Psychology - Research methods

Why are experiments important?

• They are the gold standard of research in psychology

• they can establish cause + effect relationships and help remove the cause if causing negative effects

What are three characteristics of experiments?

• IV - Independant variable, this is the variable you manipulate and change in an experiment

• DV - Dependant variable, this is the variable you measure in an experiment

• A controlled enviroment, in order to have control we need to identify all the extraneous variables in an enviroment that can influence ppts behaviour and keep the same these are called control variables

Control variables fit into three catagories; ppts variables, researcher variables and enviromental/situational variables

What are the three control variables and how can they impact a researchers experiment?

• Ppts variables, mood - biggest influence on behaviour that researchers can’t control, motivation, events in ppts life e.g daily demands

• Enviromental variables, other peoples behaviour/moods, heating/lighting/comfort

• Researcher variables, treatment of indivduals, ppts bias on age, gender, appearance changing their behaviour

What is a hypothesis?

This a prediction of the outcome of your experiment that you should write before the experiment begins, an experiment should prove/disprove your hypothesis and provide evidance

What are three types of hypothesis?

• Null hypothesis

• experimental/research hypothesis - non directional

• experimental/research hypothesis - directional

What do you not do when writing a hypothesis?

Do not explain why

What is the IV in the stroop test?

What is the DV in the stroop test?

whether the print colour + word meaning match is the experimental condition

correct or incorrect + reaction time in milliseconds

What does the stroop test investigate?

cause + effect relationship, is there a relationship between print colour and word meaning incongurency and time taken to name colour and font

What are the ppts, enviromental and researcher variables in the stroop test?

• ppts variables - eyesight, colourblindness, tiredness, focus, mood, motivation

• enviromental variables - distractions, lighting

• researcher - expectations of results

What is standardisation? how can it be used for the stroop test?

keeping things the same for all ppts

• ratio of congruent and incongruent words

• equipment setup e.g screen size

• room used for testing

• the number of words

• instructions given to ppts

What is a laboratory study/experiment?

They are in a controlled environment with;

• reliable findings

• control of extraneous variables

• and we can be confident the IV will influence the DV

What are some weaknesses of Lab studies/experiments?

• They lack ecological validity where the experiments do not reflect how people act in everyday life

• People try to work out what is expected of them in experiments - Demand characteristics

What are field studies/experiments?

• experiments in a real world setting

• natural behaviour as the person will be in a natural setting

• tend not to gain demand characteristics

What are weaknesses of field experiments/studies?

• the results are less reliable as the environment is not controlled so we can’t be confident IV is influencing DV

What are Quasi experiments?

These are experiments where the researcher does not choose the IV, the IV “chooses itself”

• They take advantage of natural differences between groups of people like age;

• Age can influence attitudes, personality, intellectual ability and memory

• we cannot randomly allocate ppts

What is one problem with quasi experiments?

• there is an issue drawing conclusions because of extraneous variables

• the IV is not the only thing that results will measure that will influence the researchers findings

What is another example of a quasi experiment?

• social class, people find themselves in either upper, middle or working classes

• the DV would be testing intellectual ability whilst the IV would be the different people we test

• however, an extraneous variable would be someone’s social capital giving them more opportunities for academic enrichment

What are natural experiments?

• they are rarely used by psychologists

• they are uncontrolled experiments that run themselves

• the experiment won’t choose the IV or DV

• the researcher analyses differences in data

What is an example of a natural experiment?

The twin towers terror attack in 2001. Flashbulb memory - where you have a vivid memory of an event that is shocking

• People were naturally different distances from the twin towers during the attacks

• the researchers developed an IV based on the distance from the towers and a DV based on peoples recollections of the event

What is the issue with natural experiments?

• there are so many uncontrollable variables that researchers have to be cautious of about drawing conclusions

• they are often unethical and unreliable

what is the idea of experimental designs?

• It is about the allocation of ppts to a control and experimental condition

What are independant group designs?

Different ppts will appear under different conditions

What is a repeated measures design?

the same ppts will appear under each condition

What is a matched pairs design?

Same as independant groups but researchers identify shared traits of ppts and pair them up. Then, the pairs are split up and tested in different conditions

How is reliabilty important in an experiment?

• Method can be reliable by being standardised so it can be repeatable

• findings can be reliable results from experiments are the same each time an experiment is ran.

How is validity important in an experiment?

when we are considering whether a piece of research is valid we are asking whether we have measured what we intended to measure

we could aslo ask is this behaviour a normal representation of life - ecological validity

How can demand characteristics impact validity?

where your ppts try to meet the researchers expectations and try to be a ‘good ppts’ therefore reducing ecological validity

How can investigator effects impact validity?

when the researcher unconsciously passes on their expectations. E.g verbally through tone of voice, facial expressions and gestures. This can be avoided through a set of instructions. This creates demand characteristics

What can we do about investigator effects?

• Run blind studies where the researcher does’nt know what the hypothesis is

• Drug trials, double blind studies, neither the researcher or ppts knows who is given the drug and who is given the placebo

What is sampling? How are samples created and why?

Sampling is about deciding who will take part in research, researchers will have in mind a target population in mind for example an age group or people of a paticular ethnic group

We cannot conduct research on complete groups so we must take samples that will hopefully represent a target population

Whar are 5 different methods of sampling? What do they depend on?

Samples will take place depending on the resources available

• Random

• Systematic

• Oppurtunity

• Volunteer

• Stratified

What is random sampling? what are the strengths and weaknesses?

• There is a list of all possible ppts as a sample frame

• The researcher will randomly select the ppts e.g through an online random name generator

• Strength - the researcher does’nt choose their ppts

• Weaknesses - there is a chance the sample isn’t representitive , there must therefore be a sample frame

What is systematic sampling? What are the strengths and weaknesses?

• List of all ppts, the researcher will then pick every nth ppt

• We decide the nth ppts by division, we divide the total number of ppts by the number needed for sample e.g 40 divided by 10 = 4, so we take every 4th ppts

• strengths, unbiased sellection

• weaknesses, the order of ppts could be biased so the names must be randomised

What is oppurtunity sampling? what are the strengths and weaknesses?

• Non random, taking people who are willing and able by “asking them”

• Strengths - quick and easy process for researcher, large samples

• weaknesses - unrepresentative as the sample is subject to researcher bias

What is volunteer sampling? what are the strengths and weaknesses?

• Non random

• Researchers will advertise their study and wait for people to sign up

• strengths - Eliminates researcher bias, is easier for the researcher and the volunteers will be motivated

• weaknesses - unrepresentative of target population, volunteers may have a hidden motive or be more interested or confident

What is stratified sampling? what are the strengths and weaknesses?

• Step 1, calculate the % of catagories of ppts in the target population and translate that % in your sample

• step 2, use a sampling method like random sampling to create a sample

• strengths - unbiased, most representative

• weaknesses - it can be more time consuming, have to use external date to work out sample proportions

What are ethics in psychology and how have they been set?

• Ethics are an agreed set of rules about how we treat participants whilst conducting research

• The code of conduct has been desingned by the British psychological society B.P.S

• A psychologist must attain approval from their university to carry out their research

What are the four guiding principle in the B.P.S code of conduct?

• Respect

• Responsibility

• Competance

• Integrity

What ethical issues can rise when conducting research?

• Informed consent, ppts need to know “as far as possible” what they are letting themselves in for such as the goal and tasks

• Right to withdraw, can leave research at any time. Ppts will have to be reminded at the beginning, middle and end, if the ppts decide they don’t want their data used it is withdrawn and destroyed

• Debrief, after the experiment the exact nature of the research is explained

What should researchers avoid during their experiments?

• Deception, sometimes necessary for the experiment to take place so it must be kept to a minimum (leave out information rather than lie)

• No physical harm, or psychological harm e.g excessive stress, anxiety and shame. Making sure ppts leave experiment reletivley the same as when they joined

What rules should researchers uphold when collecting data?

• Anonymity, data shouldn’t be identifiable so they must replace name with ppts numbers

• Confidentiality, data collected and ppt behaviour should not be discussed with anyone outside research team

What are descriptive statistics?

• Mean average

• mode

• median

• range

We calculate these 4, they are measures of central tendancy, they have a single value that represents your data set

• Standard deviation, which we interpret

We shouldn’t really draw conclusions from descriptive statistics because our results could’ve arisen by chance

What are inferential statistics and tests?

They allow us to test a hypothesis and decide whether our results are based upon chance or the experiment (IV affecting DV)

There are lots of different inferential statistics tests which one you choose depends on the type of data you have collected

What is nominal data?

It is catagorical data, places ppts into catagories e.g frequency table, where you can count the indivduals in each catagory like hair colour

What is ordinal data?

rating scale data, measure the quality of something on the scale. could be making ppts answer how much they agree from 1-10. It is a subjective measurement

What is interval data?

continuous data which is an objective measurement. The intervals between data are consistant

e.g heart rate measured in BPM or reaction times measured in seconds

What types of research are there, what designs could they be in?

Experiments, an experiment which has an IV influencing the DV

• There are 3 experimental designs they could use; independant groups with an unrelated design, matched pairs design and repeated measures design (both of these are related designs

Correlations, looking for an association between two variables, a positive and negative association

What inferential statistic test would you use for unrelated nominal, ordinal and interval data from an experiment?

• Chi-squared

• Mann Whitney

• unrelated T-test

What inferential statistic test would you use for related nominal, ordinal and interval data from an experiment?

• sign test

• willcoxen

• related T-test

What inferential statistic tests would you use for nominal, ordinal and interval data from a correlation?

• Chi-squared

• Spearman’s Rho

• Pearson’s R

give an example of a hypothesis, what are one and two tailed tests?

“Ppts will recall more digits when digits are read out in pairs compared to being read out one at a time” this is a directional hypothesis, it is one tailed and we run a one tailed test on it. For a non directional hypothesis we run a two tailed test, however one tailed hypothesis are easier to use as it’s easier to get a significant result

What do all statistical tests do? What can happen with a significant result?

All statistical tests will test the null hypothesis to see if any observed differences are due to chance, if you get a statistically significant result you may reject the null hypothesis meaning there will be no differences with both conditions and accept your own research hypothesis

What is a significance level, what are the two types of errors?

Before running statistical tests we must choose a significance level (error rate), we are happy to accept this level

• optimist error, we reject the null hypothesis and conclude a difference in conditions or an association (type 1 error)

• Pessimists error, we accept the nulll hypothesis when we should reject it

What are the two significance levels?

There are two error rates (significance level)

• 1% significance level, an error occurs 1/100 times, this shows 99% confidence in hypothesis. Only used in drug trials and one off experiments

• 5% significance level, 5/100 times we run experiment we face error. We are 95% confidant used for standard research

How do we complete a sign test?

• we take the numerical value of the least frequent sign and ignore the zeros e.g 2 (This is the S value)

• We calculate the N value, this is the total number of ppts minus the number of ppts who got zeros e.g 12

• We “ look up” our observed value of S on a critical values table. We look up if it is a one or two tailed hypothesis, it being at significance level 5% and the N value of 12

• The we apply this rule to see whether our result is due to chance or not

How can we conclude sign tests?

We can conclude that our statistical test has given us a statistically significant result. Our observed value of S = 2 is equal to the critical values table. We can reject the null hypothesis and accept our one tailed directional hypothesis

Data distribution, what do parametic tests require?

These tests require a normal distribution of data. Some examples of parametic tests are; T tests, pearson’s R and Chi-squared

Data distribution, what happens with a skewed distribution of data when running parametric tests?

If you have a skewed portion of data we run a non-parametric test

Non-parametric tests include; sign tests, wilcoxen, Mann Whitney, spearman’s Rho

Some have a positive skew and a negative skew

What are observation studies and what are the two types of observation?

For instance the Bobo doll experiment is a good example of controlled observation done through covert observation and non participant. Behaviour can be samples

• covert = ppts are unaware of observation

• Overt = pots are aware of observation

What is triangulation?

This is where we use more than one method to research the same topic. If results are the same, we can be more confident of our conclusions e.g we may pair observations with interviews

What types of obervation are there?

• naturalistic, everything is left as normal and researchers don’t interfere in any way like in ofsted inspections

• Controlled, aspects of environment are manipulated with and IV affecting the DV, mostly done in a lab setting

• Covert

• Overt

• Participant

• Non-participant

Usually, psychologists are controlled, convert and non participant

What are unstructured, structured, and behavioural categories?

• unstructured, all relevant behaviour is recorded and no systematic plan is made before the observation starts. Useful for a pilot study

• Structured, systematic obervation, recording plan is made beforehand meaning behavioural categories are devised and a behavioural sampling method is chosen

• Behavioural categories are precise definitions that will be recorded, (operationalism). Behaviour has to be defined so that each observer can be as objective as possible. A tick list is created so the observer can quickly record them

  • for example making a draft tick list for rats behaviours in skinners experiment

What are sampling procedures?

• what are the three types?

• continuous observation, not really possible so sampling behaviour is crucial

• Event sampling, counting the amount of times a target behaviour occurs in a target individual

• Time sampling, recording behaviours in a given time frame, e.g noting a behaviour every 5 minutes

How can we evaluate naturalistic and controlled observation?

Naturalistic

• strengths

  • good ecological validity

• Limitations

  • no control over extraneous variables

  • Hard to replicate

Controlled

• strengths

  • limits extraneous variables

  • Improves reliability

• Limitations

  • lacks ecological validity, no natural behaviour

What is an interview?

• a structured conversation using open ended questions, they are usually face to face where the interviewer can pick up non-verbal cues

• They provide qualities data that can be audio recorded and transcribed either manually or through softwares

• Analysis either involves content or thematic analysis

What types of interviews are there?

• structured, pre set with open ended questions, asked in a fixed order, e.g market research

• Semi-structured, same as the above, but with follow up questions, job interview

• Unstructured, no set questions on a particular topic and interaction is free flowing, cognitive interview

How can we evaluate interviews?

Strengths

• idiographic; unique experiences are explored

• In-depth info

• sensitive info can be explored

• Starting point for questionnaires

Limitations

• only suitable for a small sample

• Can’t easily find patterns and trends as it is qualitative data

• Subjective data analysis

• Takes 1 hour + transcription

What are questionnaires?

• self report method which produce quantitative data like %, many are easy to describe data due to technology

• Many do rely on honesty

• Allow us to analyse data with bar charts, pie charts, etc

• They use closed ended questions with forced choice response like yes/no and likert scores 1-10

How can we evaluate questionaries?

Strengths

• nomothetic and scientific

• Can have very large sample sizes and can be distributed easily

• Collecting very reliable data that can be generalised to sample size and they have standardised questions

Limitations

• Self report, we rely on others to be honest some may lie for social desirability which reduces the validity of our findings

• Reductionist, quantitative data can’t show depth

• Cannot be used for sensitive topics

What is thematic analysis?

Conducted on qualitative data and is left as qualitative data, there is no conversion to quantitive data

• a method for uncovering the intended meaning in a piece of text/image

How?

• By thoroughly reading the text to try and understand what the writer was expressing. This is read 3-4 times

• Analysing language and phrases used by ppts and interpreting what they meant

• Unlike content analysis there are no coding units

• We are looking for themes that come up to the surface, e.g anger, desperation, joy

• Once theme is established, we write a text summary explaining the main message by using the themes to support what we say

(Subjective)

How can we evaluate thematic analysis?

Strengths

• in-depth analysis provides a holistic picture of someone’s experience

• Idiographic, valid

Limitations

• subjective interpretation of ppts data meaning analysis may be bias

• Non scientific approach, rarely used in psychology

• Only can do them in a limited sample and they are not generalisable

What are case studies?

An in-depth analysis on a single individual

• it’s an idiographic approach using a variety of methods like brain scans, interviews, observations, and performance tests

• usually conducted over a long time e.g 20 years

How can we evaluate case studies?

Strengths

• In-depth analysis allows researcher to gain holistic insight

• Useful starting point for the future nomothetic approach

• Can use abnormal to inform us about the normal

Limitations

• can’t replicate

• Can’t generalise findings

• Researchers may become over-involved work over many years so loss of objectivity and professionalism

What are pilot studies?

• small scale studies carried out before the study proper, to test procedures and materiel you want to use

How can pilot studies be used in;

• observation

• Lab studies

• Questionnaires

• Interviews

• in observations, behavioural categories are clearly defined, data sampling method - event or time intervals? Where to observe from, how to record data?

• Lab studies, instructions to your ppts, what variables you need to control. IV and DV, do they work? Trying the task, making sure ppts can do it

• Questionnaires, clarity of questions, response choices, format, etc

• Interviews, clarity of questions, would it offend someone, is it a comfortable setting?

How can we evaluate pilot studies?

Strengths

• saves time and money as we can identify mistakes meaning they shouldn’t come up in the actual experiment

• Results are more reliable as method is standardised and replicable

Limitations

• could take time and money however

• Might lead to unconscious researcher bias, they will try and replicate the results in the study proper that they got in the pilot study

• To avoid this we use double blind procedures in the study proper where neither the researcher nor the ppts know the condition they are in

How do we construct psychological reports?

• Title, a statement or question that would interest the reader

• Abstract, brief outline of your research procedure and findings. Aim, method, findings, and conclusions

• Introduction, aim, other peoples research and your hypothesis, references

• Method and procedure, writing this in enough detail so someone can replicate your experiment. This includes the experimental design, materials used, ppts who took part, ethical considerations

• Findings and results, descriptive and inferential statistics

What else do we include in psychological reports?

• Discussion

  • Explain findings

  • Relate findings to previous research

  • Evaluation of study

  • future research

  • conclusion

• References, Full details of source material referred to in introduction and discussion (reducing playgrism and boosting researchers careers)

How is a harvard reference structured?

Author surname, Initial. (year) book title. City : publisher

Abel, T. (2028) psychology. York : university of York

How do peer reviews work?

It’s a quality control process where we aim to ensure that research that is published is good quality

• Researchers who work at universities want to publish their work in an academic journal

• These are mostly online and are overviewed by ‘peers’ who are on a board of editors, specialists in fields

• Peers decide if research is good by; looking for any errors and mistakes, checking data analysis, checking for plagrism. Is it good quality? Is it worthwhile?

Does peer review work?

Yes

• prevents research being published that has errors in it

• does work as a plagerism check as the experts in your field have 20/30 years experience and know the area well

No

• Board of editors are not unbiased, they will have preferred theories and research methods they don’t want challenged

• May be less inclined to publish research that challenges established theories

• May be more inclined to publish positive results

What is content analysis? How does it work?

A method for converting qualitive data into quantitive data

How?

• By systematically identifying the thoughts, feelings, and behaviours being expressed in text and counting the number of times each one occurs

• Each type of thought, feeling, and behaviour is called a coding unit

• Before we conduct content analysis we decide on the coding units we are going to use

• Thus, coding units are pre-determined

Whats an example of conducting content analysis?

text message analysis research

• This is qualitive data analysis in the form of texts and images/emojis

• In this case, our experiment is to investigate gender differences in text messages

• Example hypothesis; text messages from females will be more emotionally expressive compared to those of males

• Pre-determined coding units in this example: exclamation marks, use of emojis, words in caps

Sampling techniques

• measuring last 5 texts

• texts sent in the last week

• randomly selecting 5 text exchanges

How can we evaluate content analysis?

Strengths

• Patterns and trends can be established

• Standardised coding units mean different comparisons made between groups

• Reliability can be established

• Coding units decided before means we can test a hypothesis

Limitations

• Subjective process, person who analyses data will be biased

• unscientific but not as bad as thematic analysis

• reductionist, misses out on thoughts and feelings we haven’t pre-determined

Inferential stats table - NUC?

Nominal unrelated - chi squared

Inferential stats table - OUM?

Ordinal unrelated - Mann Whitney

Inferential stats table - IUT?

interval unrelated - unrelated t-test

Inferential stats table - NRS?

Nominal related - sign test

Inferential stats table - ORW?

Ordinal related - Wilcoxen

inferential stats table - IRR?

interval related - related t-test

Inferential stats table - NCC?

Nominal correlation - chi-squared

Inferential stats table - OCS?

ordinal correlation - spearmans rho

inferential stats table - ICP?

interval correlation - Pearsons R

What is counterbalancing?

This is the process in which ppts are split up into half and experience conditions in a different order. This is done to reduce order effects. For example, half the ppts will experience condition A first, and then condition B second. The other half will experience condition B first, and then condition A second. This in turn should reduce the order effects.

What is another term for a directional hypothesis?

One tailed test

What is another term for a non-directional hypothesis?

Two tailed test

What is mundane realism?

This is concerned with whether the task that people do is very reflective of something that they would do in real life e.g could be argued for Peterson + Peterson’s consonant trigram experiment

What is the name for being specific about those variables you are measuring?

Operationalisation

What do you want to be able to reject in research?

The null hypothesis

What is a type 1 error?

When we accept the hypothesis but should reject (we think it is significant but its not). This is known as a FALSE POSITIVE and is most likely to occur if we use p≤0.10 (10%).