Lecture 18 - Graphing Data Flashcards
What graphing method is best for categorical data frequencies?
Histograms and percentage histograms.
What graphing method is best for binned numerical data?
Histograms and cumulative histograms.
When should error bars be used?
Whenever a mean of several measurements is displayed.
What do error bars indicate?
The variability or confidence interval for a data point.
How is using error bars different between SPSS and Excel?
In SPSS, they are a type of plot, whereas in Excel they can be added to a graph.
What values can be used for error bars?
SD, SEM & CI
What source of variability are error bars often associated with?
Inter-subject variability, but not necessarily - can be the biggest source of noise in an experiment.
When doing an error bar plot on SPSS with multiple IVs and a within-subject design, which options should be selected on the menu?
Clustered and ‘summaries of separate variables’.
For one IV - simple.
For a between-subject design - ‘summaries for groups of cases’.
In SPSS, what is the default setting for what the error bars represent?
95% CI.
When are scatter plots useful?
When there are several levels for an IV.
How do you add a line of best fit to a scatterplot in SPSS?
Click on a data point, goes blue, then Chart|Add Chart Element|Fit Line at Total.
How do you overlay additional data on a scatter plot in SPSS?
Graphs/Scatter/Overlay
Move 2 variables over at once for an x-y pairing.
How do you create a transform of some original data in SPSS?
Transform/Compute
Insert correct variable and enter appropriate equation, e.g. for a z transform.
How should multidimensional data be grouped?
However is appropriate given what you want to show. E.g. looking for gender differences or task improvement?
What is a con of 3-D bar plots?
It is difficult to see items at the back, especially if they’re smaller than bars in front.
What’s the problem with presenting multidimensional data in a surface plot?
They require elaborate plotting packages.
What different kinds of surface plot can be used to represent multidimensional data?
Mesh plot (grid-like) Mesh combined with intensity (grid-like and colour grading) Surface with intensity and contours (lines and colour grading)
What medium is a popular way to represent intensity?
Colour.
What is a problem with using colour to represent intensity?
It can potentially introduce effects that aren’t really there, e.g. identical MRI data painted with different ‘colour maps’ and thresholding can look very different.
How is it possible to represent four dimensional data?
Using a movie
What is the problem with 4-D data?
It’s so confusing that it’s probably better to devise a simpler experiment.
What should you do before plotting a graph?
Decide the key points to be conveyed and what you want to focus on.