Flashcards in Experiments Deck (35)
What is the purpose of experiments?
To examine and test causal relationships or ceteris paribus relationships
What are the rules of causality?
- To establish that variable X causes Y, all there conditions must be met:
- 1. Both X (cause) and Y (effect) should covary
- 2. X should precede Y. There must be a time sequence in which the two occur.
- 3. No other factor should possibly cause the change in the outcome Y. Other variables that might covary with Y must be controlled for
What kinds of variables are there in experiments?
* Confounding variable
* Extraneous variable
What are independent variables?
What are dependent variables?
What are confounding variables?
- When there is an alternative explanation beyond the experiment variables for any observed difference in the dependent variable
- Effects the DV
- DOES effect
What are extraneous variables/
- Variables that naturally exist in the environment that may have some systematic effect on the V
- Could affect the DV but has been controlled for so it doesn’t
- MAY effect
What are the two essential aspects of experiments?
- 1. Manipulation of the independent variables
- The changes in X cannot occur naturally, they have to be deliberalty created, manipulated or arranged by the experimenter
- 2. Control of the extraneous variables
- The method can only be used when subjects and conditions to be studies can be practically and ethically manipulated by the researcher
What is the method to conducting experiments?
* Identify the causal factors
* Plan and Administer the manipulations
* Plan and apply controls
* Observe and Measure the DV
What are treatments in experimentation?
- Manipulations of the IV are also known as treatments which are administered to the experimental groups whereas control groups do not receive these treatments
What are the ways in which the IV can be manipulated?
* Presence vs Absence technique
* Amount technique
* Type technique
How are controls planned and applied in the experimentation method?
* Proper sampling techniques
* Standardisation of conditions
* Controlling EVs and CEs
How can EVs and CEs be controlled for?
- Confounding effects and extraneous variables can be minimised through good sampling
* Matching groups
What is randomisation in experimentation?
- For controlling for the EVs and CEs
- Involves randomly assigning subjects to experimental and control groups
- Doing so distributes the effects of counfounding variables across groups equally
- Any errors caused by say age or gender are now more or less similar across groups
What are matching groups in experimentation?
- For controlling for the EVs and CEs
- More precise strategy
- On the basis of confounding variables
- e.g. the effects of gender and age can be distribute across groups
- The risk is that we may fail to match some critical factors across groups
What is the treatment effect?
In observing and measuring the DV, the difference between pre-testing and post-testing
What are the types of experimental design?
* Two randomised groups design (classic design)
* Two matched groups design
* Two within-subjects groups design
* Multiple Groups designs
* Facotrial designs
What is the two randomised groups experimental design (classic design)?
Random assignment of participants are made to the experimental and control groups
What is the two matched groups experimental design?
Participants in the experimental and control groups are matched with one another in terms of potential EVs
What is the two within-subjects group experimentals design?
Participants in the experimental and ctonrol groups are the same people, measures before and after the treatment - the ultimate in matching
What is the multiple groups experimental design?
IV has more than 2 conditions; participants can be randomly assigned, matched or repeated across the conditions
What is the factorial experimental design?
More than one IV is investigated, allows experiments to test a more complex causal hypothesis as DVs can have multiple causes; allows exploration of interactions effects of IVs too
What are the two types of experiments?
- Contrived but standardised environment
- When manipulations and controls are introduced to establish cause-effect relationship in an artificial setting
- Conducted in a natural setting, often for a longer period of time
- Not possible to control for all extraneous variables in the field because it may not be possible to assign subjects to groups or match them
How can the internal validity of an experiment be measured?
- Did the experiment measure what is set out to do
- The ability of an experiment to establish whether the experimental treatment (IV) was the sole cause of changes in a DV
- Experiment is invalid if there are rival hypotheses or alternative explanations for the results
How can the external validity of an experiment be measured?
- Can the findings be applied to other situations outside the experimental situation
- The extent to which the results found are generalisable to actual field or organisational settings
What is true of the validity of the different types of experiments?
- Lab experiments have high internal validity but low external validity compared to field experiments
- To ensure both types, some researches first test causal relationship in a laboratory setting, and then examine whether such relationship are generalisable to the field setting
What threats to experimental validity are there?
* History effects
* Maturation Effects
* Testing Effect
* Instrumentation Effect
* Selection Bias Effects
* Demand Effects
* Hawthorne Effect
* Morality or Attrition
What is the history effect threat to experimental validity?
- Specific events that occur while the experiment is in progress that could impact on the DV, but are beyond the control of the experiementer
- e.g. a natural disaster or an economic downturn occurs; new laws are passed; the company downsizes
What is the maturation effect threat to experimental validity?
- The passage of time over the course of the experiment can influence the DV
- Subjects can get tired, gain experience, get bored, grow older and these processes can affect the DV