L8 - Reliability, Validity & Control Of Extraneous Variables Flashcards
Reliability definition
Reliability refers to the consistency of a research study especially when a study is repeated again and the same results are gained on both occasions
Two types of reliability
- External Reliability
- Internal Reliability
External Reliability
- This is whether a test and the results gained are consistent over time.
- The test-retest method can be used to assess external validity.
- A research study is conducted once, and then it is conducted again in the future.
- If the results gained from the both tests are similar then the study can be said to be reliable
Internal reliability
- This is whether a test and the results gained are consistent within itself.
- The split-half technique assesses the internal reliability of questionnaires.
- The questionnaire is split in half and if participants score similarly on both halves of the questionnaire then the questions are measuring the same factors and the questionnaire has internal reliability.
Validity
Validity means that a study is measuring what it intends to measure when referring to the aim of the study
External Validity: (Ecological validity)
This is the extent to which the findings of a study can be generalised to other settings
Participant validity
The results from the participants used in the study can be generalised to the target population
Temporal validity
The results from the study can be generalised to people in today’s contemporary society
Internal Validity
- This is when the outcome of the study is a direct result of the manipulation of the independent variable (IV) upon the dependent variable (DV) and has not been affected by extraneous variables (EV).
- In order to ensure that a study has internal validity, extraneous variables must be well controlled
3 EV categories
- participant variables
- situational variables
- experimenter variables
Participant variables
- These are characteristics of the participants which may affect the DV (e.g. intelligence, age, gender, personality etc.).
- Choosing an appropriate experimental design can help to try and overcome these type of extraneous variables.
- Matched pairs and repeated measures design can help to avoid participant variables.
- However, repeated measures can lead to order effects, so counterbalancing should be used to avoid this.
- Random allocation of participants to conditions (e.g. by drawing names out of a hat) should also ensure that groups are not biased.
Situational variables
- These are factors in the environment where the experiment is conducted that could affect the dependent variable (e.g. temperature, time of day, lighting, noise etc.).
- A way to resolve this issue is to use standardisation (i.e. making sure that all the conditions, materials, and instructions are the same for all participants).
Experimenter variables
- These are factors to do with the experimenter which can affect the dependent variable, for example personality, appearance, and conduct of the experimenter.
- Standardised instructions should ensure that the experimenter acts in a similar way with all participants and follows a script and speaks to everyone in the same manner and tone.
Investigator effects
- Investigators may inadvertently influence the results of their research.
- Certain physical characteristics of the investigator, such as age, gender and ethnicity can influence the behaviour of participants which therefore affects the data that is collected from the research.
- If investigators know the hypothesis they may also inadvertently be biased in their interpretation of the results.
- Observer bias and interviewer effects are a type of investigator effects
Investigator effects can be overcome by the….
……double blind technique
This is when neither the participants nor the investigator knows the aim of the study and hopefully this will mean the data collected will be more valid.