PSY294 Statistics for Psychology

1
PSY294 Statistics for Psychology
Deadlines
Assignment C (C1-C2) to be handed in by 4.30pm
Thursday 15th (TODAY) of December (in
Psychology Department) Your answers/discussion
of Practical exercise (P1-P5) to be handed-in by
4.30pm on Friday 16th of December (in School of
Maths and Stats (Merz Court)). Please use the
cover sheet handed out in todays lecture. N.B.
Please hand your Practical Folder in to the MS
General Office on the 5th Floor of Merz Court)
2
PSY294 Statistics for Psychology
Revision SPECIMEN EXAM PAPERSOLUTIONS (handed
out today) PSY294 (2004-2005) PAPER PSY297
(2002-2003) PAPER www.ncl.ac.uk/exam.papers Week
12 lecture (Revision Week) PSY294 (2004-2005)
paper anything else you would like to cover email
to suggest topics
3
PSY294 Statistics for Psychology
Example of interactions

• Consider a manufacturing process
• If we increase the temperature by 1 degree
• the process works 5 times more efficiently
• If we increase the pressure by 1 unit the
  process works 8 times more efficiently
• If we simultaneously increase the temperature
  by 1 degree and the pressure by 1 unit the
  process works 20 times more efficiently.

4
PSY294 Statistics for Psychology
Extensions to ANOVA
Last week we discussed factorial experiments and
these can be extended to several factors and
their interactions Minitab can deal with data
based on up to 5 factors and their
interactions
5
PSY294 Statistics for Psychology
Extensions to ANOVA
We may have to deal with nesting or structural
limitations If we have unbalanced experiments,
we might have to use Generalised Linear Models
(GLMs) for our analysis
6
PSY294 Statistics for Psychology
Transformations
The methods of correlation, regression and ANOVA
are based on assumptions of Independence,
Normality and Common Variance Non-parametric
variants are available when these assumptions
are deemed to be invalid
7
PSY294 Statistics for Psychology
Transformations

• If the data is deemed to be non-Normal, sometimes
  it can be made Normal by a simple transformation
  of the data.
• Typical transformations are
• square root
• reciprocal
• logarithm

8
PSY294 Statistics for Psychology
Multivariate methods
Cluster analysis used to classify
individuals and separate them for further
analysis Discriminant analysis similar to
cluster analysis Principal Component Analysis
(PCA) method for reducing the number of
variables used to characterise a data set
9
PSY294 Statistics for Psychology
Overview

• I hope that this course has helped you to
• appreciate the need to use statistical methods
  in the analysis of data arising in investigations
  in Psychology and allied subjects

10
PSY294 Statistics for Psychology
Overview
On a practical note, you should now be able to

• state and test hypotheses relating to
  correlation, regression and the various types of
  analysis of variance and their non-parametric
  equivalents.
• understand the purpose of these techniques and
  be able to distinguish the circumstances in which
  each is used.

11
PSY294 Statistics for Psychology
Overview
And finally

• Be able to state conclusions both in statistical
  terminology and in non-statistical terms related
  to the context of the experiment.

12
PSY294 Statistics for Psychology
EXAM DATE
Thursday 26th January 2.00pm Freemen's Hall