Week 3

Question 1

What is reliability?

Answer 1

consistency or repeatability of measurements (should get the same results every time)

Question 2

What is validity?

Answer 2

accuracy of measurements, you are measuring what you are supposed to measure

Question 3

Types of Reliability?

Answer 3

Intra-rater reliability - same measurement by same person every time

Inter-rater - measurement by different people at same time point

Test-retest reliability - same measurement undertaken at different time points by different instruments

Question 4

Types of Validity

Answer 4

Face validity - does it appear to measure what is aims to measure

Content validity - does the measurement cover everything it should?

Criterion validity - how does one measure relate to an outcome

Construct validity - how well does the measurement truly reflect what it claims to measure

EXTERNAL VALIDITY (generalisability) 
how well can you take the research findings of the sample and relate it's to the population and let it represent it.

INTERNAL VALIDITY (relates to the methodology of the research)
can you be confident that the research was well conducted, where the results because of exposure/intervention? 

internal validity is affected by
- CHANCE
this is a random error
it can’t be eliminated only minimised through sample size - increase it

• BIAS
  this is a systematic error, error in the way the research was undertaken (can be eliminated)
  **many types
  reporting bias: positive only published)

selection bias: who gets selected for sample, sample should represent wider population

allocation bias: who gets into the intervention and control group

attrition bias: dropout from the study

measurement bias: errors in measuring

maturation bias: changes which occur naturally over time

• COFOUNDERS (these are variables that weren’t taken into account and hence influence outcome)

Question 5

Why do we need reliability and validity?

Answer 5

if the evidence is not underpinned by reliability sand validity then it is prone to errors and misinterpretation

Question 6

What is Normal Distribution?

Answer 6

• means it is a naturally occurring phenomena
• there is a symmetric distribution, equal on both sides of the central peak
• it is a unimodel (mean, median, mode)
• it is the most important probability distribution in statistics

Question 7

Population vs Sample

Answer 7

Population is everyone you are interested in, represented by a capital N

Sample is a subgroup of the population, represented by lower case n

Question 8

Types of Data

Answer 8

CATEGORICAL

• ordinal data (can be logically ordered or ranked eg. academic grades, clothing size)
• nominal data (opposite of ordinal, cannot be ordered)
  eg. gender, culture, religion)

NUMERICAL

• continuous data (values are between are certain set of numbers eg. height, weight, temperature)
• discrete data (measured as whole units, numbers eg. people, number of children)

Question 9

Measures of Centrality and Dispersion

Answer 9

First of all, it depends on whether the data is symmetric or skewed

Symmetric
- mean (centrality) and standard deviation (dispersion)
**mean can’t be used for categorical data
Skewed
- median (centrality) and interquartile range (dispersion)