4- sampling and data collection

Methods of Data Collection in Psychology

Population

Entire group of people of interest e.g all psychology students

Sample

subgroup of population. e.g psych students in the room

what should I include in sample

Age, pattern of life e.g if ur researching alcohol are they students or alcoholic.

Effects may differ depending on the population

Who do you want to make inferences about?  Appropriateness of the sample

depends on the population definition

sample appropriate if?

Goal of statistical inference. Gaining knowledge about the population

A sample is appropriate when we can generalise from the sample to the population

Generalization is possible if the sample is representative of the population

Sample A is representative of the population

Sample A is the right answer because it is an exact mini-version of the main group.

Think of it like tasting a spoonful of soup to see if the whole pot needs salt:

• The Population (The Whole Pot): Has an equal mix of 50% men and 50% women.

• Sample A (A Good Spoonful): Also has 50% men and 50% women. Because the mix is identical, whatever you find in Sample A applies to the whole population.

• Sample B (A Bad Spoonful): Has 20% men and 80% women. This sample has way too many women and not enough men, so it does not reflect the real world.

Higher sample

better representative less likely to have random luck skewing your results.

Representativeness

• Samplpes are more representative of the population if their composition is similar to the population

• - large samples are more likely to be representative of the population that small samples

selection bias

Selection bias occurs when the sample is non-representative and this lack of representativeness influences results.

happens when you pick the wrong mix of people for a study, and that bad mix messes up your results

selection bias examples volunteers

more likely to the females, highly educated, from higher social class

Most psychology students are WEIRD participants

WEIRD participants aren’t representative of the worlds population —— DANGER OF SELECTION BIAS

How can we ensure that a sample is representative?

Sampling methods

Sampling methods

The technique used by researchers to select a smaller group of participants from a larger population. How individuals from the population are allocated to the sample

probability sampling - high degreee of representativeness

Each member of the population has a specific probability of being chosen as a participant , making it possible to know how likely it is that a sample will be similar to the population before drawing sample.

Non-probability sampling low reprresentativness

NOT RANDOM Unknown probability of being selected, impossible to know how likely it is that a sample will be similar to the population

Probability sampling - simple random sampliing

individuals have a known and equal probability of selection, .

Samples drawn via simple random sampling are perfectly representativ

A method where every individual in a population has an equal chance of being selected, like drawing names from a hat.

Sampling methods: Probability sampling- CLUSTER PROBABILITY SMAPLING

Divide population into clusters, select random, sample of clusters, every individual within cluster is part of the sample

Probability sample e . g How good are British primary school students at reading

Cluster- schools

Population- all school kids in British primary school s

Rando,ly select schools

Cluster sampling advantages in this eg

• No need for a list of all students

• Known probability of selection (1/all schools)

• on average, Sample will be representative of School level characteristics (e.g size, rich/poor)

Non probability sampling, oppurtunity sample

ppl that happen to be available, restricted sample e.g family and frinds

Probability of inclusion unknown (what is the probability for being your friends?)

Selection bias likely

Non prob Purposive sampling

ppl included/excluded based of researches judgement

A non-probability method where the researcher uses their own judgment to pick specific participants who fit the exact criteria of the study.

Purpose e.g

Example: “How do senior managers cope with work-related stress?”

 Paul works as a senior developer at Microsoft and leads a team of 10 people.

 Michelle is a professor in biostatistics.

 Richard is the headmaster of a primary school.

 Anna is the CEO of a start-up company with 30 employees.

Probability sampling uses random chance (giving fairer, more representative results), while non-probability sampling relies on convenience or judgment (easier to do, but higher risk of bias).

Physiological data e.g

EEG, FMRI, Eye tracking, EMG, skin conductance

Physiological data is

• quantative

• Natural response of the body to the study enviroment

• requires specific apparatus (e.g leg device, firm scanner)

• usually continuous over time, million data point per preson

• incentive preprocessing to make it interpretable

Physiological data advantages

Most objective type of data

No verbal communication needed

helps tp connect psychological theory with findings from medicine

Physiological data disadvantages

• needs expensive equipment

• Heavy preprocessing

• can’t measure someones social behaviour

behavioural data

Data collected by directly observing and measuring what participants actually do, such as their reaction times, choices, or actions during a task or experiment.

Behavioural data is

• Quantative

• Measurable

• Small range of observable behaviours in controlled enviroments

• Amount of data collected from each participant ranges from one data point to hundreds

• Encountered in cognitive psychology, social psychology, developmental psychology, consumer psychology

Behavioural data advantages

Breaks down complex human behaviour into observable interpretable peices

Allows for testing of psychological theories

Cost efficient

Disadvantaegs behavioural data

Behaviour may not be representative of real world behaviour

if in person- needs lab space

If online - data quiality may be poor

Survey data

questionare, structued intervuew

Survey data is

Typically quantative

predetirmed questions with limited answer options

Can be collected via mail, online, on phone, face 2 face

one datapoint collected per question

Efficient way of data collection, but sl=election bias needs to be considered

Encountered in social psychology, personolairt psychology clinical psy

advantages survey

cheapest, most efficient data collection method

Huge samples posisble== high reprasentiveness

Good for asking about simp;e well understood real life behaviours / opinions

Survery disadvantages

Limited answer options leave little space for self expression

Seecion bias possible

Participnts may give socialy desirable instead of true answers

test data

test data is

Quantative

predetirmined questions with limited answer option that together measure someone ability or psychological contrsuct

Test data is combined to a single test score per person.

Subdisciplime of psychology for Deve;oping good tests

test data advantages and disadvantages

:)= + Making invisible (“latent”) constructs

visible / measurable (e.g., “math

ability”)

+ Can (often) be used for individual

diagnosis

+ Usually easy to collect and analyse

once test has been constructe

Bad= Limited application domain

(measuring invisible personal

characteristics)

- Requires lots of validation before a

test is good enough

(psychometrics!)

- Test data alone is usually not

enough to test psychological

theories

Interview data

Semi structured interview, focus group interviews

Interview data is

Qualitative

 Largely predetermined questions (“interview guide”), answers are completely

open

 Interview data emerges from transcribing verbal information from audio or video

recordings

 Many pages of text per participant (e.g., 1 hour interview ~ 20-25 pages)

 Most often encountered in clinical psychology, organizational psychology,

forensic psychology

Advantages interview

+ Participants have freedom to

express themselves

+ Opportunity to explore complex

phenomena that are not well

understood yet

+ Building personal rapport with

participants

Disadvantages interview data

Requires participants’ self-insight in

studied phenomena

- Time intensive (interviews &

transcriptions)

- Results are dependent on

interviewer (skills, opinions,

interviewer-interviewee dynamic)

and not replicable

Observational data

Structured observation, participant observation

Observational data

Can be quantitative or qualitative

 Observation of natural behaviour in a natural or artificially created environment

 Behaviour is not restricted, but researchers may focus their attention on specific

behavioural aspects (e.g., does a baby cry if their parent leaves the room?)

 Quantitative data: Behaviour is coded (on the spot, e.g., using checklists, or after

the fact based on video recordings)

 Qualitative data: Behaviour is described verbally (e.g., researcher diary)

 Most often encountered in clinical psychology, developmental psychology,

cultural psychology

advantages

Participants are not restricted in the

range of possible behaviours

+ Opportunity to explore complex

behaviours that are not well

understood yet

+ No self-insight of participants

required

Disadvantages

Often time intensive (e.g. coding of

hours of video material)

- Overt/covert observation: Balancing

observer effects and ethical data

collection

- Behaviour categorization can be

subjective / dependent on observer

Cross country studies

A research design that compares different groups of participants (such as different ages) at one single point in time.

Longitudinal studies

A research design that studies the same group of participants repeatedly over an extended period of time to observe changes.

Over what period of time is data collected from

participants? Cross sectional studies-

reasr4chers collect data from participants at a single point in time ( most psych studies are cross sectional)

In longitudinal studies,

researchers collect data from participants on multiple occasions over a longer period of time.

Always repeated measure data, observe trends over time and someones development, distinction between cohort studies, panel studies a.s.o

Cohort studies (longitudinal)

Participants share common experience or demographic trait (e.g., being born in the same year in Liverpool)

 Participants are followed forward and regularly assessed

Distinction between cohort effects and age/time effects

used in developmental psych

E.g to understand

• Pick a Cohort: Researchers gather 1,000 babies all born in the UK in the year 2000.

• Follow Them Over Time: They check in on these same people every few years as they grow up (at ages 5, 15, 25, and 40).

• Look for Patterns: Researchers compare those who experienced something (like playing video games daily) against those who didn't to see how it affects their health or careers later in life.

Why it is useful

• Finds Causes: Because you follow people from start to finish, you can see what happened first and what outcomes followed later.

• Real-World Observation: Researchers simply watch life unfold naturally without forcing people into experimental groups.

Panel studies (longitudinal)

A longitudinal study that collects data from the exact same group of individuals repeatedly over time to track changes in their attitudes, behaviors, or outcomes.

Participants who drop out can often be replaced with other people with the same

demographic characteristics

E.g panel

understanding Society (The UK Household Longitudinal Study)

• What it is: One of the largest panel studies in the world, run by the University of Essex.

• How it works: Researchers survey the exact same 40,000 UK households every single year.

• What they track: They ask the same people about their physical health, mental wellbeing, jobs, family life, and income year after year to see how changes in the UK economy or society affect individual lives over time.

Insensitive longitudinal studies

A longitudinal design where participants complete the same short survey at least once a day over weeks or months (often via smartphone apps) to track real-time behavior and thoughts.

Advantages and uses of intensive longitudinal studies

• Collects data in a person's natural environment, which reduces inaccuracies caused by memory bias. Commonly used in clinical, social, and health psychology.

E.g intensive

Example: Studying University Exam Stress

• The Setup: A psychologist wants to know what triggers stress in university students during exam season.

• The Pings: Participants download a study app on their phone. For 3 weeks, the app sends them a quick notification ("ping") 4 times a day (at 10 am, 2 pm, 6 pm, and 10 pm).

• The Quick Survey: Every time their phone buzzes, the student spends 30 seconds answering three questions right there in the moment:

  1. "How stressed do you feel right now on a scale of 1 to 10?"

  2. "What are you doing right now (studying, scrolling social media, hanging out with friends)?"

  3. "Have you drunk any caffeine in the last two hours?"