Lecture 2- Experimental Design 1

Question 1

What 2 out of the four science goals are key to experimental research?

Answer 1

Explanation

Control

Question 2

What does causation mean in an experiment?

Answer 2

One factor directly affects another factor

Question 3

To demonstrate causation we must demonstrate that…

Answer 3

1) Changing the first thing produces a change in the second thing
2) There is no other possible cause for the change (an intermediate)

Question 4

Representative sampling is achieved by….

Answer 4

Randomizing selection so that all members have an equal chance

Question 5

What is descriptive statistics?

Answer 5

Summarizing the data you receive from the sample e.g. averages

Question 6

What is inferential statistics?

Answer 6

Generalizing from the sample to the population

Question 7

What is the dependent variable in an experiment?

Answer 7

• The thing you are recording/ the measurement

- This depends on what the participant does

Question 8

What is an operational definition?

Answer 8

When there is no direct way to measure what you want e.g. happiness, intelligence.
Therefore, you need to chose a dependent variable/ measure that can indirectly stand in for your property of interest e.g. IQ scores for intelligence

Question 9

What does it mean if a dependent variable is valid?

Answer 9

It measures what it is supposed to. If an unintended component is influencing the score then you dependent variable is not valid. So a poor operational definition can result in a DV that is not valid.

Question 10

What does it mean if a dependent variable is reliable?

Answer 10

In the same conditions it consistently gets the same result with a small margin of error. If a DV is not reliable than there is a data error or bias and the data is also not valid.

Question 11

What does it mean if a dependent variable is bias?

Answer 11

When it is consistently inaccurate in one direction (too high or too low)

Question 12

A good DV is….

Answer 12

• valid
• reliable
• unbiased

if it not then any claims made from the data is pointless

Question 13

What is the ceiling effect?

Answer 13

When a task is too easy and so all scores are very high masking any difference between two groups

Question 14

What is a floor effect?

Answer 14

When a task is too hard and so all the scores are too low masking the difference between the scores of two groups that may have otherwise appeared.

Question 15

What does a data type determine?

Answer 15

What sort of analysis can be performed on data and therefore what type of conclusions can be drawn.

Question 16

What data types can data be spilt into (think tree)?

Answer 16

Data can either be numerical or categorical (not numbers)
categorical data can then be spilt into ordered (e.g. military rank, with each step up it gets better/ gain importance) and unordered (e.g. gender, political parties, there are different groups but none are objectively better than the others)

Question 17

What scale is numerical data associated with? What does this scale mean?

Answer 17

Interval or ratio.

Interval scales categorize, order or establish an equal unit of measurement. Means you can know how much more. 0 is just another point on the scale and does not mean an absence of the thing being measured.

Ratio scale categorize, orders and establishes an equal unit in the scale, and contains a true zero point where there is an absence of the thing being measured. This means you can make comparisons like ‘twice as long’ which you can’t do with interval because there is no true zero.

Question 18

What scale of measurement is categorical, order associated with? What does this scale of measurement mean?

Answer 18

Ordinal.
This categorizes + orders categories, bigger is better but distance between points in the scale are not considered equal. Therefore can’t say how much better.

Question 19

What scale of measurement is categorical, unordered associated with? What does this mean?

Answer 19

Nominal scale

categorizes without ordering