Aims and hypotheses
Aim: Outlines the general area the researcher is interested in investigating.
Hypothesis: A precise prediction about the expected findings of a study. It is not merely an idea.
For scientists, it is not precise enough as it doesn't specify exactly what is to be measured and how. To do this the researcher will operationalise the hypothesis, they will manipulate variables to be identified, measured and used in the experiment.
Null hypothesis: There will be no difference between the conditions
Experimental hypothesis: There will be a difference between conditions. Termed experimental if study based on an experiment. Alternative such as a questionnaire.
One-tailed hypothesis: The direction of the influence or relationship is predicted. (less, more, faster or slower).
Two-tailed hypothesis: does not specify a direction difference. It simply states that one factor affects another.
Variable: Any object, thing or event that changes or varies.
IV: Variable that the experimentor manipulates in some way.
DV: Variable which changes as a result of the manipulation of the IV.
Extraneous variables: All the other variables that could cause the DV to change that are controlled. (Control variables).
Confounding variable: An extraneous variable that has been identified after the experiment is conducted that has not been controlled for.
Random errors: Any uncontrolled or inadequately controlled variable which could effect the results in an unpredictable way.
Constant errors: Something other than the IV which consistantly affects the DV.
Extraneous variables = subject / situation variables
Subject variables: Are individual differences of a participant (intelligence / gender)
Situation variables: Any differences that exist in an experimental situation. (rooms).
Population: Any study has a population to which the results of the study are intended to relate.
Sample: A selection of participants taken from the population that is studied so that the researcher can make generalisations about the population as a whole.
Representative sample: A sample that is as similar as possible as the target population.
Sampling error: occurs if the sample of participants that have been selected contains an over-representation of a selection of the target population.
Random sampling: Every member of the identified target population has an equal chance of being selected.
- It is unbiased
- It may not be representative
- It is extremely time consuming to identify whole population.
Opportunity sampling: Selects those participants that are willing and available at the time of the study.
- Practical and time efficient method.
- Non-representative sampling technique.
Quota sampling: Researcher identifies a selection of the population (e.g. certain age). Selects participants who are willing and available
- Quick and efficient.
- Not all participants in population will be selected, not representative.
Systematic sampling: Random sampling method, every nth person is selected from population.
- It is random, helps reduce sampling bias.
- Not representative.
Stratified sampling: Randomly selecting participants from each stratum.
- Likely to be representative.
- extremely time consuming.
Lab experiments enable researchers to: control variables, enables replicability, experiments are seen to be objective. Weakness:ecological validity
Field experiments: Using participants in their normal surroundings.
- Very high level of ecological validity, participants are usually unaware they are being studied. Demand characteristics are reduced.
- Researcher doesn't always have complete control. It is very difficult to replicate.
Natural / Quasi experiments: The researcher has no control over participants.
- High ecological validity, demand characteristics are reduced.
- Lack of control over IV, difficult to replicate.
Related / Repeated measures design: Participants take part in both or all conditions of the experiment. Their performance in the different conditions is cmpared and analysed.
- Practical convenience: Don't need to recruit as many participants. Absence of bias.
- Order effects, Practice effects and demand characteristics.
Unrelated measures design: Different participants are tested in each condition of the experiment and participants are randomly allocated to conditions.
- Eliminates order effects.
- reduces demand characteristics.
- Individual differenes.
- Cost and convenience - needs twice as many people.
Matched Pairs design: Each participant matched with another on characteristics. One of the pair allocated one condition, and their associated pair allocated to other condition.
- Eliminates order effects.
- Reduces demand characteristics.
- Time consuming to match pairs accurately.
- If one person decides to withdraw, researcher loses two people.
A pilot study is a small scale study carried out with a few participants to highlight any possible problems. It enables the researcher to test their design, procedure and materials used.
Consent: All perticipants, where possible, should be informed of the full details before the researcher seeks their consent for the study.
Right to withdraw: They have the right to withdraw at any time. A participant may withdraw their consent retrospectively and require their data to be destroyed.
Deception: Withholding info or misleading participants is unacceptable. Deception may be unavoidable in the study of psychological processes. In these, participants should be informed of the full details of the study at the earliest opportunity.
Confidentiality and anonymity: Results treated in the strictest confidence. Participants identity should remain anonymous..
Researchers must protect participants from physical or mental harm. They should be asked about any pr-existing medical issue.
Brief and De-brief
Brief: Consent, their right to withdraw, confidentiality of data.
De-brief: Reassure participants of their right to withdraw consent, reassured of confidentiality. If a single-blind technique is used, participants must be fully informed of the details of the investigation.
Single-blind technique: Participant is naive. Not fully informed of details of investigation. Reduces effect of demand characteristics.
Double-blind technique: Both the researcher and participant are naive. The researcher merely collects the data. Avoids introducing demand characteristics.
Levels of measurement
refers to differences in precision using quantitative studies.
Nominal level: Each piece of data is put into a category. Measurement is basically a head count. Gives very little info.
Interval level: more precise, measures time, temperature and length.
Ratio level: Similar to interval level but scale has a genuine zero or starting point.
Ordinal level: Researcher has collected data in form of a score. If two scores are equally represented, the distribution of scores is bi-modal. If there are 3 'modes', it's tri-modal.
Mode - advantage - Suitable when data not affected by outliers.
- Disadvantage - Not useful if there are many different modes.
Median - advantage - Suitable when data collected is at an ordinal level and also when data produces outliers.
- Disadvantage - Not useful if sample is very small
Mean - advantage - Suitable measure of central tendancy, when data is at interval/ratio level.
- Disadvantage - If outlier occurs it will skew the mean, cannot be used on categorical data.
Range - Advantage - Quick and easy to calculate.
- Disadvantage - Can be distorted by anomolous results, with a large range we dont know how the scores are dispersed.
Standard deviation - Advantage - Most sensitive measure of dispersion, considers all data, most appropriate with interval data.
- Disadvantage - an anomolous score will distort SD.
Tabular display: A table of results.
Graphical display: a graph of results.
Bar charts: Display frequencies of discrete data: drawn next to each other if continuous, apart if not continuous.
Histograms: Displays frequencies of continuous data, no gaps.
Polygon: mid-points at the top of each bar are joined. Useful to show results of 2 or more conditions at once.