Psychological research and scientific method - A2 Psychology

Popper (1934) What is falsification? (Science) A01

'No matter how many white swans we observe, we can't justify that all swans are white' so we seek to disprove the theory beginning with a null hypothesis that not all swans are white, if we see no black swans we can say with reasonably certain the NH is false so reject and develop an alternative H which we can then accept with reasonable certainty (best approximation to the truth)

What is peer review? (2) (Validating new knowledge) A01

'Refereeing' = assessment of scientific work by others who are experts in the same field to ensure high quality Internet pages such as Wikipedia are a form of peer review where the info is policed by 'wisdom of crowds' approach (reader's decide the validity and post comments/edit articles accordingly)

What is the purpose of peer review according to the parliamentary office of science and technology (2002) ? (3) (Validating new knowledge) - H - A01

1 - Allocation of funding (government and charity bodies) e.g. Medical research council (2008/9) had £605 million to spend as Hansord (2007) states, and they require PR to decide which studies are worthwhile 2 - Publication = prevent incorrect or faulty data entering the public domain 3 - University department rating e.g. research excellence framework = future funding depends on good rating from RAF peer review

2 examples of recently invalid data - J.E.A - (Validating knew knowledge) A01

1 - Professor Marc Hauser - Harvard Uni (2010) was unable to provide evidence for many published conclusions 2 - John et al (2012) 2,000 psychologists, anonymous report and 70% admitted to cutting corners in reporting and 1% falsifying data

Report your own study (6)

1. Abstract (summary of aims, H, method, results and conclusions) 2. Intro (review previous research, aims and H) 3. Method (design, participants, materials, procedure and ethics) 4. Results (descriptive statistics = tables/graphs with frequencies, measures of dispersion and central tendency, inferenrial = reject ot accept NH, QLT = Categories and themes) 5. Discussion (summarise results, relationship to past research, methodology e.g. criticisms and improvements, implications and suggestions 6. References

What are the 2 types of graphs? (Descriptive statistics)

1. Bar chart (height = frequency) 2. Scattergram (correlational data, bottom left to top right = positive, top left to bottom right = negative, no detectable pattern = zero or no correlation)

What are the 3 levels of measurement? (Inferential statistics)

1. Nominal data (categories) 2. Ordinal data (ordered) 3. Interval data (units of equal intervals)

Informed consent (2) (Dealing with ethical issues)

1. People under 16 years or those with impaired capacity to consent need addtional consent of parents or those legally responsible for them 2. It's reccomened that one pilot test informing and debriefing is carried out on those with lower end literacu levels than the range expected in the planned research

An example of the 7 general principles in qualitiative research

1. Read and re-read transcript to undertand the meaning and perspective of participants 2. Break into meaningful units 3. Assign a label or code to each unit 4. Combine the simple codes into larger categories or themes 5. Check for emergent categories by collecting new data and applying the categories 6. Final report should discuss and use quotes or other material to illustrate emergent themes 7. Conclusions can be drawn

Three steps to choosing inferenrial statistics

1. Selecting the right test 2. Justifying choice (spearman = correlation, chi-square = categories ans independent, mann-Whitney = difference, independent and ordinal, wilcoxon = difference, repeated and interval) 3. Stating conclusions (include observed value, greater or less than critical value, NH reject or accept e.g. Spearman and chi-square = greater then restate accepted H)

Morals and non-human animals (3) - S - R -

1. Sentient beings = respond to pain and self-awareness (brain damaged people lack sentient but still need consent) 2. Speciesm = Singer "Utilitarian view" (ethically acceptable if can alleviate pain and suffering) but is no different to racial or gender discrimination 3. Animal rights = Reagan (1984) it's not acceptable under any circumstances even though the definition of "rights" is dependent on responsibilities in society and animals don't have any

What are the aims and hypothesis of a study (types of H) ?

A = What is intended to be studied H = formal statement of expectations, either directional (one-tailed) or non-directional (two-tailed) which is operationalised to make the variables easily tested

What are lab experiments?

A controlled environment, with high internal validity as the extraneous variables are controlled but low in external validity (mundane realism) and may also be skewed by investigator effects and demand characteristics

What is a pilot study?

A small-scale trial run to test aspects of a design with the view of making improvements in the main study

What are demand characteristics?

A subtle cue that makes participants aware of what the experimenter expects to find or how participants are expected to behave

What is replicability? (Science) A01

Ability to record procedures in a way so they can be repeated and verify the original results

Inductive vs deductive research in qualitiative analysis

Aim to be inductive or bottom-up (categories emerge from data which may lead to New theories = emergent theories) Deductive (top-down) starts with preset categories of themes generated from previous studies so research aims to see if the data are consistent

Design your own study - method: ethics

Any ethical issues and how they will be dealt with

What is protection from harm and consequences for research?

Any negative physical or psychological effects with no greater risk than their everyday lives, should remain in the same state as when they start R = Not possible to estimate

What do ethical committees do and do they always work? (3) (Dealing with ethical issues)

Approve research before it begins by a cost-benefit approach weighing up the potential consequences with meaningful findings X = Judgements are subjective, costs aren't always apparent until after research is conducted and it's difficult to quantify the costs and benefits

What are the 3 meqsures of central tendency? (Descriptive statistics)

Averages 1. Mean (adding up all scores and dividing it by the number of scores = unrepresentative if there are extreme values NOT NOMINAL) 2. Median (middle value in an ordered list = not affected by extreme scores as not all values are reflected, NOT NOMINAL) 3. Mode (mosy common, only method appropriate when nominal)

What 2 techniques do observational studies include, and 1 advantage and disadvantage of them?

Behaviour categories record particular instances of behaviour and time (every 30 secs) and event (certain behaviour occurs) sampling Good = give rich picture of what people actually do over what they say they do Bad = Observer bias (influenced by their expectations, thus not objective = lack validity)

What is the Cyril Burt affair? (Validating new knowledge) - B - K - G- M - A01

Burt (1955) reported identical correlation throughout twin studies over 11 years which appeared to have suspicious consistency of correlation coefficients, Kamin (1977) called it 'invented data' Gillie (1976) couldn't find 2 assistants which confirmed underlying fraud and led to the research being publicly discredited Mackintosh (1995) called Burt 'astonishingly dishonest' and even shaped social policy, 11+exams judging whether a child will go to Grammar or secondary modern school based on their IQ

What is content analysis, cross-cultural research and meta-analysis?

CA = Observational study C-CR = Compare effects of different cultural practices on behaviour M-A = Results of many studies on the same topic are reviewed to find an overall conclusion

What is systematic sampling?

Choosing every nth person

What is theory construction? (Science) A01

Collection of general principles explaining observations and facts

What is confidentiality and anonymity and their consequeces for research?

Communication of personal info and the trust for it to be protected (data protection act = legal right for personal info to be protected with no connection able to made from results to the particpant) R = Not always possible as details may lead to identification such as case studies as they're recognisable by unique features

What is informed consent and consequences for researcher?

Comprehensive info on nature and purpose of the study and the role in which they will play to make an informed decision R = Can reduce meaningfulness as it may chage their behaviour

What is probability? (Inferential statistics)

Conclusions are based on the probability of a pattern being chance ot not

Designing your own study - method: design/materials (3)

Consider validity and reliability (generally O/Q/I) Experiments = lab, field or natural and which design? Q/I = Structured, semi or unstructured, open or closed A? O = Naturalistic or controlled, overt, covert, participant or non-participant?

What is objectivity? (Science) A01

Controlled conditions where a cause-and-effect relationship can be established e.g. the IV effects the DV

Interative oricess in qualitative analysis

Data are gone through repeatedly to impose some order to represent the participants perspective, a bottom-up approach ensures order emerges from the data and not preconceptions

What are case studies and 1 advantage with 2 disadvantages?

Detailed studies of an individual, event or institution, often longitudinal Good = Enables researchers to study complex interactions of many factors Bad = Difficult to generalise the results as the case study holds unique characteristics and research often includes recollection of past events which is unreliable being retrospective, they may recall memories incorrectly

What is empirical evidence? (Science) A01

Direct observation/experiment

What are 2 advantages and 2 disadvantages of self-report studies?

E.g. questionnaires or interviews Good = structured so more easily repeated and open questions allow rich insight with qualitative data Bad = Social-desirability bias (aim to present themselves in a 'good light') and difficult to analyse

What is opportunity sampling and is it useful?

Easily available, however only represent a small part of the target population e.g. if you choose participants from town on a Monday morning, research would likely lack professional people as they'd be at work

Are self-report studies reliable? (2)

External = Find the same results over different occasions Internal = Consistent within itself e.g. Are all the questions measuring the same thing?

Are experiments valid?

Extraneous variables cause experiments to lack internal validity, lab experiments control the EV's but field experiments do not, thus lack internal validity

Why would we use non-human animals in research? (2) - H -

For greater control and objectivity or when we can't use humans such as Harlow (1959)

What is right to withdraw and consequences for research?

Freely decide to not continue and withdraw data at the end if unhappy R = If paid or rewarded may feel as though they can't and it can lead to bias as those who withdraw may be more confident of intelligent this leaving a bias sample behind

Preserving the status undermining validating new knowledge - H - A02

Go with existing theory not dissenting or unconventional one's as Horton (2000) the former editor of medical journal, The lancet stated PR is discovering acceptability of new findings, not validating them

Psychology lacks objectivity and control (Science) - H - A02

Heisenberg (1927) can't measure subatomic particle without altering its behaviour in doing the measurement, uncertainty principle = presence of experimenter changes the b of what is being observed

Are observational studies valid? (3)

High external validity as they observe natural behaviours however lack internal validity if the coding system or behaviour checklist is flawed for example, and can also suffer observer bias which is when the observer's influence the results by their expectations thus meaning they do not remain objective

What is induction and deduction? (Scientific process) A01

I = reasoning from particular to general (observe natural phenomenon and create a general law) e.g. Newton's Law observed physical objects behaviour then created a law D = reasoning from general to particular (theory looks to confirm) e.g. Darwin's theory of evolution was a theory which he then observed animals in nature to confirm

What is internal and external validity?

I = what goes on inside a study, whether it tested what it intended to Ex/Ecological = how generalizable it is to other situations/populations

What is an independent and dependent variable?

IV = Varied to researcher effects DV = Dependent on the IV

What is correlational analysis, and 1 advantage and disadvantage?

Identifies relationship between co-variables Good = Large data sets are easily replicated Bad = Unknown, intervening variables may explain why co-variables are linked

What is stratified and quota sampling, are they useful? (2)

Identify strata and choose a predetermined number from each in proportion to their representation in the target population, for example if there are twice as many males in the target population there will be twice as many in the research S = random sampling whereas Q uses opportunity sampling It creates a representative sample as it's proportional however opportunity sampling can be biased

How do we draw conclusions? (Inferential statistics)

Inferenrial tests help us draw inferences (conclusions) through desriptive tests which provide a summary but not conclusion as we cant go beyond the sample, so they allow us to infer whether a pattern is likely or not due to chance, and if not then this is seen as significant

How do interviewers increase reliability?

Inter-interviewer reliability (2 interviewers find the same results)

French (2011) and Ritchie et al and publication bias (Validating new knowledge) A02

Journals love positive results thus leading to bias and misperception of true facts as they avoid publishing straight replications (although fundamental in validity) F = paranormal phenomena not considered for PR R.E.A = Journals as bas as newspapers seeking eye-catching stories

Are experiments reliable?

Lab experiments can be easily replicated

Designing your own study - introduction

Look at findings from last research, create aims and hypothesis, directional or non-directional related to past research

What is random sampling and is it useful? (2)

Lottery method or random number generator It creates an unbiased sample as all members of the target population have equal chance of being picked, however it could become biased if some refuse to take part, as these people would likely be similar

Wikipedia as a way of validating new knowledge (2) A02

More likely point out bias that would unchallenged in traditional reference work and permits instant revision when mistakes are spotted

What are field experiments?

More natural environment, with lower investigator effects as those being studied are often unaware however demand characteristics may affect the research as the operationalised IV may convey the hypothesis

Does psychology use the scientific method? - M - A02

Most psychs generate models that can be falsified and conduct well-controlled experiments to test these models BUT Miller (1983) states psychs who attempt to be scientists are only 'dressing up' taking on the tools of science such as quantified measurements but this doesn't make it a science Perhaps it's a pseudoscience which is dangerous because psychs can then claim their discoveries are fact

4 requirement of the UK Animals (scientific procedures) act (1986)

Must be done in a liscensed lab with a liscensed researcher which are granted if: 1. Have potential important results 2. Use the minimum number of animals 3. Research can't be done using non-animal methods 4. Discomfort or suffering is kept to a minimum e.g. given the appropriate painkillers or anaesthetics

What is a null and alternative hypothesis? (Inferential statistics)

N = (H0) states no difference or correlation A = (H1) If there is an effect then the NH is rejected and an AH is created

What is deception and consequences for research?

Not told true aims of a study or deliberately misled preventing informed consent R = Relatively harmless and compensated by debriefing

Qualitative research support psychology as a science A02

Qualitative methods are scientific as they aim to be valid e.g. interviews and observations which are then triangulated (compare findings w/ other methods to verify them and make them objective)

Summarising qualitative vs quantitative data (Qualitiative analysis)

Quantitative data is readily summarised with measures of central tendency and measures of dispersion and use of graphs but qualitative data needs to be summarised by identifying repeated themes

The scientific approach is reductionist and determinist A02

R = Complex phenomena reduced to simple variables to study casual relationships from, telling us little about 'real' behaviour yet without R it would be difficult to pick out patterns D = Search for causal relationships thus oversimplifying relationships but provides insight into important factors

Design your own study - method: procedures

REPLICABILITY Outline standardised instruction including informed consent Where, when, what materials Pilot study with group similar to the target population

How is reflexivity and validity shown in qualitative research

Reflexivity refers to a process that reflects researcher's values and thoughts Instead of trying to minimise the subjective bias qualitative research acknowledges the subjective nature as part of the research itself Validity may be demonstrated through triangulation (comparing results of different studies on the same topic using different methods)

What is falsifiability? (Inferential statistics)

Related to whether we should accept or reject a null hypothesis by seeking disproof e.g. Popper

Brooks (2010) and already published data undermining validating new knowledge A02

Remains in public view even if subsequently fraudulent, B = subsequently debunked research is continued to be used in debates in parliament, members of parliament have little critical understanding of the process of science, Brooks emphasises the need for increased vigilance by scientist's quality of PR

What are the 3 types of experimental designs?

Repeated measures = participant tested in all conditions Independent groups = Separate group for each IV Matched pairs = participants with matching key variables such as age and race are tested in different conditions to compare results

Anonymity and validating new knowledge A02

Researcher uses the veil of anonymity to settle old scores or bury rival research, compete for funds and grants, social relationships inevitably affect objectivity

BPS Code of ethics and conduct 2009 (4)

Respect (dignity and worth e.g informed consent and privacy) Competence Integrity (honest and accurate) Responsibility (protection from harm and debriefing)

What is Spearman's Rho test? (3)

Rho 1. Hypothesis states a correlation 2. Data is related 3. Ordinal or interval data

What is privacy and consequences for research?

Right to control into about themselves (who knows personal info) R = difficult e.g. observations don't want to alert they're being studied to avoid affecting their behaviour

Designing your own study - method: participants

Sampling techniques, independent groups design = how to assign to groups? Observational techniques include selection of partipants and how observations will be recorded e.g. time or event sampling

Scientific research is desirable supporting psychology as a science A02

Scientific research enables us to produce verifiable knowledge about behaviour as distinct from commonsense psychology which is used as proof of a theory

British medical journal undermining validating new knowledge - S - A02

Slow, expensive, profligate to academic time and prone to bias, it's almost useless at detecting fraud Smith (1999) Can't always find appropriate expert so poor research passed as reviewer didn't understand

What two methods are used to assess reliability?

Split-half method (performance on 2 halves of a test should show the same correlation in scores (internal R) Test-retest method (Repeated after a reasonable interval such as a week or a month and find the same results

What are the 2 measures of dispersion? (Descriptive statistics)

Spread of data 1. Range (difference between highest and lowest, affected by extreme values) 2. Standard deviation (spread of data around the mean, more precise as all values taken into account but some characteristics of the data such as extreme values are not expressed)

What is Wilcoxon test? (3)

T 1. Hypothesis states a difference between 2 sets 2. Pairs of scores e.g. matched pairs or repeated groups design 3. Ordinal or interval data

Designing your own study - method

Target population and research method related to aims and advantages/disadvantages Experiments/c orrelational analysis = identify IV/DV or co-variables and operstionalise

What are observed values? (Inferential statistics)

Test statistic e.g. T (wilcoxon) U (Mann-whitney) and Rho (Spearman's rho) = CALCULATED VALUE

What is hypothesis testing? (Science) A01

Testable expectation to prove a study is valid, if fail to support it then the hypothesis must be modified

What are critical values? (4) (Inferential statistics)

The observed value is compared to the critical value to see whether to accept or reject a null hypothesis, based on: 1. Degree of freedom (N) 2. One-tailed or two 3. Significance level (P < 0.05) 4. OV greater or less than CV for significance

What is Mann-Whitney test? (3)

U 1. Hypothesis states a difference between 2 sets 2. Independent-groups design 3. Ordinal or intern data

What are natural experiments?

Use existing IV, involve deliberate manipulation of variables so causal conclusions can't be drawn, participants are also not randomly allocated thus decreasing the validity however this is sometimes the only way to study a certain group or situation

What is debriefing and does it always work? (Dealing with ethical issues)

Used to assess the effects of procedures and offer counselling of needed, and if deception is involved they will also be offered the right to withdraw X = If research has caused distress, it can't be undone

What is snowball sampling and is it useful? (2)

Used when it's difficult to identify suitable people for a sample, such as those with eating disorders, once one is studied they can give insight into other's who are suitable for the research, although it's useful to find otherwise unidentifiable samples, bias can evolve as they tend to contact people within a limited section of the population

What is presumptive consent and does it always work? (Dealing with ethical issues)

Used when it's not possible to gain informed consent from the participants so they gain it from others with similar key variables X = Saying they wouldn't mind is different from actually experiencing it

Is volunteer sampling useful? (2)

Usually include similar groups of people thus result in volunteer-bias e.g. highly motivated individuals and people with extra time on their hands

What happens if research is published but discredited? (Validating new knowledge) A01

Usually publish retractions (e.g. if flawed) but some may still use faulty data not knowing it's been discredited

Design your own study - method: results

What statistics to use Q/I = analysis focuses on individual questions Qualitative = open questions

What are investigator effects?

When a researcher unintentionally affect the results due to the procedures of the research itself e.g. communicating their expectations

Laing (1960) Nomothetic vs idiographic goals of psychology undermining it as a science A02

When discussing causes of schizophrenia, inappropriate to view a person experiencing distress as a physical-chemical system gone wrong, treatment only succeeds if each person is treated as an individual case (idiographic approach) whereas science makes generalisations about people and finds similarities (nomothetic approach)

How do you get a type 2 error? (Inferenrial statistics)

When there is a difference but a significance level that is too stringent (1%) so we accept the null hypothesis although it's false = False negative

How do you get a type 1 error? (Inferenrial statistics)

When there is no difference, and a lenient level of significance (10%) leading to a rejection of a null hypothesis yet it's true = False positive

What is chance? (Inferential statistics)

Whether a pattern or effect is showing a real difference/correlation in the population, usually based on a probability of 0.05 (95%) but 0.01/0.001 when replicating a study or considering the effects of a new drug on health

What is Chi-squared test? (3)

X2 1. Hypothesis states a difference or association between 2 sets 2. Independent-groups design 3. Nominal data