Research methods

Types of investigations:
Click the card to flip 👆
1 / 49
Terms in this set (49)
Extraneous variables - Variables other than the IV that might affect the DV (e.g. individual differences of participants).

1. Participant variables: e.g. age/intelligence

2. Situational variable: the surrounding environment e.g. temperature/noise levels

3. Experimenter variables: concern changes in the personality, appearance and conduct of the researcher.

Confounding variables - Uncontrolled extraneous variables that negatively affect results.
Features of a piece of research which allow participants to work out its aim and/or hypothesis, participants may change their behaviour.

- Participants may try to act socially desirable

- Try to guess the purpose & try to ruin the experiment (screw you effect)

- May act unnatural due to nervousness or fear of evaluation.
Participant observation: Involves observers becoming actively involved in the situation being studied to gain a more 'hands-on' perspective (e.g. Zimbardo's prison study).

Non-participant observation: Involves researchers not becoming actively involved in the behaviour being studied (e.g. Ainsworth's 'Strange Situation' study).

Overt: Where participants are aware they are being observed (e.g. Zimbardo's prison study)

Covert: Where participants remain unaware of being observed ('Strange Situation' study).

Unstructured observation - Records all behaviour which can be seen, this is difficult without the use of recording equipment, it also requires training & can be difficult to analyse because it provides a lot of detailed information (qualitative data).

Structured observation - Designing a type of coding scheme to record behaviour; these generally provide quantitative date (numerical). An example of this is behavioural categories, this is dividing target behaviours into subsets of behaviours through the use of coding systems.
The difference between experiments and correlations:An experiment isolates & manipulates the IV to observe its effect on the DV while controlling the environment by limiting extraneous variables, experiments establish cause and effect, correlations do not. In comparison, correlations identify variables and looks for a relationship between them.What is a content analysis?This is studying an individual using secondary data. This is data previously existing which is used for a psychologist's purpose which wasn't why that data was made. With this data, a researcher may: Coding: The researcher then 'codes' this information in quantitative data. Therefore, this is objective and in a measureable form. Thematic Analysis: This is a qualitative approach to analysis that involves ideas within the data. This is subjective and overall, more detailed.Case studies:In-depth investigations of a single person or small group. Strengths - Case studies create opportunities for a rich yield of data, and the depth of analysis can in turn bring high levels of validity (i.e. providing an accurate and exhaustive measure of what the study is hoping to measure). - Studying abnormal psychology can give insight into how something works when it is functioning correctly, such as brain damage on memory (e.g. the case study of patient KF, whose short-term memory was impaired following a motorcycle accident but left his long-term memory intact, suggesting there might be separate physical stores in the brain for short and long-term memory). - The detail collected on a single case may lead to interesting findings that conflict with current theories, and stimulate new paths for research. Weaknesses - There is little control over a number of variables involved in a case study, so it is difficult to confidently establish any causal relationships between variables. - Case studies are unusual by nature, so will have poor reliability as replicating them exactly will be unlikely. - Due to the small sample size, it is unlikely that findings from a case study alone can be generalised to a whole population. - The case study's researcher may become so involved with the study that they exhibit bias in their interpretation and presentation of the data, making it challenging to distinguish what is truly objective/factual.Hypothesis'sThe Null hypothesis - predicts that the IV will not affect the DV 'There will be no significant difference in X as a result of Y' The alternative/experimental hypothesis - Predicts that differences in the DV will be beyond the boundaries of chance (X will significantly affect Y - general hypotheses).Two types of hypothesis:Directional 'one-tailed' hypothesis - Predicts the direction of the results (there will be a significant reduction/increases in X as a result of Y). Non-directional 'two-tailed' hypothesis - Predicts that there will be a difference but does not predict the direction of the results.How can you construct a hypothesis?Refer to IV and DV Operationalise both Use the word 'significant'Types of sampling:Random Opportunity Volunteer Systematic sampling (every nth person) Stratefied (small-scale representation of a populationWhat is a Pilot Study?Are small-scale practice investigations, carried out prior to research to identify potential problems with the design, method or analysis, so they can be fixed. Participants may also suggest appropriate changes, for example, if participants admit that they guessed the purpose of the study and acted accordingly (demand characteristics), changes could be made to avoid this. Why do we need pilot studies? To ask participants about their experience of taking part & finding out what changes can be made to save time & money. To check if the IV is manipulated correctly. To check if you have selected the best method to measure the DV. To check whether it is appropriate. To check the participants understand instructions. The findings will be irrelevant, just the procedure needs to be checked.What experimental designs are there?Independent groups design Repeated measures design Matches pairs designWhat observational designs are there?Event sampling: Counting the number of times a behaviour occurs in a target individual or individuals Time sampling: Counting behaviour in a set time frame, for example recording what behaviour is being exhibited every 30 secondsWhat is Inter-observer reliability?Where observers consistently code behaviour in the same way. This must be established before an observation begins and it is easier to achieve if behavioural categories are clearly defined & do not overlap with each other (this will reduce bias).What is counterbalancing?Technique used to deal with order effects when using a repeated measures design. With counterbalancing, the participant sample is divided in half, with one half completing the two conditions in one order and the other half completing the conditions in the reverse order. E.g., the first 10 participants would complete condition A followed by condition B, and the remaining 10 participants would complete condition B and then A. Any order effects should be balanced out by this technique.Ethical issues with studies?Informed consent: signing a consent form should be completed before an experiment. Deception: When participants are misled or wrongly informed. A debrief: Must occur at the end of every study, tell them what they were investigating & why, must be told if they were deceived & if so why. They must be told they can ask questions & these need to be answered honestly & as fully as possible. Protection of participants: Investigators have the right to protect participants. Confidentiality: participants & their data must be kept anonymous unless they give their full consent. Right to withdraw: Participants must be aware they have the right to withdraw from the start right through to the debrief/end of the study. Incentives: There should be no incentives to take part as they may feel forced to do so. CRPIIDDWhat is a peer review?Peer review is papers before publication are subjected to independent scrutiny by other psychologists. It is the assessment of scientific work by others who are experts in this field It checks validity, filters for bad research, asses' the quality of design and methods, judges the importance of research & its originality (not been done before), the paper is recommended to be published or reviewed. It's 3 main purposes? Allocation of funding research (ensures money is only spent on worthwhile research) Publication of research in journals (prevents faulty or inaccurate data entering the public domain) To suggest amendments or improvementsImplications of ethical research:People will spend less time in institutions where tax pays pay for this if cures and treatments are found. When conducting research psychologists need to be aware of the ethical guidelines and not exploit people. Research can help the population.What is reliability and how to measure it?The degree to which the results can be dependent on being accurate. Test-retest - Giving a test at two different occasions in hope both results are the same. Inter-observer - Two experimenters observing the same test, in hope they get the same results.How can you improve self-report reliability?More specific questions, like closed questionsHow can you improve experimental reliability?Standardisation, and controlled environment (control of extraneous variables).How can you improve content analysis / observational reliability? -Making clear categories and chose quantitative over qualitative data as it is more specific.What is validity and the various types?Ecological validity - Weather the test would test the same in real life scenarios. Face Validity - Assesses if a test appears to measure what it is designed to measure. This can be assessed by an expert to see if something has face validity. Concurrent validity - Where a newer test is expected to get the same results as the older one. Can be assessed by comparing 2 tests to each other. Temporal validity - The extent to which research findings can be generalised to other historical times.How can you improve validity?Experiments - Have a control group to compare with (Lombroso should've done this). Questionnaire / self-report - Make the questionnaire anonymous.What are the steps when conducting an empirical method?1. Theory construction 2. Hypothesis generation 3. Empirical testing 4. Significance testing 5. Accepting or rejecting hypothesis 6. Replication research 7. Altering theoryWhat is falsification?The act of disproving a proposition or hypothesis.Paradigms and paradigm shift?An important change in the basic concepts and experimental practices of a scientific discipline.What goes into a report?Abstract - Concise summery of the investigation and findings. Introduction - Other similar studies, aim and hypothesis. Method - Participants, resources, procedure, variables. Results - No raw data, graphs and analysis. Measures of central tendency. Discussion - Limitations and what could've been made better. Conclusion of investigation. Further research which could be conducted. Referencing - Things which have been read I order to conduct the study. Appendices and raw data. (AIMRDM)How do you reference a book?Second name, dot, first letter of first name, Date, Name of book (in italics), Publisher.What types of data can there be?Qualitative data: Looks at the quality of written data, harder to analyse, not numerical, usually about attitudes, beliefs & behaviours, often has a greater real world meaning. Quantitative data: data in numerical form, can be put into categories or in rank order or measured in units of measurements. Primary data: Data given to the researcher by the participants themselves e.g. field experiments, original data, questionnaires, interviews, and observations. Secondary data: Data collected by someone other than the person conducting the research.What is a meta-analysis?Data gathered from a large number of studies, perform a statistical analysis of the combined data. Could be subject to researcher bias, may leave out research with negative or non-significant results.Measures of central tendency:Methods of estimating mid-point scores in sets of data. Mean: Average (add all numbers up and divide by how many things there are). Median: The middle number (order the numbers from highest to lowest and select the middle number). Mode: The most common occurring number.Measures of dispersion:Measurements of the spread of scored within a set of data. Range: Difference between the highest & lowest number. Standard deviation: The average distance of each score from the mean (the variation of data) High standard deviation = the greater spread of data, not all participants were affected by the IV in the same way. Lower standard deviation = data points tend to be very close to the mean (participants responded in a similar way). 1. Work out the Mean (the simple average of the numbers) 2. Then for each number: subtract the Mean and square the result. 3. Then work out the mean of those squared differences. 4. Take the square root of that and we are done.What to include in a graph?A title X-axis & Y-axis title Key (if necessary) Only draw a line of best fit if it specifically says toWhat types of graphs are there?Scatter graphs: Are only used for correlations. Bar chart: Used for discrete data, IV must be on the X axis & DV on the Y axis & do not put bars touching each other Histograms: Columns should be touching each other because data is usually collected over a period of time(At continuous intervals). Line graph: Join up the dots, often used to show how something changes overtimeWhat types of distributions are there?Normal distribution: Data with an even distribution of scores either side of the mean. Skewed distribution: Data that does not have an even distribution of scores either side of the mean. - Positive skew: Will contain more low scores than high scores, and will be distributed by the high scores. - Negative skew: Will contain more high than low scores, the skew being produced by outlying low scores.What levels of measurement are there?Nominal: States the number of people / things falling into each category. (mode- most common number) Ordinal: Scores are on a numerical scale and can be put into order from the lowest to the highest. This scale can be subjective (10 to one person may be different to another person). (Median - middle number) Interval: Allows things to be put in order, this time we can be certain that the gaps between numbers are the same. (Time, height, weight, number, heart rate). (Mean - Calculate the average number).What is a statistical test?The Sign Test To see which hypothesis should be accepted. If the difference between two groups is big enough to say that the results were not caused by chance How do you calculate the sign test? 1. Subtract 1st set from 2nd set (put a + or a -) 2. Cross out any with no change 3. Add up the symbols + vs - (ignore those with no change) 4. Find n (number of participants - those crossed out) 5. Find S (Smallest of the + or - totals)What is a Type 1 and 2 error?Type 1 (False positive) = Incorrectly rejecting the null hypothesis, where in fact the results were due to chance. This error can be reduced if you change the P value to a lower value. Type 2 (False negative) = Incorrectly accepting the null hypothesis, when actually the hypothesis was true. You will be more likely to make this error if you have a two-tiled hypothesis.Related or unrelated?Independent groups is unrelated Repeated measures is related Matched pairs is relatedTests for a difference:R - NOMINAL - SIGN U - NOMINAL - CHI SQUARED R - ORDINAL - WILL COXAN U- ORDINAL - MANN-WHITNEY R - INTERVAL - RELATED T U- INTERVAL - UNRELATED T (Sucking cock will make rory unhappy)Test for a correlationNominal - Chi squared Ordinal - Spearmans Interval - Pearson (Charlie snaked pete)