Everything (Semester 1) Flashcards
A research question must have
- A research problem/ construct
- Context
- Unit of analysis
Primary objective
To determine….
Secondary question
To determine…
Will have categories
e.g. gender, age, place of residence
Constructs
theoretical concept, theme, or idea based on empirical observations
e.g. level of trust, brand loyalty, brand awareness, company transparency
Univariate
- Tests a single population parameter.
-1 construct and a mean
e.g. H1: The average IQ of BUS3008W students is greater than 100.
Difference between groups
- Comparing results found from 2 or more different groups who were measured on the same variable.
- 1 construct and 2 or more group
e.g. H1: There is a difference in the brand perception of users and non-users of Samsung Mobile Phones
Difference between measures
- Comparing results found between different measures/variables within the same group.
- 2 constructs and 1 group
e.g. H1: High-income consumers differ in their brand loyalty and perceived quality of granulated coffee brands.
Relational/Correlation
- Tests for correlation between 2 (or more) constructs or groups.
- Requires 2 (or more) constructs/groups.
e.g. H1: There is a relationship between warm coloured packaging and purchase intent.
Directional Relational
used to predict whether one variable causes an increase or a decrease in another variable.
-1-tailed is a relationship that is either positive or negative.
Non-directional relational
2-tailed is when the direction of the
relationship is not indicated.
Static Group
- pre-experimental design
-No Randomisation
- Two groups of test units: the experimental group (EG) and the control group (CG); only the EG is exposed to a treatment.
- The treatment effect is the difference between the two measurements.
- observation taken once for both groups
- Need to control for Selection Bias and Mortality.
-LOOK AT PICTURE
Pretest-Posttest Control Group
-true experimental design
- Randomisation
- Includes an EG and CG, measured before and after treatment.
Treatment effect is measured as: (O1 – O2) – (O3 – O4) - Controls for all extraneous variables with the exception of Interactive Testing Effects.
–LOOK AT PICTURE
One-group Pretest- Posttest
- No randomization
- One test unit is measured twice, before and after exposure to a treatment
- The treatment effect is the difference between the two measurements
- Need to control for extraneous variables
- LOOK AT PICTURE
History (H)
Specific events that are external to the experiment but occur at the same time as the experiment
Maturation (MA)
Similar to history, but refers to the changes in the test units themselves that occur with the passage of time.
Main Testing Effect (MT)
An effect of testing occurring when a prior observation affects a later observation
Interactive Testing Effect (IT)
When a prior measurement affects the test units response to the treatment
Instrumentation (I)
When there is a change in the measurement instrument or observers
Statistical Regression (SR)
When test units with extreme scores move closer to average scores during the course of experimentation.
Selection Bias (SB)
Occurs when there is an improper assignment of test units to treatment conditions
Mortality (MO)
When test units are lost during the course of experimentation
Ways to control extraneous variables
- Randomisation
- Matching
- Statistical Control
- Design Control
5 steps of the sampling design process
1.define the target population
2. Determine the sampling frame
3. Select the sampling technique
4. Determine the sample size
5. Execute the sampling process
Bayesian Sampling approach
Respondents are selected sequentially and data on previous respondents is used
Traditional approach
Sample is collected before the data collection begins and it is the most common approach
Probability sampling
Select respondents by chance with a fixed probability of each being selected
Non-probability sampling
Sample is selected based on the personal judgement of the researcher
Three basic research designs
- Exploratory= nonprobability
- Descriptive= non-probability and probability
- Causal = probability
Convenience sampling technique
Samples collected at the convenience of the researcher
-non probability
Judgemental sampling
Samples collected at the judgement of the researcher
-non-probability
Quota sampling
A two stage restricted judgement sampling
1. Develop control categories
2. Selected sample elements based on judgemental sample
Snowball sampling
Carried out in waves
1. Select an initial group of respondents
2. Those respondents provide referrals
3. The referrals provide referrals
-non-probability
Primary scales of measurements
-nominal= gender (male or female)
- ordinal = placement (1st or second or third etc)
- interval = temperature (no true zero)
- ratio= length
Transitivity of preference
Given three objects A, B and C. If a consumer prefers a to b and b to c then according to the concept they should also prefer a to c. Therefore, A>B>C.
Unbalanced scale
It may be required when you expect responses to be skewed
Three elements used to evaluate a scale
Reliability, validity, generalisability
Paired comparison
Which chocolate do you prefer? X or Y
Rank Order
Rank the following chocolate bars from 1 being your least favourite to 4 being your favourite.
Constant Sum
Below are 5 chocolate brands. Allocate 100 points among the different brands. The number of points you allocate determines the importance of the brand.
Likert itemised rating scales
Requires a statement with an uneven number of categories
Semantic differential itemised rating scale
requires suitable bipolar labels and 7 points
Stapel itemised rating scale
it should be vertical with a 10 point scale