A hypothesis is a testable statement about the expected outcome.
A Null hypothesis is a hypothesis which predicts no effect of the IV on the DV.
A operationalised hypothesis is where the exact nature and means of measuring variables.
A One tailed (directional) hypothesis is a precise hypothesis which specifically states the direction of the results.
A Two tailed (non-directional) hypothesis is a less precise hypothesis whcih does not predict the outcome or direction of the results..
There are 2 types of variables that are involved in designing an experiment:
- Independent variable (IV), a varibale that is manipulated by the researcher.
- Dependent variable (DV), a variable that is affected by the change in the dependent variable.
When designing an experiment you need to control extraneous varibles which are any variables other than the independent variable which might affect the Dependent variable.
Lab experiments - where the IV is directly manipulated, and all extraneous variables are controlled.
+ High levels of control, replicability, can conlude cause and effect.
- Lack ecological validity, high chance of participant and investiagtor effects, lack mundame realism,
Field experiments - The researcher controls the independent variable but the experimenter cannot control other extraneous variables.
+ Can conclude cause and effect, high levels of ecological validity, reduction in participant effects.
-Less control over extraneous variables, often more time consuming,
Natural experiments - the researcher does not control extraneous variables or manipulate the independent variables.
+ Useful when it is unethical or impossible to manipulate the IV, high ecological validity.
- Low internal validity, low extraneous varianles.
Correlation - Technique for analysing data by measuring strength of relationship between 2 variables. They can show a positive correlation; a negative correlation; or no correlation.
+Can establish strength of relationship between 2 variables.
- Cannot establish cause and effect, can only measure linear relationships.
Questionnaires - a set of questions used to collect data.
+ can use large sample, can collect large amounts of data, time efficient, reduces investiagtor effects, replicable.
- Social desirability, low postal response rate,, closed questions can limit depht, open questions difficult to analyse.
Observations - data is collected by observing behaviour. There are 2 types - naturalistic and controlled.
+ N - high ecological validity, participants behave more natural. C - higher levels of control over EVs
- N - no control over EVs, C- may not act naturally
Interviews - researcher asks participants questions direcly face to face. These can be stuctured or unstructure.
+ S - Less risk of investigator effects, able to clarify ambgiouty, US can follow up, expand answers, more informal.
- S - cannot follow up answers, formal - may inhibit answers. US interviewer effects, time consuming, social desirability.
Independent Measures - participants are randomly allocated to wither one or other of conditions.
+ No order effects, reduce demand characteristics,
- Little control over participant variable, more particiants required.
Repeated measures - the same participants are use in both conditions
+ No participant variable, fewer participants needed
- Order effects, demand characteristics
Matched pairs - participants are matched as closely as possible with another participant and then pairs are randomly allocated to wither one or the other conditions.
+No order effects, good attempt at controlling participant variables
- difficult to match participants exactly, more participants required.
The BPS code of ethics:
- Consent - for someone to participate in research they have to provide consent.
- Deception - information must not be witheld from participants
- Debreifing - after an investigation participants have to be debriefed to inform them.
- Withdrawal from investigation - participants have to be aware of their right to withdraw.
- Confidentiality - Participants have the right to confidentiality
- Protection of participants - psychologists have responsibility for protecting participants from physical and pschological harm.
Methods for dealing with ethical issues:
Deception can be dealth with through debriefing, and retrospective informed consent
Informed consent can be dealt with by giving prior general consent, presumptive consent, and for children gaining consent of parents or those in loco parentis.
Protection from harm can be dealth with by reminding participants of their right to withdraw, terminating any research which appears to cause distress, debriefing, and offering advice/support.
Reliability and Validity
Reliability is where 2 or more measurements of the same psychological event will be consistent with each other. We can test for reliability by repeating the experiment and using the same observation and behaviour checklist,
Validity is where we measure what we claim to be measuring.
Internal validity is where the outcome is the result of the variable that are manipulate in the study.
External validity is whether the findings can be generalised to other settings. (e.g. population validity, cultural validity etc.)
You can test for validity by using face validity which is a simple eyeball test where you check that it looks as if it is measuring what it is supposed to be measuring. You can also use concurrent validity which involved comparing the results yeilded by a new test with those from an older test known to have good validity.
There are different types of sampling which are:
Random - A sample whcih every member of the target population has an equal chance of being selected. This sample has high population validity.
Oppourtunity - A sample that consists of thoe people available to the researcher. This sample has a high chance of bias and little population validity.
Stratisfied - the participants are classified into categories and then chooses a sample which consists of participants from each category.
Self selecting (volunteer) - a sample that consists of people who volunteer to take part in the study. This sample has a high chance of bias and low population validity.
Qualitative data is data that is focused on numerical data. The advantages of this is that it provides rich, detailed information but it can provide too much information to analyse quickly.
Quantitative data is data that is focused on numerical data. The advantages were that the data can be analysed quickly, however, it provides limited information.
Measures of central tendancy are the mean, mode and median.
+ The mean is the most sensitive measure taking all scores into account. The median is unaffected by extreme scores.
- The mean can be distorted by extreme scores, the median only takes one or two scores into account.and can be affected by a change in one score,
Measures of dispersion is the spread of data around a central value.
+ The range is easy to calculate, the standard deviation takes all scores into account and is a sensitive measure of dispersion.
-The range can be distorted be extreme scores, and SD can be difficult to calculate.
There are different types of data:
Nominal data: Data that is placed in categories, e.g. number of people with blue eyes.
Ordinal data: data that can be put into rank order, e.g. exam marks.
Interval data: data that is measure in terms of equal intervals e.g. minutes.
Each graph can be used for different types of data:
- Bar - used for nominal data and non-continuous data.
- Histograms - used for ordinal/interval data which is continous.
- Frequency polygrams- two sets of continous data
- scatergrams - relationships between 2 variables.