research planning and design lecture 2 Flashcards
main goals and features of experiments
- establish a causual relationship between variables
- measure the effect of one variable on another via manipulation
- must involve the active intervention of the researcher
- controlled environment
- control of extraneous variables
systematic manipulation
deliberate and controlled alteration of one or more variables in an experiment to observe its effect on other variables.
- changing the level/ condition of independent variable systematically.
single factor two level design
= one independent variable, two levels
single factor multi level design
= one independent variable, more than two levels
within subject design/ repeated measures design
= all participants are exposed to all levels/ conditions of the IV at different times
between subject design/ independent design
= different participants are exposed to different levels of the IV
extraneous variables
= any other variable apart from the IV and DV that could affect your outcome
- if it is not controlled can become a confounding variable
- example, age, time of testing
confounding variables
= a extraneous variable that varies systematically across levels of IV
Random assignment + block randomisation
way to control for confounding variables
1.) each participant has the same chance of being assigned to each condition, assignment of each participant is independent of all the others.
2.all the conditions occur once in the sequence before any of them is repeated.
benefits of within-subject design
- maximum control of extraneous variables
- more powerful
- smaller sample size needed
disadvantages of within subject design
- not practical with all designs
- carryover effects (repeated practice, fatigue, order effects) -> solution to this is counterbalancing
benefits and cost of between subject design
- no carryover effects
- independent scores
- less time per participant
- practical for designs where its not possible to have participants in more than one condition
- needs a larger sample size
counterbalancing
complete counterbalancing = equal number of participants complete each possible order of conditions -> might not be possible with large number of conditions (lots of different orders)
random counterbalancing = randomly assorted, not paying attention to how many people in each group
latin square = each condition appears exactly once in each position, means that each condition has an equal chance of being experienced in every position -> allows for fewer participants
randomised control trial
= experiments done in the field to investigate the effectiveness of a treatment/ intervention
- only difference we should be able to see is the outcome of the variable being studied
placebo effect solution > double/ single blind
internal validity
- ## tests if we can we confidently infer a causal relationship between variables without the influence of other factors or variables.