A) Error sources in experimental research

Question 1

two main sources of bias we need to control for

Answer 1

• extranous variables
• confounding variables

Question 2

extranous variables

Answer 2

extra variables that are also varying adding noise in our data (eg ppt doing a cognitive test at different times, differences in lighting or differences in ability) this type of variable adds unsystematic variation in the data (adds noise)

Question 3

confounding variables

Answer 3

other variables that vary between our conditions (eg getting everyone in one group to take part in morning and everyone in the second group to take part in the evening). this type of variable adds systematic variation in the data (can cause ambiguous effect)

Question 4

example of a variable that adds systematic variation

Answer 4

confounding variables

Question 5

example of a variable that adds unsystematic variation

Answer 5

extraneous variables

Question 6

what is an unsystematic error caused by

Answer 6

• caused by all things that are varying other than manipulation, eg ppt mood, differences in ability, condition variations etc
• unsystematic variance cannot be totally removed, all we can do it make the experiment powerful enough to overcome it.

Question 7

how can we combat effects of unsystematic (random) error

Answer 7

• create more powerful manipulation (eg increase dosage of nicotine administered)
• testing large number of people
• using matched-pairs design

Question 8

what is a systematic error causes by

Answer 8

where we have a systematic bias between two groups that may affect the results (make it appear that there is. difference when there isn’t one, or no difference when there is one)
eg one group taking experiment outside in noisy place, second group indoors in quite room
makes it harder to interpret the difference between conditions (is the difference due to manipulation or the confound?)

Question 9

4 ways to control for confounding variables

Answer 9

-standardisation
-randomisation
-counterbalancing
-blind design

Question 10

standardisation meaning in terms of controlling confounding variable

Answer 10

to ensure environmental condition do not systematically vary between groups

Question 11

randomisation meaning in terms of controlling confounding variable

Answer 11

to ensure participants are split between groups without bias

Question 12

counterbalancing meaning in terms of controlling confounding variable

Answer 12

to avoid order (practise or fatigue) effects

Question 13

blind design meaning in terms of controlling confounding variable

Answer 13

to prevent the ppts (and researchers) expectations from confounding your results

Question 14

quasi experiment

Answer 14

comparing groups eg boys and girls, or individuals with developmental disorders to a neurotypical population

Question 15

natural experiment

Answer 15

comparing naturally occurring situations eg comparing individuals that have gown up in care to those who have not or those with a single parent to those with both parents etc

Question 16

why is it difficult to control extraneous or confounding variables in quasi or natural experiment

Answer 16

because they may have been naturally occurring as a product of belonging to that group