Summary Research Methodology 1 - Exam period 4

Chapter 4 Andy Field - Introduction to SPSS
- IBM SPSS is software used most often to analyse quantitative research data
- Available to buy and download and freely available at the RU computers
- SPSS consists of 3 windows
o Data editor
o SPSS output viewer
o Syntax editor  for every line there is an underlying syntax
 A kind of code which you could use if you want to conduct the exact same
analysis again, without having to click on all the separate windows, and if you
want to keep track of what specific analyses you did
 You can view the syntax in your output file, but If you want to work with the
syntax, you can also create a separate syntax editor
- Data from different things go in different rows
- Data from the same things go in different columns


Chapter 6 Treadwell – Chapter 1 + 19 Field
Measures can be…

- Nominal – labels
- Ordinal – rank order
- Interval – numeric scale
- Ratio – numeric scale with a 0 point

Statistics

- Look at the variation in your participants’ data
- The design of your study determine how you will analyse this variation
- The difference between within-subject and between-subjects designs
o Between subjects design
 Subject 1  pre-test  treatment A  post-test
 Subject 2  pre-test  treatment A  post-test
 Subject 3  pre-test  treatment B  post-test
 Subject 4  pre-test  treatment B  post-test
 Etc.
o Within subject design (repeated measures)
 Subject 1  pre-test  control  test  A  test  B  post-test
 Subject 2  pre-test  control  test  A  test  B  post-test
 Etc.

- Systematic variation: variation in the data due to a manipulation
- Unsystematic variation: variation resulting from random factors

, o It is preferable to keep it as low as possible, so that most of the variation will be
caused by your manipulation
 Use a within-subject design
 Random assignment of your participants to the different conditions
 Randomising the order of conditions of your study
- Systematic variation / Unsystematic variation = Effect / Error
o This ratio is called a test statistic
Univariate Analysis

- Measures of central tendency
o Mean = add scores: divide total number of scores
 Can be used for interval/ratio variables
o Mode = most frequent score
 Only measure of central tendency which can be used for nominal variables
 Not sensitive to outliers
 There may be more than one mode
 2  bimodal
 2+  multimodal
o Median = middle score of ranked scores if number of scores is uneven
 Can be used for ordinal and interval/ratio variables
 Not sensitive to outliers

Measures of dispersion

- Minimum: lowest score
- Maximum: highest score
- Range: maximum value minus minimum value (affected by extreme values)
o Interquartile range = cut off bottom and top 25%, calculate middle 50%
Deviance = looking how different each score is from the centre of the distribution

o Variance = average squared deviation from the mean
o Can be used for interval/ratio variables, but hard to interpret
o Standard deviation = square root of variance
o Often used because it is easier to interpret than variance
o The larger the SD, the wider the range of values of either side of the mean
o Can be used for interval/ratio variables
o Only comparable with variables that use the same type of measurement
o Z-score = the number of units of SD any one individual value is above or below the mean
o Standardized scores to allow comparison between different variables