Psychological Research And Scientific Method

Scientific method Major features of science - Empiricism - Control - Objectivity - Replicability - Theory construction

Scientific process Induction- reasoning from particular to general. Deduction- reasoning from general to particular.

Commentary- Is Psychology scientific? Scientific research is desirable. Psychology shares the goals of science. Kuhn- no single paradigm. Lack of objectivity and control leads to experimenter bias and demand characteristics. Commentary- Are Goals of science appropriate? Nomothetic versus idiographic. Qualitative research- triangulation- quantitative research.

Synoptic links Scientific approach is:

- Reductionist- reduces complex phenomena to simple ones; and - Determinist- searches for causal relationships.

Validating knowledge Peer Review Serves three main purposes: - Allocation of research funding. - Publication in scientific journals. - Research Assessment Exercise.

Research published on the internet requires new solutions.

Commentary May be an unachievable ideal. Anonymity allows honesty and objectivity. Publication bias favors positive results. May lead to preservation of status quo.

Conventions of scientific reporting Abstract- summary of study. Introduction/aim- literature review and research intentions. Method- procedures and design of study. Results- descri ptive and inferential statistics. Discussion- outcomes and implications of study. References.

Synoptic links Some changes in science are not logical changes but represent a shift in perspective (paradigm shift). Burt research- an example of scientific fraud.

Designing investigations Research methods Experiments IV varied to see effect on DV. Lab experiment- high on internal validity, low on external validity. Field experiment- more natural environment but more issues of control than lab. Natural experiment- uses naturally occurring Ivs but cannot conclude causality. Experimental designs- repeated measures, independent groups, matched pairs.

Observational studies Observing behaviour through behavioral categories. Sampling methods- time and event sampling. Open to subjective bias- observations affective by expectations.

Correlational analysis Concerned with relationship between two variables. Does not demonstrate causality. Other variables may influence any measured relationship.

Case studies Detailed study of individual, institution of event. Generally longitudinal, following individual or group over time. Allows study of complex interaction of many variables. Difficult to generalize from specific cases.

Design issues Reliability Experimental research- allows for replication of study. Observations- inter-observer reliability can be improved through training. Self-report- internal reliability (split-half) and external reliability (test- retest).

Validity Internal validity- does study test what it was intended to test? External validity- can results be generalized to other situations and people? Lab experiments not necessarily low in external validity. If low in mundane realism, reduces generalisability of findings. In observations, internal validity affected by observer bias. Self-report techniques, issues of face and concurrent validity.

Sampling techniques - Opportunity- most easily available participants. - Volunteer- e.g. through advert, but subject to bias. - Random- all members of target population must have equal chance of selection. - Stratified and quota- different subgroups within sample, leads to more representative sample. Snowball- researcher directed to other similar potential participants.

Ethics Ethical issues with humans Informed concent and deception. Harm- what constitutes too much?

Code of conduct Respect for worth and dignity of participants. Right to privacy, confidential, informed consent and right to withdraw. Intentional deception only acceptable in some circumstances. Competence- retaining high standards. Protection from harm and debriefing. Integrity- being honest and accurate in reporting. Use of ethical guidelines in conjunction with ethical committees. Socially sensitive research- potential social consequences for participants.

Ethical issues with non-humans Reason for animals use- offers opportunity for greater control and objectivity; can`t use humans; psychological similarities. Moral issues- sentience (experience pain and emotions). Specieism- form of discrimination against non-human species. Animal rights- Regan (1984), no animal research is acceptable. Do animals have rights if they have no responsibilities? Animal research subject to strict legislation (Animals Act; BPS guidelines). The 3Rs- Reduction, Replacement, Refinement.

Data analysis Probability and significance Probability= likelihood that a pattern of results could arise by chance. If probability extremely unlikely, then result is statisticsically significant. Inferential tests determine whether chance or real trend in data. Probability levels represent acceptable level of risk (e.g. P<0.05) of making a Type 1 error. More important research, more stringent significance level. Type 1 error= null hypothesis rejected when true. Type 2 error= null hypothesis accepted when false.

Inferential tests Different research designs require different tests. Different tests for different levels of measurement (nominal, ordinal, interval, ratio). Tests yield observed values, and then compared to critical values to determine significance level. One- tailed test= directional hypothesis. Two-tailed test= non-directional hypothesis.

Inferential tests Spearman`s RHO Used when: - Hypothesis predicts correlation between two variables. Each person is measured on both variables. Data is at least ordinal (i.e. not nominal).

Mann-Whitney U Used when: Hypothesis predicts difference between two sets of data. Independent groups groups design. Data at least ordinal (i.e. not nominal).

Wilcoxon T Used when: Hypothesis predicts difference between two sets of data. Related design (repeated measures or matched pairs). Data at least ordinal (i.e. not nominal).

Chi- Square Used when: Hypothesis predicts differences between two conditions or associated between two variables. Data is independent. Data in frequencies (nominal). Expected frequencies in each cell must not fall below 5.

Central tendency Indicates typical or `average score`. Mean, median, mode. Measure of dispersion Range- highest/lowest difference. Standard deviation- spread of data around mean. Precise measure but influence of extreme scores not taken into account.

Graphs Bar chart= illustration of frequency, height of bar represents frequency. Scattergram= illustration of correlation, suitable for correlational data. Indicates strength of correlation and direction (positive and negative).

Qualitative data Key Points Quantitative methods not relevant to `real life`. Qualitative represents worlds as seen by individual. Emphasizes collection of subjective information from participant. Data sets tend to be large. Qualitative data cannot be reduced to numbers. Can be examined in themes.

Methods of analysis Coding using top-down approach (thematic analysis)= codes represent ideas/themes from existing theory. Coding using bottom-up approach (grounded theory)= codes emerge from data. Behavioral categories used to summarize data. Reflexivity indicates attitudes and biases of researcher. Validity demonstrated by triangulation. Reliability checked by inter-rater reliability.

Quantitative vs qualitative Quantitative: Easy to analyze and produces neat conclusions. But: Oversimplifies reality and human experience. Qualitative: Represents true complexities of behaviour through rich detail of thoughts, feeling etc. But: More difficult to detect patterns and subject to bias of subjectivity.

Nominal= Data separate categories, grouping class into tall, medium and short Ordinal= Data are ordered in some way, for example lining up your classmates in order of height. The `difference` between each item is not the same. Interval= Data are measured using units of equal intervals, such as when counting correct answers of measuring your classmates` heights.