RMA: WEEK 7

Question 1

Variable

Answer 1

• Something that varies/changes

Question 2

Independant variable

Answer 2

• The variable which is manipulated by the researcher

Question 3

Dependant variable

Answer 3

• The variable which is measured to see how much of an effect the IV had on it

Question 4

Operationalised variable

Answer 4

• Specific statement about how a variable will be measured to represent the concept in the study
• e.g: “Ppt listens to music while studying” is too broad
  “Ppt studies for 4 hours a day seated at a table listening to the Beatles using headphones at volume X

Question 5

Problems with defining variables

Answer 5

• Objectivity/Subjectivity: not all variables can be measured objectively
• e.g. mood (happiness), social behaviour (like shyness) > can only be measured subjectively
• Testability: how far can we measure the variable for an example anxiety, activation of words in mind

Question 6

Types of measurement: Scales

Answer 6

• Measurement: way to describe real life factors using numbers
  1. Nominal scales
  2. Ordinal scales
  3. Interval scales
  4. Ratio scales.
• Scales differ in relationship based on the properties of the number + property of what is being measured

Question 7

Nominal scales

Answer 7

• Least amount of detail + in named categories + is discrete (only 1 item appears in 1 cat) > most basic form of data. Does not look at individual ppt, groups them together (boys or girls, not ppt 1,2,3)
• EG: no. of boys + girls who did and did not conform in a line test
• number assigned is just a label
• No relationship between size of number and attribute measured > e.g: number 4 bus is a label + does not mean it is twice as much as the no.2 bus or is bigger

Question 8

Ordinal scale

Answer 8

• Order of size of no. = order of size of attribute measured
• Data is put in order but only relative ranking + distance between scores can vary > e.g: medals from different schools can be ranked but the difference between 1 and 2 may differ from that between 3 and 4

Question 9

Interval scale

Answer 9

• Equal intervals on the scale = equal intervals in the property measure > equal space between 1,2,3
• Measure of magnitude/size
• Zero point on the scale doesn’t mean anything as values can be minus as well (doesn’t stop at 0)
• e.g: difference between 10 degrees + 20 degrees is the same as that between 20 and 30 degrees but we cannot say 20 degrees is twice as warm as 10 dg > 0 is not meaningful as we can go to -10 dg for example

Question 10

Ratio scale

Answer 10

• Interval scale but zero has meaning + cannot go below 0
• e.g: reaction time to pressing a button cannot go below 0
• can say if one person take 500 milliseconds and another person takes 1000 milliseconds, it is possible to say the second person was twice as slow.

Question 11

How to choose what scale to use

Answer 11

• Nominal scales only show a difference between the number but has no significant magnitude, equal intervals or meaningful zeros
• Ordinal scales show significant difference and magnitude as the numbers can be ranked but no equal interval or meaningful zero
• Interval scales show significant difference, magnitude and equal intervals as the size can change + distance between the gap is equal but no meaningful 0
• Ratio scale shows significant difference, magnitude, equal intervals and meaningful 0 as the number has size and equal gap between them + cannot go below 0

Question 12

Measures of central tendency

Answer 12

• Use this when we need an average of overall results
  Mean- add all no. and divide by amount of no. present
  Mode-most frequent no.
  Median- midpoint of sample (50% above + below N), Order numbers + find the one in the middle, if 2 add both and divide by 2

Question 13

Outliers

Answer 13

• score which vary a lot from other scores > extreme values

- if the researcher thinks there may be outliers, they should use the median + not the mean

Question 14

Advantages + disadvantages of mean

Answer 14

+ Includes all data (more representative) > should use mean if all values in dataset are important
- Strongly affected by outliers so if included in mean the data wont be accurate

Question 15

Advantages + disadvantages of median

Answer 15

+ Less affected by extreme values > use median if there are outliers
+ best for ordinal data as it is an ordinal measure (ranked)
- Doesn’t work with small data > doesn’t incorporate all data (generally less representative if no outliers)

Question 16

Advantages + disadvantages of mode

Answer 16

+ Less affected by extreme values
+ Good for data on nominal scale > e.g most frequently used bus is easy to see
- Distribution can be bimodal > 2 numbers occur equally as most frequently
- Only uses one value > not a good representation
- Some dataset doesn’t have a mode > e.g: in reaction times, individual times may only come up once
- Mode can be atypical > for an example, it may be not be central to the data + not accurately reflect average

Question 17

What measure of central tendency should be used for what scales?

Answer 17

• Nominal scale is best used with mode
• Ordinal scale prefers median but mode can be used
• Interval scale prefers median or mean but mode can be used (depends on data)
• Ratio scale prefers median or mean but mode can be used (mean is best unless data doesn’t fit w/ it)