# W3 - Descriptive Statistics Flashcards

1
Q

Define nominal scale

A

Subjects grouped into mutually exclusive categories.

2
Q

Examples of nominal scale

A

Eye colour

Year of birth

3
Q

Define ordinal scales

A

Ranks subjects/scores in order.

Doesn’t indicate how much better 1 score is to another.

4
Q

Example of ordinal scales

A

University rankings

5
Q

Define interval scales

A

Equal units or intervals between data points on a scale, but there’s no zero point.

6
Q

Example of interval scales

A

Temperature

7
Q

Define ratio scales

A

Equal units of measurement

Established 0 point.

8
Q

Example of ratio scales

A

Mass

Length

Height

9
Q

What are measures of central tendency used for?

A

To summarise or provide meaning to a whole dataset

To provide a ‘central’ value

10
Q

What doesn’t central tendency show?

A

Distribution of scores

11
Q

What is the SD?

A

Avg distance away from mean

12
Q

What allows you to summarise data?

A

Measures of central tendency + dispersion

13
Q

What graph would you make if you had a lot of data?

A

Frequency distribution plot/graph

14
Q

Properties that all normally distributed data show

A

Symmetrical

Equal mean, median + mode

15
Q

What do Z-scores do?

A

Describes a values relationship to the mean of a group of values.

16
Q

How is the z-score measured?

A

In terms of SD from the mean

17
Q

Equation to calculate z-score

A

(X - mean) / SD

18
Q

Where would z-score of 0 be on a normal distribution curve?

A

In the middle of the bell curve

19
Q

What do z-scores + normal distribution curve allow you to identify?

A

Ind. score rel. to pop

Probability of scoring a given z-score or higher/lower

What score is needed to be in a given % of the pop

20
Q

What are the important z-scores?

A
1. 96
2. 58
3. 29
21
Q

Z-score

1.96

A

95% of scores fall w/in 1.96 Z scores or SDs above/below mean

22
Q

Z-score

2.58

A

99% of scores fall w/in 2.58 Z scores or SDs above/below mean

23
Q

Z-score

3.29

A

99.9% of scores fall w/in 3.29 Z scores or SDs above/below mean

24
Q

Type of deviation from normality

Skewness

A

A measure of lateral deviation from normality

25
Q

Are skewed distributions symmetrical?

A

NO

Most scores are high/low with small % of scores away from the majority

26
Q

+ively skewed

A

“Tail” of distribution positioned at upper end of score continuum

27
Q

-ively skewed

A

“Tail” of distribution positioned at lower end of score continuum

28
Q

Type of deviation from normality

Kurtosis

A

Measure of vertical deviation from normality

29
Q

What are the distribution curves under kurtosis (type of deviation from normality)

A

Leptokurtic

Platykurtic

30
Q

KURTOSIS

Leptokurtic curve

A

More peaked than a normal distribution curve

31
Q

KURTOSIS

Platykurtic curve

A

More flat than a normal distribution curve

32
Q

KURTOSIS

What is meant by mesokurtic

A

Level/Distribution of what you would expect from the normal distribution

33
Q

What is a good guide for the coefficient of skewness?

A

-1.0 - 1.0

34
Q

What is a good guide for the coefficient of kurtosis?

A

-1.0 - 2.0

35
Q

What are the measures of central tendencies?

A

Mode

Mean

Median

36
Q

When would you use the mode

A

When data is categorical

37
Q

When would you use the mean?

A

If you want the total

or

If distribution isn’t skewed

38
Q

When would you use the median?

A

If distribution is skewed

39
Q

Define a predictor variable

A

Variable thought to predict an outcome variable.

(Like the independent variable).

40
Q

What can variables be split into?

A

Categorical

Continuous

41
Q

What comes under categorical variables?

A

Binary variables

Nominal variables

Ordinal variables

42
Q

What comes under continuous variables?

A

Interval variables

Ratio variables

43
Q

Define criterion validity

A

Establishing that an instrument measures what it claims to measure

• by comparison to objective criteria.
44
Q

What is said to assess concurrent validity?

A

Comparing a new test to an existing one to see if they produce similar results.

45
Q

What is said to assess predictive validity?

A

When data from new instrument is used to predict observations at a later date.