Research Methods
- Evaluating research
- 1. Identification of the source
- when, where, peer reviewed?
- 2. Analysis of argument
- Theory? alternatives? concepts defined? hypothesis and questions
appropriate?
- 3. Examination of methodology
- Particpants representative? sample? design? reliability and
validity
- 4. Evaluation of results
- Appropriate stats? reliable and clearly presented results? results support
hypothesis?
- 5. Analysis of conclusions and interpretations
- Conclusions logical? overly speculative? alternative explanations explored? does it just re-state the
results?
- Experimental designs
- Independent and dependent variables
- 1 tailed vs 2 tailed hypotheses
- 1 tailed - know the direction of outcome
- 2 tailed - don't know direction of outcome e.g. is there a difference at all?
- Error - influence of other variables other than the IV
- experimental designs involving groups of participants
- Between participants
- 1 factor e.g. mood level and 2 or more conditions e.g. 30 mins exercise vs no
exercise
- Matched - when there are likely to be errors
- Within participants/repeated measures
- Each participant takes part in all conditions - looking at changes within person
- More complex designs
- Multi-level
- more than 2 conditions from the IV
- Factorial - more than one IV
- Multivariate - more than one DV measured
- Mixed - within and between participants measured
- Covariate - covarying factors on DV subtracted out
- Single participants
- AB design - A (baseline) B (intervention)
- ABAB design. Gap of time in between and repeated
- Multiple baseline - lots of baselines taken then measured over time with
intervention
- Changing criterion design - changing intervention and measuring over time
- Describing and displaying data
- Descriptive vs Inferential statistics
- Nominal data - labels, names
- Ordinal - 1st, 2nd, agree disagree
- Interval vs ratio - equal data. ratio has
true zero
- Measures of central tendancy - mean, median,
mode
- Measures of spread - range, variance,
SD
- Frequency distributions
- Statistical tests of difference
- 1 independent variable 2 conditions
- Wilcoxon matched-pairs signed ranks test
- nonparametric
- ordinal data
- experimental design - repeated
- Mann-Whitney U test
- experimental design - independent
- ordinal data
- nonparametric
- Related t-Test
- experimental design -
repeated
- interval/ratio data
- normal distribution
- unrelated t-Test
- experimental design -
independent
- normal distribution
- interval/ratio data
- there are others but ill have a job remembering those>>>
- Correlation and regression
- positive vs negative vs no correlation
- regression
- predicting score using the correlation (once you know X you can
predict Y)
- Multivariate analysis
- ANOVA
- Analysis of Variance
- A method of statistical analysis for analysing experiments where there are two or more variables. Also examines any interactions between
variables
- one-way related
ANOVA
- repeated measures or matched perform under all conditions of the
variable
- one way unrelated
ANOVA
- unrelated participants perform under each
condition
- two-way related
ANOVA
- complex parametric test examining interaction between 2 variables e.g noise level of child am vs pm and how well
they do maths
- two way unrelated
ANOVA
- unrelated - which treatment works best and is the SLT experienced
or not?
- two way mixed
ANOVA
- Statistical tests for Nominal data
- Binomial sign test
- each condition 50% - better or worse? yes or no?
- Binomial test
- 2 conditions + expected and observed - significant difference?
- Fisher's exact test
- 2 independent samples compared for a correlation
- Chi-square test
- Like fisher's exact test but bigger data and goodness of fit
- tests of correlation/association
- chi-square
- nominal
- non-parametric
- spearman's rho
- ordinal
- non-parametric
- pearson's r
- interval/ratio
- parametric