Quantitative Research: Erik Geelhoed
- Questionnaires : Online forums that collect data from users
- Technology Logs: When an app was started ended, what buttons were selected
- Observations When, How Long, How Often
1.Nominal Data (Frequency) Data
eg: Tally Red or yellow cars in an hour.
25 red cars and 10 yellow cars
Is this statistically significantly different?
2. Ordinal Data – Prioritise “what do you like better apples, bananas or clementines
Forced to make a choice and order them.
This is the 3 alternative forced choice (3 AFC)
Statistical Analysis – a great many
Correlations which can lead to a market segmentation.
We do not know how much one option is preferred over the other choice.
3. Scale Data – ratings “where 1 is low and 5 is high”
Scale data gives the most powerful statistical analysis.
Analysis of Variance (ANOVA)
Evaluate more than one variable at a time
correlations between two variables
cluster analysis/market segmentation
Eg: A question in a questionnaire.
- Central Tendency
Mean (average value,scale data) What is the spread ?
Median (The 50% mark, scale and ordinal data)
- Measures of spread
standard deviation – used most often
Mau (in the middle)
Sigma (standard deviation)
either side is normal….
- Normal Distribution (Gaussian Distribution)
We are looking for a significant of 5% or less which indicates the level of confidence we can have in that the result is significant and not by chance.
p value of 0.05
Non Parametric Date
Carry out Usability Research
UX research informs interface design. So for example, if the designer needs to design an interface which young children will use then using a font size of 14 is generally considered to be the optimum font size. This is informed by quantitative research in which the user is observed and different font sizes tested.
In order to keep up with current discourse in UX Design research I follow key theorists and designers in the field. This includes the Nielson Norman group.