BDO 373
SU4: Research Design
Research designs
Steps that take you to answer your research Q
Should tell you:
▪ If you will observe 1/ more groups
▪ If there are more groups , how will groups allocation take place
▪ Where the observations will take place
It constitutes the blueprint for the collection, measurement, and analysis of data
Different approaches to research
1. Quantitative
o Quantity- elementally an amount of something
o Counts and measures of things (Berg, 1998)
o Collect numerical data and test hypothesis (Jackson, 2006, p. 8)
o Large samples
2. Qualitative
o Emphasise the approaches focus on processes and meanings that are rigorously
examined but not measured in terms of quantity, amount or frequency
o Data provide depth and detail through direct quotation and careful description of
situations, events, interactions and observed behaviours (Labuschagne, 2003)
o Small sample
o Interviews
Important terminology: Laboratory studies and field research
Field (within natural context)
Location
Laboratory (simulated enviromnent)
Laboratory- any controlled environment where the researcher can manipulate the variables
1 of the most famous psychology experiments in history is Milgram’s obedience study
Design by time period
Cross sectional (at one collects data from
point in time) groups
Time taken to collect data
at least 2
Longitudinal (over a
mesuarements at diff
long period of time)
times
©Megan Meiring
,Different types of quantitative designs !NB!
1. Non-experimental Experimental
X- a natural event X- a manipulated intervention
2. Pre-experimental 3. Quasi-experimental 4. True experimental
Cross-sectional survey One-group post-test Randomised one-group Randomised pre-test
(data collection design) only post-test only post-test control group
(classic experiment)
Randomised cross- Multigroup post-test Comparison group Randomised Solomon
sectional survey only post-test only four-group
Replicated randomised Longitudinal case Comparison group pre- Randomised post-test
cross-sectional survey study test post-test only control group
Retrospective case One-group pre-test Interrupted timeseries
study post-test and single system
Group comparison Ex post facto
Correlation
Non-experimental designs (Describe, explore, explain) :
©Megan Meiring
, Pre-experimental
➢ One-group post-test only- event that 1 group was exposed to (IV), and 1 measurement after the event (DV)
➢ Multigroup post-test only- Same as 1gpto design but uses multiple groups to compare in terms of the
comparison
➢ Longitudinal case study- Same as 1gpto design, but with many measurements over a longer period (DV)
➢ One-group pre-test post-test- Same as the 1gpto design with an added pretest for a baseline
measurement
➢ Ex post facto- Study naturally occurring events based on secondary data
True experimental
Quasi-experimental
3 types of Quantitative variables:
1. Independent variable- effects dependent variable
2. Dependent variable- outcome of independent variable
3. Confounding variable- other variables hidden that influence relationship btwn variables
(situational variables, participant variables, experimenter effects, and demand characteristics)
Internal vs external validity (IV vs EV)
Internal External
Degree to which changes in the dependent Degree to which the results can be generalised
variable are indeed due to the independent to the whole population
variable rather than something else
Experimental designs- stronger IV and weaker Non-experimental designs- stronger EV and
EV because so controlled not representation of weaker IV because little control & manipulation.
real world. Little control over confounding Likely reflect the relationship that exist in the
variables real world. Takes place in a natural environment.
©Megan Meiring
SU4: Research Design
Research designs
Steps that take you to answer your research Q
Should tell you:
▪ If you will observe 1/ more groups
▪ If there are more groups , how will groups allocation take place
▪ Where the observations will take place
It constitutes the blueprint for the collection, measurement, and analysis of data
Different approaches to research
1. Quantitative
o Quantity- elementally an amount of something
o Counts and measures of things (Berg, 1998)
o Collect numerical data and test hypothesis (Jackson, 2006, p. 8)
o Large samples
2. Qualitative
o Emphasise the approaches focus on processes and meanings that are rigorously
examined but not measured in terms of quantity, amount or frequency
o Data provide depth and detail through direct quotation and careful description of
situations, events, interactions and observed behaviours (Labuschagne, 2003)
o Small sample
o Interviews
Important terminology: Laboratory studies and field research
Field (within natural context)
Location
Laboratory (simulated enviromnent)
Laboratory- any controlled environment where the researcher can manipulate the variables
1 of the most famous psychology experiments in history is Milgram’s obedience study
Design by time period
Cross sectional (at one collects data from
point in time) groups
Time taken to collect data
at least 2
Longitudinal (over a
mesuarements at diff
long period of time)
times
©Megan Meiring
,Different types of quantitative designs !NB!
1. Non-experimental Experimental
X- a natural event X- a manipulated intervention
2. Pre-experimental 3. Quasi-experimental 4. True experimental
Cross-sectional survey One-group post-test Randomised one-group Randomised pre-test
(data collection design) only post-test only post-test control group
(classic experiment)
Randomised cross- Multigroup post-test Comparison group Randomised Solomon
sectional survey only post-test only four-group
Replicated randomised Longitudinal case Comparison group pre- Randomised post-test
cross-sectional survey study test post-test only control group
Retrospective case One-group pre-test Interrupted timeseries
study post-test and single system
Group comparison Ex post facto
Correlation
Non-experimental designs (Describe, explore, explain) :
©Megan Meiring
, Pre-experimental
➢ One-group post-test only- event that 1 group was exposed to (IV), and 1 measurement after the event (DV)
➢ Multigroup post-test only- Same as 1gpto design but uses multiple groups to compare in terms of the
comparison
➢ Longitudinal case study- Same as 1gpto design, but with many measurements over a longer period (DV)
➢ One-group pre-test post-test- Same as the 1gpto design with an added pretest for a baseline
measurement
➢ Ex post facto- Study naturally occurring events based on secondary data
True experimental
Quasi-experimental
3 types of Quantitative variables:
1. Independent variable- effects dependent variable
2. Dependent variable- outcome of independent variable
3. Confounding variable- other variables hidden that influence relationship btwn variables
(situational variables, participant variables, experimenter effects, and demand characteristics)
Internal vs external validity (IV vs EV)
Internal External
Degree to which changes in the dependent Degree to which the results can be generalised
variable are indeed due to the independent to the whole population
variable rather than something else
Experimental designs- stronger IV and weaker Non-experimental designs- stronger EV and
EV because so controlled not representation of weaker IV because little control & manipulation.
real world. Little control over confounding Likely reflect the relationship that exist in the
variables real world. Takes place in a natural environment.
©Megan Meiring
