Lecture 1 How to Interpret ‘TABLE 1’

Question 1

Numerical data types may be what 2 types?

Answer 1

continuous (many decimal places – e.g. height)

discrete (‘whole’ observations - e.g. Age in years, count of people in the room)

Question 2

what is the mean?

Answer 2

•The mean gives a sense of the centrality of the data

The mean is the average of a data set

•Mean = Sum of the observations/n

Question 3

what is SD?

Answer 3

•The SD describes the spread of the data

the average ‘spread’

Question 4

what is the median?

Answer 4

middle of the set of numbers.

Question 5

what SD cover how much of the population?

Answer 5

• ± 1SD : about 68% of the population
• ± 2SD : about 95% of the population
• ± 3SD : about 99% of the population

These found as the area under curve within each section, just like for a histogram

Question 6

• The mean is not a good measure of ‘most’ of the information
• Often the SD is not sensible either – because it relies on the mean

what can you use instead?

Answer 6

•Use medians and interquartile range (upper and lower quartiles) instead

Question 7

Categorical data:

•Data that can be put into boxes

Ordinal = ???

Nominal = ???

Dichotomous = ???

Answer 7

Ordinal – there is a sense of order (cancer stages)

Nominal – no order (blood groups)

Dichotomous – only 2 options

Question 8

what is Ordinal Data?

Answer 8

Data that has an order but may not have a consistent ‘linear’ scale*. Eg do you find this lecture helpful on a Likert Scale ranging from 1 (no help) up to 5 (really helpful)

*The gap between 1 and 2 may not always be the same in respondents’ mind as between 4 and 5

Question 9

how should ordinal data be summarised?

Answer 9

• The mean and SD rely on the data having a distribution – so no good unless there are many ordinal points that you are certain are equally spaced
• The median and IQR would be better but again only really meaningful if there are several ordinal points
• If only 3-4 may have to just report proportions of each category

Question 10

Nominal data has many categories represented as n as a _____________

Answer 10

percentage

Question 11

Consider the Visual Analog Score (VAS):

•A Visual Analogue Scale (VAS): an instrument that tries to measure a characteristic or attitude that is believed to range across a continuum of values and cannot easily be directly measured e.g. pain

What data type is VAS?

A)Numerical – continuous

B)Numerical – discrete

C)Categorical – dichotomous

D)Categorical – nominal

E)Categorical – ordinal

Answer 11

E

Spaces between a discrete variable is the same = parity of 3 to 4 is the same as a parity of 2 to 1, evenly spaced

An ordinal variable may not be evenly spaced

Question 12

Most appropriate summary of VAS

A)Mean and standard deviation

B)Median and standard deviation

C)Mean and inter- quartile range

D)Median and inter- quartile range

E)Proportions and percentages

Answer 12

D

Could be E as it is ordinal data but because you have several points maybe more sense if you use the median