Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
100 Cards in this Set
- Front
- Back
|
Validity |
Is an observed effect genuine, and does it measure what it's meant to? |
|
|
Improving validity |
Control investigator effects, confounding variables and demand characteristics so that only the participant's behaviour is being recorded |
|
|
Internal Validity |
The study's ability to test the hypothesis it was meant to test, aka the extent an IV has affected the DV |
|
|
External validity |
The study's ability to be applied outside of the study |
|
|
Temporal validity |
The extent to which findings can be applied to different dates in history |
|
|
What things should be considered in validity? |
Demand characteristics Experimenter bias Participant variables |
|
|
What should be considered in assessing external validity? |
Ecological validity Population validity Temporal/historical validity |
|
|
Ecological validity |
The study's ability to be applied to real life scenarios |
|
|
Face validity |
A deft assessment of a test's validity to see if it assesses what it intends to assess |
|
|
Concurrent validity |
Does a test gain similar results in a different test, and test the correlation |
|
|
Content validity |
Experts in the study's field check the methodology to see if it measures the desired behaviour |
|
|
5 things researcher must change to improve validity |
Investigator effects
Demand characteristics
Confounding variables
Social desirability
Poorly operationalised behaviour categories |
|
|
How can you improve ecological validity? |
Give participants a test with high mundane realism, and if possible make it a field or a natural experiment |
|
|
How can you improve population validity? |
Use a large sample, use a sampling technique that is likely to produce a representative sample |
|
|
How can you reduce demand characteristics? |
Single blind |
|
|
Single Blind |
Participants do not know which condition they're in meaning they cannot alter their responses to suit or foil the researcher |
|
|
How can you reduce demand characteristics and researcher bias? |
Double blind |
|
|
Double blind |
Neither the participant or researcher know what each condition represents. Eg. Researcher splits groups into X or Y but does not know which group represents what |
|
|
Reliability |
Will results be the same if the procedure is repeated? |
|
|
Internal reliability |
Consistency within a test a participant takes, most common in personality tests as personality must be consistent throughout Consistency |
|
|
External reliability |
The ability to be able to produce the same results every time the task is carried out Same results |
|
|
3 ways to improve reliability |
Standardisation Take more than one measure Pilot studies |
|
|
Standardisation(reliability) |
Procedures must be the same each time otherwise participant's performance cannot be compared |
|
|
Take more than one measure(reliability) |
Take more than one measurement from a participant and create an average of their results, reducing the frequency of anomalous scores |
|
|
Pilot studies(reliability) |
Finds issues with research design such as inaccurate instructions |
|
|
Improving reliability in observations |
Behavioural categories Pilot studies Standardisation |
|
|
Behavioural categories(observations) |
Observational categories need to be operationalised to ensure researchers don't misinterpret behaviour |
|
|
Pilot studies(observations) |
Discover problems such as poorly defined behavioural categories or inadequate training |
|
|
Standardisation(observations) |
Training to ensure the criteria is clear |
|
|
Improving reliability in self report studies |
Reduce ambiguity Pilot studies Standardisation |
|
|
Reduce Ambiguity(self report studies) |
Make questions understandable, and ensure they have a certain definition and purpose |
|
|
Pilot studies(self report studies) |
Ensures that the questions makes sense to the average participant |
|
|
Standardisation(self report studies) |
Reduces investigator effects by creating a standard by which questions must align by |
|
|
Inter-interview reliability |
Get multiple interviewers and compare their results |
|
|
Split half method |
Results on one half of a test and results on another half are compared, common in questionnaires as results are easy to compare. This assesses a test's consistency |
|
|
Inter-observer reliability |
The extent to which observers tasked with observing behaviour agree on the observations they record. This allows the observations of multiple observers to be correlated, and allows for reliability to be assessed. If there is a correlation of 0.8, the data will have high inter-observer reliability |
|
|
Test retest |
Test is used multiple times with the same group of participants, with a short gap to ensure participants don't remember questions.(a week) The previous scores are compared with the new ones and a correlation is created. 0.8 correlation=good |
|
|
What reliability does split half method assess? |
Internal reliability as it measures consistency by comparing results on one test |
|
|
What reliability does test retest assess? |
External reliability as it assesses whether the test can be repeated and gain the same results |
|
|
Content analysis |
Turns qualitative data into quantitive data |
|
|
Case study |
A detailed study of an individual or a group allowing for individualistic data that gives researchers a realistic perspective of a group or individual. This may be done with methods like interviews or observations |
|
|
Case study example |
London Riots- Psychologists assessed behaviour in the London Riots to figure what caused the riots |
|
|
Coding |
Categorising data into specific categories to allow researchers to conduct content analysis easier |
|
|
Thematic analysis |
Records qualitative data |
|
|
Nominal evaluation |
(+)Generates a lot of data quickly (-)Data is crude and does not allow for analysis |
|
|
Ordinal data |
(+)Participants can respond differently to questions, allowing for more sensitive data (-)Based on subjective opinion, thus lacks precision |
|
|
Interval |
(+)More precise than nominal or ordinal because it is based on scientific measurements |
|
|
Is the study testing difference or association? |
An experiment is a test of difference, whereas a correlation is a test of association |
|
|
What type of data is it? |
Ordinal Nominal Interval |
|
|
Sign test |
1.Repeated measure designs 2.Nominal data 3.Looking for a difference |
|
|
Mann-Whitney |
1. Independent groups 2. Ordinal data 3. Looking for a difference |
|
|
Wilcoxin |
1.Repeated measures 2.Ordinal data 3.Looking for a difference |
|
|
Unrelated t-test |
1. Independent groups 2. Interval data 3. Looking for a difference |
|
|
Related t-test |
1. Repeated measures 2. Interval data 3. Looking for a difference |
|
|
Spearman's Rho |
1. Ordinal data 2. Looking for a correlation |
|
|
Pearson's R |
1. Interval data 2. Looking for a correlation |
|
|
Chi Squared |
1. Independent groups 2. Nominal data 3. Looking for a correlation or difference |
|
|
Paradigm |
A shared set of assumptions that that distinguishes scientific disciples from one another |
|
|
Falsifiability |
Can another researcher prove the hypothesis wrong(Is it operationalised)? |
|
|
Objectivity |
Is the theory free of bias, and is it based on observable phenomena? |
|
|
Replicability |
Can the method be repeated by the research or another researcher and gain the same results |
|
|
Empiricism |
Is the information for the study gathered through observable methods? |
|
|
Theory construction |
Inductive and deductive methods used to create a theory |
|
|
Hypothesis testing |
A hypothesis needs support by other studies to be efficient, thus if it doesn't it must be altered |
|
|
Correlation coefficient
|
Number between -1 and +1 that demonstrates the strength of a relationship between variables |
|
|
Positive Correlation |
Perfect correlation=+1 +0.5 and above=strong +0.4 and below=weak |
|
|
Negative Correlation |
Perfect correlation=-1 -0.5 and above=strong -0.4 and below=weak |
|
|
Are the results significant?
|
Did the IV affect the DV? |
|
|
Results are significant |
Accept experimental hypothesis, reject null hypothesis |
|
|
Results are not significant |
Accept null hypothesis, reject experimental hypothesis |
|
|
One-tailed hypothesis |
Hypothesis predicts the expected direction of the results |
|
|
Two tailed hypothesis |
Hypothesis predicts the expected difference of the results |
|
|
Example of one tailed hypothesis |
As IQ goes up, participants happiness will go down |
|
|
Example of two tailed hypothesis |
There will be a difference in happiness score depending on IQ |
|
|
Why is 0.05/5% level of significance used? |
0.05/5% is used as it strikes a balance between the risk of Type I/Type II errors |
|
|
Type I error |
Results look significant but they're not meaning that the expected hypothesis was accepted when the null hypothesis should've been accepted |
|
|
Type II error |
Results don't look significant, but they are |
|
|
Why do the errors occur? |
Type I occurs cause probability is too high and the test was too easy, and type II happens cause probability was too low and the test was too hard |
|
|
How to check for errors |
To find a Type I error use test retest and set the probability lower, and for a type II error use test retest and set probability higher |
|
|
Abstract |
A short summary of the whole report that outlines one sentence for all aspects of the scientific report |
|
|
Aspects of scientific report |
Aim Hypothesis Sample Procedure Results Conclusion |
|
|
Study design |
Hypothesis Sample Procedure IV+DV Research design Ethics Conclusion Method Covert or overt |
|
|
Self report study aspects |
Open or closed questions |
|
|
Interview study aspects |
Structured, unstructured or semi structured |
|
|
Pre-science |
No paradigm, one must be created
|
|
|
Scientific revolution |
Evidence against old paradigm creates paradigm shift |
|
|
Kuhn(1970) |
Psychology is a pre-science as it doesn't hace a pradigm |
|
|
Things a theory requires |
Falsifiability Empiricism Objective Replicability Generalisation |
|
|
Falsifiability |
Can a theory be falsified |
|
|
Empiricism |
Does the theory use observable data? |
|
|
Objective |
Data that has no bias |
|
|
Replicability |
Standardised procedure |
|
|
Generalisation |
Can results be applied to general population |
|
|
Significance |
A statistical term that refers to how certain we are a correlation or difference exists |
|
|
Nominal data |
The researcher counts how many people are in each category to assess whether a change in participants has occurred, participants can only be in one category and the data is presented in a bar chart |
|
|
Ordinal data |
Each participants has a rough numerical score without measurble units, and is data that is put in order on a scale |
|
|
Interval data |
Participants are placed on a scale with recognised equal units |
|
|
Two tailed test |
A non-directional test that states there will be a difference whether it be positive or negative |
|
|
One tailed test |
A directional test that states there will be a positive or negative difference |
|
|
Finding the critical value |
Based on level of significance and whether it's a one tailed or two tailed test, look for the table row that has the number of participants in the study, or the degrees of freedom or the number of scores changed in a sign test |
