# How to analyze quantitative data gained from repeated measure experiment?

I am new here, so I hope this question is on topic and comprehensive. I understand a little bit about statistics, but my knowledge of what methods to use in real case scenarios is a rather limited.

I am wanting to statistically analyse whether a physical keyboard is faster and more accurate to type on than a touchscreen keyboard. I have completed a cross-over repeated measure experiment that contained a control (membrane keyboard and smartphone touchscreen kb) and a experiment (mechanical keyboard and tablet touchscreen kb) where subjects were randomly divided into groups, like the following:

                 Session 1  |   Session 2
-----------------------------------------
Group A ---> Control    |   Experiment
Group B ---> Experiment |   Control


The data acquired provided me with a single metric that was the word per minute minus the percentage of errors (accuracy).

With the data how can i statistically analyse whether or not the physical keyboard is faster and more accurate to type on than a soft / touchscreen keyboard?

Is it even relevant to even have tested on the mechanical and tablet keyboard if all i want to obtain information regarding which input method has greater accuracy and speed?

Thanks in advance.

• Can you expand on you last paragraph? – mdewey Oct 19 '16 at 7:16

## 1 Answer

considering your research example I would describe it as a 2x2 design with one depended variable.

2x2 stands for two factors (independend vars) on 2 levels each - one iv is your experimental condition and the other iv is session.

how can i statistically analyse whether or not the physical keyboard is faster and more accurate to type on than a soft / touchscreen keyboard?

Look for a two-way repeaded measure anova (in SPSS it's under General Linear Model), where in fact the main effect of experimental condition is the only thing you are interested it.

Also, please consider that what I just wrote is a simple answer to "what analysis you could use - among others, without any indication on other aspects of design methodology etc".