Please enable JavaScript.
Coggle requires JavaScript to display documents.
A2 Research Methods, = Determines if a significant difference/correlation…
-
= Determines if a significant difference/correlation exists in an investigation than by chance & whether it accepts/rejects null hypothesis
- Look for aim of investigation, if correlation or difference
- Difference describes directional/non-directional/null hypothesis
- Correlation describes associate +ve/-ve
- If correlation IGNORE this step
- Could be independent groups, unrelated as pp's in each condition are different
- Or repeated measures & matched pairs (pp's aren't same but have been matched), so both are related
- Quantitative data split into 3 types:
- Nominal data = In categories e.g. nbr of boys & girls in class & is discrete 1 item appears in only 1 category e.g. student votes for fav food in one section
- Ordinal data = Ordered e.g. on scale 1-10, has no equal intervals between each unit e.g. saying one who rated psych 8 likes it x2 than one rating 4, lacks precision as it's subjective to not measuring physical aspects like height/reaction time and is found in e.g. questionnaires, unsafe data isn't used in ST & converted into ranks
- Interval data = Based on numerical scales with equal units, most precise e.g. using stopwatch measures time/ thermometer measures temp
= Consistency of a study's findings, usually when replicable
E.g. a ruler is always consistent, measurement only changes if the object changes/ IQ one day would be same another day unless a person becomes more intelligent
- Same test/questionnaire (Interviews too) given to same person on diff occasion, to show same results can be obtained = reliable
- Time between each T/Q shouldn't be too long/short, scores then are correlated & if +ve the method is reliable
- 2 observers must carry out observations, pilot study is done 1st to see if they agree on the behavioural categories used
- Observers then watch same event (by event/time sampling), but record separately
- Their data is then compared & correlated to measure the reliability, greater than +.80 = reliable
- Same method for content analysis & interviews too
- When comparing test-retest data the correlation must exceed +.80
- But the same person may answers open questions differently each time, so closed questions can solve this
- Using same interviewer or structured interviews allows reliability in interviews
- Lab experiments are reliable as extraneous variables are controlled allowing replication, yet slightly diff conditions pp's tested in could affect this
- Behavioural categories must be operationalised & non-overlapping, if not different observers record diff behaviours affecting reliability e.g. older adults --> adults aged 60+
- Data from the same T/Q are split and compared, if similar it shows reliability
= The extent to which a test measures what it intends to measure (internal) & whether it can generalise beyond the research setting (external) e.g. a broken scale is reliable as it gives the same weight each time, but not valid as it's giving a higher weight than it actually is
- IV to do with experiments effects when independent variable is manipulated, but demand characteristics affect this as pp's may behave in an expected way.
- EV links to factors out of a study affecting it e.g. type of people/era
- Ecological validity = Whether findings can be generalised to everyday life from one setting to another e.g. lacking EV due to testing memory by getting pp's to learn list of words (EV due to memory MR due to word list)
- Temporal validity = Whether findings can be generalised to other historical times/eras e.g. an old studies results may not be same today
- Face validity = Whether a research measures what it appears to measure e.g. does an IQ test actually measure intelligence
- Concurrent validity = Whether a psychological measure relates to an existing similar measure e.g. IQ test results may be compared to another type of IQ test, if correlate a high CV exceeds +.80
- Lie scales used to control respondents response, avoiding social desire ability bias & data kept anonymous = producing valid data
- Covert observations tend to have high ecological validity as observed behaviour is natural & non-overlapping behavioural categories allows a high validity too
- Use of control group shows any effect on DV due to change in IV
- Standardisation used to reduce participant reactivity (pp's acting unatural) & investigator effects
- Single & double blind procedures control demand characteristics & investigators effects increasing validity
- Qualitative Methods (interviews & case studies) have high ecological validity than quantitative, more detail shows a pp's reality
- Yet, researcher would further develop by interpretive validity of interpretation of data, so researcher matches pp view & triangulation involves many sources develops this too = valid
- Abstract = Key details of research project e.g. aims, methods, results & conclusion which a psychologists reads to decide if they want to further examine the investigation
- Introduction = Review of the investigation & any concepts related to the study, includes aims and hypothesis
- Method = Describes what the researcher did so others can replicate: Type of design (experiment/observation), sample/pp's involved, assessment materials used, procedure of study (brief/debriefing & instructions) and ethics
- Results = Findings of study: statistics of graphs/charts/central tendency, inferential statistics of statistical test/which hypothesis rejected & analysis of qualitative data
- Discussion = Summarising findings verbally based on the intros evidence & limitations are discussed based on method and how to improve, real life implications considered too
- Referencing = Book reference: author(s),date, title of book (in italics), place of publication, publisher
- Journal reference: author(s), date, name of journal (italics), volume & issue nbr (italics) & pg nbr(s)
- Hypothesis written 1st of investigation & statistical test determines whether to accept/reject null & vice versa for alt hypothesis
- Null H has no difference, Alternative (directional/non-directional)
- Significance level = If a correlation/difference exists, if significant null H can be rejected & accept alt
- SL value is 0.05, if hypothesis is accepted less than 5% due to chance than 95% due to change in IV affecting DV
- 0.01 more stringent value that can be used to be more confident results are NOT due to chance e.g. for drug testing & ensures it's more significant
- Calculated value = Result calculated from a statistical test
- Critical value = value that allows us to accept/reject null H & vice versa for alt H
- To check significance in a test calculated value compared to critical
- To find critical: check if hypothesis is one/two tailed, nbr of pp's (N value)/ or degrees of freedom df & level of significance 0.05/ more stringent level 0.01
- Can be a 5% possibility that a wrong hypothesis is accepted
- Type 1 error = When a null H rejected & alt is accepted (false positive), when really the test is non-significant (more likely when significance is too high e.g. 0.1 than 0.05)
- Type 2 error = When null H accepted & alt rejected (false negative), when really it's significant (more likely when 0.01 used)
- Replicability = Whether a scientific experiment can be repeated by other researchers, allowing generalisability & gives validity of findings (To occur psychologists must report their investigations in detail)
- Falsifiability = When a theory cant be said scientific unless the idea of it being proved is not true, so strongest theory would be one that hasn't been proved false yet. (Popper)
- Paradigm = Shared assumptions & methods in scientific discipline, psychology lacks universal acceptance of paradigm as approaches conflict each other
- Paradigm shift = Change in a dominant theory within a scientific discipline, too much evidence against theory
- Objectivity = When bias is minimised so research isn't influenced, where investigator effects are controlled e.g. lab studies are objective
- Empirical method = Scientific approaches based on observation/experience, which develops objectivity (To claim scientific theory --> must be empirically tested)
- Theory = Set of laws to explain behaviours (forms through theory construction by gathering evidence by observation)
- Hypothesis testing = Makes a clear prediction based on theory so it can be scientifically tested
= Value between -1 and +1 that shows the strength & direction of a relationship
- +1 perfect +ve correlation & -1 perfect negative correlation
- Closer values are to +1/-1 stronger the relationship, but closer to 0 weaker relationship
- Coefficient = weak relationship = still can be statistically significant
= Analysis of an individual,group institution in detail
- E.g. On rare disorder, crimes or normally an old persons background
- To do so qualitative data gained by interviews/observations & experiment can be conducted to test their behaviour --> quantitative data (forms over a long time where data can be gathered from other too)
+ve Gives detailed data of unusual be, that wouldn't be obtained from an experiment
Provides normal functioning understanding e.g. HM found normal memory stores in STM & LTM
-ve Findings for case studies lack generalisation due to small samples, researcher may have bias to what makes it on it & data from people may not be accurate affecting validity
= Indirectly observes behaviour of people through their communications of e.g. Conversation, text or TV show to draw a conclusion of that person
- 1) Coding when the communication is analysed into categories as nbr of words/phrases = quantitative data e.g. Nbr of times person is called mad/crazy in newspaper
- 2) Thematic analysis qualitatively identifies explicit/implicit ideas (clear/unclear) once themes have developed from data e.g. Newspaper says mentally ill are threat to children's wellbeing --> categorises into stereotype or treatment of them. Then researcher collects new data to test validity of same themes & categories = write report
+ve Content analysis avoids ethical issues of research as data used is already in public, so easy to gain permission
Flexible too as it produces qualitative & quantitative data
-ve As people are indirectly studied researcher may form opinions that shouldn't be formed
Lacks objectivity as analyst produces bias on behaviours findings