Lecture 3 - Variables

Question 1

What are independent variables?

Answer 1

Something we change in the experiment or is naturally changing (e.g. training user to do something and seeing if improved).

it can be a characteristic or circumstance

Question 2

Experiments with independent variables are often called ….

Answer 2

factorial experiments

Question 3

Some examples of independent variable characteristic?

Answer 3

e.g. screen size , feedback modality ,
e.g. gender , expertise , age

Question 4

Some examples of independent variable circumstance?

Answer 4

e.g. weather , lighting , background noise

Question 5

If we have an independent variable then what is our research question typically based on?

Answer 5

On the independent variable changing and the effect of this change on the dependent variable.

Question 6

What is a dependent variable?

Answer 6

In HCI the DV is often a measured human behaviour. The
measurement depends on what the participant does.

Question 7

Examples of dependent variables?

Answer 7

• anything that is observable and measureable that we are not changing

e.g. typing speed, eye movements , facial expressions , response to questionnaire

Question 8

In HCI of what do we pick characteristics for the independent variable.

Answer 8

These will be the characteristics of the computer or the person.

Question 9

Dependent variable is the point of the whole experiment ?

Answer 9

Yes, we want to see the effect of our changes on it, so that we can come to a conclusion

Question 10

Is it a good idea to have a lot of IDs?

Answer 10

No , usually 2 is the max , more means it will be hard to aggregate the data.

Question 11

What is an effect number?

Answer 11

The number of observed effects of the experiment, it increases as the number of independent variables increase. We have to consider every effect in the study !

Question 12

What is the effect number for 3 independent variables?

Answer 12

3
- one for id1
- one for id2
- one for interactive id1 x id2

Question 13

What are control variables?

Answer 13

Variables not under investigation byt might influence participant behaviour.

e.g. keyboard angle , chair height , display size

Question 14

What should we do about control variables?

Answer 14

Control them, as to minimise their effect on the study.

• we can also have strict recruiting and exclusion critertia

Question 15

By controlling control variables what happens to the validity?

Answer 15

The internal validity goes up, but external decreases

Question 16

What are random variables?

Answer 16

Some variables that are allowed to vary in order to generalise the results.

-> it is impossible to control everything so there will always be random variables

e.g. height , social disposition

Question 17

Is there a tradeoff between control variables and control variables in experiments?

Answer 17

Yes , we will typically have a threshold for this

Question 18

What are confounding variables?

Answer 18

a variable / condition / circumstence that changes systematically with the independent variable

e.g. practice , prior experiance

Question 19

Why are variables confounding?

Answer 19

As they prevent the possiblity of a cause and effect relationship being inferred from the results, so they need to be controlled.

Question 20

Obvious example of confounding variable?

Answer 20

Experiance with a certain keyboard layout.