Study Design

Construct

Construct

A representation of a factor that is not directly observable, done via operationalisation. For example, this could be done by IQ for intelligence.

Operationalisation

The process of converting a factor that isn’t directly observable into a construct which can be measured or observed, for example intelligence as IQ.

Reliability

A measure of how well studies can be replicated.

Internal

A type of reliability that assesses how well all measures within a single test assess the same concept, for example personality.

Inter rater

A type of reliability that assesses whether multiple observers who witness the same thing reach the same conclusions, for example whether in an anger study both observers recognise the participant is angry.

Test retest

A type of reliability that assesses whether the results of the measure are consistent for the same individual when used multiple times/in the same conditions. For example, does an individual score lower on an IQ test in the morning?

Validity

A measure of whether measures are measuring the right thing.

Content

A type of validity where the measure is checked to ensure it measures all aspects of the construct it intends to measure. For example, are all elements of personality assessed in a personality measure?

Face

A type of validity where the measure is checked to ensure participants see the test is valid, as this may effect their level of participation. For example, this is low if a memory questionnaire asks questions about exercise.

Criterion

A type of validity where the measure is checked against other related measures and the results are compared. For example, whether another personality test had a similar number of introverts.

Construct

A type of validity where the measure is checked for how well it measures the concept it was designed for. For example, whether a personality test accurately measure personality.

Correlational

Studies which investigate the relationship between two or more variables. Can test hypotheses, but cannot make causal statements due to directionality issues, a potential third variable and the potential for a coincidence.

Linear

Term for correlations that are either positive or negative.

Experimental

Studies where the IV is manipulated and the effect on the DV is measured. These aim to minimise noise and control other potential variables.

Noise

Term for extraneous variables that influence DV measurement and are evenly distributed across experimental conditions. Statistics aim to see true results through this noise, and good experimental designs should reduce it as much as possible.

Quasi

Experiments where participants cannot be randomly allocated due to plausibility or ethical reasons, for example if comparing depressed vs non-depressed individuals.

Between

Method of participant allocation where a participant only appears in one condition. This is used in experimental and quasi-experimental studies.

Within

Each participant is used in every condition. Can be used in experimental and correlational studies.

Cross sectional

Studies that only look at a single moment in time, for example hours of sleep at age of 2.

Longitudinal

Studies an individual over a long period of time, for example hours of sleep at ages 2-4.

Factorial

Study design where multiple IVs are examined, for example vaccine effectiveness (1) across ages (2).

Interaction

Term for when the effect of one IV depends on the level of the other.

Meta analyse

Studies that examine effect size by looking at multiple studies which investigate the same thing. Systematic with very strict criteria.

Case study

An in depth analysis of a person, an event or a group. An idiographic method that can utilise multiple different methods. Has limited generalisability, no causal statements but provide rich data.

Observation

Study where individuals are observed in their natural environments, either structured or unstructured. No causal statements, observer bias, but high external validity.

Structured

Type of observation where the environment is structured by the experimenters and then watched.

Archival research

AKA data mining. Answers questions with data already collected, often for independent purposes. Avoids collection bias, but researcher can still bias data collection. Data scraping is often used.

Data scraping

Method of archival research where a piece of code is developed that will look across multiple databases for data that fits certain criteria.

Survey

Type of study that directly asks participants questions. Can be an interview or a questionnaire. Can be used to generalise across a population or to assess correlations.

Basic

Type of research that is mechanistic and has little obvious applications.

Applied

Type of research that is done in order to answer a specific relevant question.

Confounds

Nuisance variables that vary systematically with IV, influence DV and threaten internal validity. There are three types.

Person

Confound where research is affected by individual differences between participants. Can be avoided by random assignment or matching groups based on these differences, such as gender.

Operational

Confound where research is affected by the measure also measuring another factor on accident, threatening construct validity. Can be avoided by a refined operational definition.

Procedural

Confound where research is affected by a researcher accidently manipulating another variable alongside the IV, which threatens internal validity. Can repeat the study while controlling for this new potential variable.

Carryover

Effect where a previous experimental condition alters a participants behaviour and alters their results in proceeding conditions. Issues include practice, fatigue, priming and interference.

Interference

Being exposed to something can alter your ability to something else, for example playing chess may be affected by past experience playing draughts.

Maturation

Issues of longitudinal studies where over time an individual will change, perhaps due to age or even symptom fluctuation.

Instrumentation

Effect where the measuring instrument has changed and is no longer identical. An example is it referring to out of date things, such as payphones.

Attrition

People dropping out of a study over time for various reasons, which may effect external validity as these people may be different overall to those who decided to stay.

Regression to the mean

Effect where an extreme score will become less extreme over repeated testing. This is because an extreme score is likely due to an extraneous variable, such as mood, so when repeated this will no longer be present.

Experimenter

Type of bias where the experimenter can act in a way that alters participant behaviour. Alternatively, they could also look at the data in particular ways, aiming for a certain result.

Participant

Type of bias where participants behave systematically in a way the study does not expect. This includes the Hawthorne effect, evaluation apprehension and good subject effects. Dealt with by double blind technique, manipulation checks, debriefs, cover stories, indirect measurement and guaranteed anonymity.

Hawthorne

Effect where people will alter their behaviour because they know they are being observed or studied.

Double blind

Technique where both the participants and the experimenters are unaware of group assignment, preventing experimenter and participant biases.

Selection

Bias where the sample is not representative due to interactions between individuals with biological, behavioural and psychological conditions and their environment, for example an in patient ward having generally more extreme cases of depression than the general population.

Sampling

Bias where the sample does not correctly represent the research population.

Ceiling and floor

Effects where the tool or measurement is too difficult or too easy, limiting variation. For example, an easy spelling test will lead almost everyone to score in the high ranges of the test.

Positive

Type of control group where another manipulation is used, such as a placebo or sham operation.

Open label

Term for when a participant knows what group they’re apart of.

Counterbalancing

When conditions are not performed in the same order for all participants or conditions, drowning out carryover effects.

External

Type of validity that measures how well the measure can generalise across times, populations and environments. Two key elements are mundane realism and experimental realism.

Experimental realism

Element of external validity that refers to whether participants feel the manipulation as real, therefore leading to real behavioural responses.

Mundane realism

Element of external validity that refers to whether the experiment is similar to everyday life, therefore allowing it to be generalised to real life.

Convenience

Type of non-probability sampling which samples available participants at a specific time and place.

Purposive

Type of non-probability sampling which samples based on predetermined criteria or characteristics. Can be either quota or snowball sample.

Quota

Type of purposive sampling where certain percentages of the sample must be made up of certain characteristics, for example age and gender.

Snowball

Type of purposive sampling where members of the group to advertise to other members within the group, as these would be otherwise inaccessible to a researcher, for example drug users.

Random

Type of probability sampling where people are selected randomly from a group. Rarely used in practice.

Stratified

Type of probability sampling where the sample must have a certain percentage of subgroups to make it similar to the real population.

Cluster

Type of probability sampling where an entire group is sampled, often used when the population of interest tends to clump together, such as hospitals.

Non-response

Bias where those who respond are systematically different from those who don’t, for example those who responded all use Instagram regularly.

Selection

Type of bias where those who chose to be in the study are different from those who don’t, for example those who chose to participant enjoy the subject.

Social desirability

Bias where people give responses based on what they think is desirable to others, for example how often they shower.

Big data

Data that is too large and complex to be processed by normal software or by individuals.