Flashcards in Sampling & Measurement Deck (20)
What is NOMINAL data?
Give an example.
It describes categories of a variable.
Latin for “name”
Independent variables tend to be nominal.
Eg. Eye Colour, Short and Tall
What is ORDINAL data?
Give an example.
It gives ranges of a variable or shows the “order” of the variables.
It cannot tell us the differences between the orders.
Has no absolute zero.
Eg. 1st, 2nd, 3rd place in a race - we know where they placed but not how much faster one person was than the other.
Or income brackets - $40-$50/hr $51-60 - you don’t know if someone is on 49 or 41 etc.
What is INTERVAL data?
Give an example
It shows us where a variable measures along a scale of equal increments.
No absolute zero - it can be negative.
Eg. Temperature - degrees are equal and we can have negative degrees.
What is RATIO data?
Give an example of both continuous and discrete ratio data.
It is the measurement along a continuous scale.
It has a true zero and cannot be negative.
Eg. Continuous - Time or age (no negative)
Discrete - number of children you have
What does it mean to OPERATIONALISE a variable?
To decide how you will measure it so it can be quantified and recorded.
What are levels of measurement?
The relationship between what is bearing measures and the numbers that represent what is being measured.
What is SCALE data?
Give 2 examples.
Is is data that can take on any value along the measurement scale.
Interval (temperature) and Ratio (Likert Scale)
What is CATEGROICAL data?
Give 2 examples.
Categorical data are distinct entities.
Eg. Nominal (eye colour) and Ordinal (annual income bracket)
List 7 ways to increase RELIABILITY of a measurement tool
- Increase number of items or observations
- Eliminate ambiguity (eg weight, lb or kg)
- Moderate difficulty (eg for ESL people)
- Minimise effects of external events
- Standardise conditions (eg library vs cafe)
- Standardise instructions
- Standardise scoring
What does RELIABILITY refer to?
Consistently of measurement. The same results can be obtained with the same input. It is essential for a good measurement tool.
What does VALIDITY refer to?
The extent to which a test/measurement tool measure the content that we are interested in
What is MEASUREMENT ERROR?
It is the discrepancy that exists between the data found and the TRUE value of measurement
Eg. Someone’s IQ score and their actual level of intelligence
How do you test the RELIABILITY of a measurement tool?
Through test-retest reliability. When you compare scores on a test from time 1 to time 2.
Pearson’s (r) - the correlation co-efficient - will tell you how similar the results are and therefore how consistent/stable the measurement is over time.
What is PROBABILITY SAMPLING?
Where the probability of any individual in the population being selected is known. Ie. you know the size of the population, the distribution of related characteristics wiring the population and have access to them.
What is NON PROBABILITY SAMPLING?
Where the likelihood of any one individual being selected from the population is unknown. Ie. you do not know the size or exact details of the population
Define SIMPLE RANDOM SAMPLING and it’s 4 steps
Where each individual in the population has an equal and independent chance of being selected as part of the sample
- Define the population
- List all members of the population in a random order
- assign each member a number
- select participants through a ... random number generator or similar
What is STRATIFIED SAMPLING and when is it used?
It ensures that the profile of the population is fairly represented in the sample if specific characteristics related to what is being studied (eg, sex, age) are not distributed randomly in the population.
Eg. In a study on male/female attitudes toward study in psyc ACAP, 60% females to 40% males if the sample was 10, would require 6 females and 4 males to reflect the actual distribution in the population.
What is the benefit of a larger sample size?
The larger the sample the easier it is to detect significant differences
What is CLUSTER SAMPLING?