Flashcards

Key Terms

Aims - the purpose of a research study.

Alternative hypothesis - any hypothesis (including an experimental hypothesis) that isn’t a null hypothesis.

Anxiety - a normal emotion similar to nervousness, worry, or apprehension, but if excessive it can interfere with everyday life and might then be judged an anxiety disorder.

Bar chart - like a histogram, a representation of frequency data, but the categories do not have to be continuous; used for nominal data.

Bias - some people have a greater or lesser chance of being selected than they should be, given their frequency in the population.

Bimodal - a distribution with two modes.

Case study - the detailed study of a single individual (or a few individuals); it typically involves qualitative data.

Class intervals - categories into which scores can be divided to summarise frequencies.

Closed questions - questions where the answer has to be selected from a limited set of options.

Concurrent validity - deciding whether a test measures what it is claimed to measure by relating the scores on the test to some other test or criterion.

Confidentiality - the requirement for ethical research that information provided by participants in research is not made available to other people.

Content analysis - a qualitative research method involving the analysis of behaviours or the written or spoken word into pre-set categories, a processing known as coding.

Control group - the group of participants who receive no treatment and act as a comparison to the experimental group to study any effects of the treatment.

Controlled observations - observations in which the researcher exercises control over some aspects of the environment in which the observations are made.

Correlation - an association that is found between two variables.

Correlation coefficient - a number that expresses the extent to which two variables are related or vary together.

Correlational analysis - testing a hypothesis using an association that is found between two variables.

Cost–benefit analysis - a comparison between the costs of something and the related benefits, in order to decide on a course of action.

Counterbalancing - used with repeated measures design to overcome the problems of practice and order effects, and involves ensuring that each condition is equally likely to be used first and second by participants.

Co-variables - the variables involved in a correlational study that may vary together (co-vary).

Covert observation - the participants are not aware they are being watched and their behaviour recorded by the observer.

Debriefing - attempts by the experimenter at the end of a study to provide detailed information for the participants about the study and to reduce any distress they might have felt.

Deception - in research ethics, deception refers to deliberately misleading participants, which was accepted in the past. Currently the view is that deception should be avoided wherever possible, as it could lead to psychological harm or a negative view of psychological research.

Demand characteristics - features of an experiment that help participants to work out what is expected of them, and lead them to behave in certain predictable ways.

Dependent variable (DV) - an aspect of the participant’s behaviour that is measured in the study.

Directional hypothesis - a prediction that there will be a difference or correlation between two variables and a statement of the direction of this difference.

Discourse analysis - a qualitative method involving the analysis of meanings expressed in various forms of language (e.g. speeches; writings). The emphasis is often on effects of social context on language use.

Double-blind technique - a procedure where neither the participant nor the investigator knows the condition to which the participant has been assigned; it reduces experimenter effects.

Ecological validity - the extent to which research findings can be generalised to naturally occurring situations.

Ethical guidelines - written codes of conduct and practice to guide and aid psychologists in planning and running research studies to an approved standard, and dealing with any issues that may arise.

Ethics - a set of moral principles used to guide human behaviour.

Evaluation apprehension - concern felt by research participants that their performance is being judged.

Event sampling - a technique for collecting data in an observational study. The observer focuses only on specified events.

Experimental group - the group receiving the experimental treatment.

Experimental hypothesis - the hypothesis written prior to conducting an experiment, which usually specifies the independent and dependent variables.

Experimental realism - the use of an artificial situation in which participants become so involved they are fooled into thinking the set-up is real rather than artificial.

Experimental treatment - the alteration of the independent variable.

Experimenter expectancy - the systematic effects that an experimenter’s expectations have on the performance of the participants.

External validity - the validity of an experiment outside the research situation itself; the extent to which the findings of a research study are applicable to other situations, especially “everyday” situations.

Extraneous variables - variables other than the independent variable that are not controlled and may affect the dependent variable (measure of behaviour) and so prevent us from interpreting our findings.

Face validity - this involves deciding whether a test measures what it is supposed to measure by considering the items within it.

Field experiment - a study in which the experimental method is used in a more naturalistic situation.

Generalisation - the extent to which the findings obtained from a given sample apply to a much larger population.

Histogram - a graph in which the frequencies of scores in each category are represented by a vertical column; data on the y-axis must be continuous with a true zero.

Hypothesis - a statement of what you believe to be true.

Independent groups design - a research design in which each participant is in one condition only. Each separate group of participants experiences different levels of the IV. Sometimes referred to as an unrelated or between-subjects design.

Independent variable (IV) - some aspect of the research situation that is manipulated by the researcher in order to observe whether a change occurs in another variable.

Individual differences - the characteristics that vary from one individual to another; intelligence and personality are major ways in which individuals differ.

Informed consent - relates to an ethical guideline which advises that participants should understand what they are agreeing to take part in. They should be aware of what the research involves, and their own part in this.

Intelligence quotient (IQ) - a measure of general intelligence with a population mean of 100 and a standard deviation of approximately 15.

Internal validity - the validity of an experiment in terms of the context in which it is carried out; concerns events within the experiment.

Inter-observer reliability - the extent to which there is agreement between the behavioural ratings of two observers.

Interview - a verbal research method in which the participant answers a series of questions.

Interviewer bias - the effects of an interviewer’s expectations on the responses made by an interviewee.

Investigator effects - the effects of an investigator’s expectations on participants’ responses.

Laboratory experiment - an experiment conducted in a laboratory

setting or other contrived setting away from the participants’ normal environments. The experimenter manipulates some aspect of the environment to observe its effects on the participants’ behaviour.

Longitudinal - over an extended period of time, especially with reference to studies.

Matched pairs design - a research design that matches participants on a one-to-one basis rather than as a whole group.

Mean - an average worked out by dividing the total of participants’ scores by the number of participants.

Measures of central tendency - ways of representing the mid-point of a set of data such as the mean, median, and mode.

Measures of dispersion - ways of expressing the spread of the data such as the range or standard deviation.

Median - the middle score out of all the participants’ scores.

Meta-analysis - a form of statistical analysis based on combining the findings from numerous studies on a given issue.

Mode - the most frequently occurring score among participants’ scores in a given condition.

Mundane realism - the use of an artificial situation that closely resembles a natural situation.

Natural experiment - a type of experiment where use is made of some naturally occurring variable(s).

Naturalistic observation - unobtrusive observations from a study conducted in a natural setting.

Negative correlation - as one co-variable increases the other decreases; they still vary in a constant relationship.

Negatively skewed distribution - most of the scores are to the left of the peak frequency and the mode is higher than the mean.

Non-directional hypothesis - a prediction that there will be a difference between two variables but the direction of that difference is not predicted.

Non-parametric tests - statistical tests that don’t involve the requirements of parametric tests.

Non-participant observation - observations in natural situations where the observer does not interact directly with the participants.

Normal distribution - a bell-shaped distribution in which most of the scores are close to the mean. This characteristic shape is produced when measuring many psychological and biological variables, such as IQ and height.

Null hypothesis - this hypothesis states that there will be no differences between the two conditions being compared (other than chance differences).

Observational techniques - those research techniques that involve observing behaviour, covertly or openly or as a participant in the activity.

Observer bias - distorted interpretations of participants’ behaviour; often based on the observer’s expectations.

Open questions - questions where the respondent can answer freely without being constrained by a predetermined set of options.

Operationalisation - defining all variables in such a way that it is easy to measure them.

Opportunity sampling - participants are selected because they are available, not because they are representative of a population.

Overt observation - the participants are aware they are being watched and their behaviour recorded by the observer.

Parametric tests - statistical tests that require interval or ratio data and normally distributed data.

Participant observation - observations in natural situations where the observer interacts directly with the participants.

Peer review - the process of evaluating the quality of someone’s research by other researchers; this is typically done to decide whether the research should be published in a professional journal.

Percentage - number or ratio expressed as a fraction of 100.

Pilot study - a smaller, preliminary study that makes it possible to check out standardised procedures and general design before investing time and money in the major study.

Population - the total number of cases about which a specific statement can be made. This in itself may be unrepresentative.

Positive correlation - when two co-variables increase at the same time.

Positively skewed distribution - most of the scores are to the right of the peak frequency and the mode is lower than the mean.

Predictive validity - assessing whether a test is measuring what it is claimed to measure by seeing whether scores on the test predict some future performance (e.g. on a related test).

Presumptive consent - a substitute for voluntary informed consent, it is presumed that if one set of people regard an experimental procedure as acceptable this applies to all people, including the experimental participants whose consent has not been obtained.

Primary data - the data that are collected directly by the researcher during the course of a study to test his/her hypotheses.

Protection of participants from psychological harm - an ethical guideline saying that participants should be protected from psychological harm, such as distress, ridicule, or loss of self-esteem. Any risks involved in the research should be no greater than those in the participants’ own lives. Debriefing can be used to counter any concern over psychological harm.

Pygmalion effect - an effect in which individuals perform surprisingly well because others expect them to; it is a kind of self-fulfilling effect in which others’ expectations turn into reality.

Qualitative data - data in the form of categories (e.g. has fun watching movies; has fun watching TV) or more complex narrative accounts.

Quantitative data - data in the form of scores or numbers (e.g. on a scale running from 1 to 7).

Quasi-experiments - research that is similar to an experiment but certain key features are lacking, such as the direct manipulation of the independent variable by the experimenter and random allocation of participants to conditions.

Questionnaire - a survey requiring written answers.

Random allocation - placing participants in different experimental conditions using random methods to ensure no differences between the groups.

Random sampling - selecting participants on some random basis (e.g. picking numbers out of a hat). Every member of the population has an equal chance of being selected.

Randomisation - the allocation of participants to conditions on a random basis, i.e. totally unbiased distribution.

Range - the difference between the highest and lowest score in any condition.

Ratio - the relative size of two classes; it can be expressed as a proportion.

Raw scores - the data before they have been summarised in some way.

Reliability - the extent to which a test produces consistent findings.

Repeated measures design - a research design in which the same participants appear in both or all groups or conditions.

Replicability - the ability to repeat the methods used in a study and achieve the same findings.

Representative sample - the notion that the sample is representative of the whole population from which it is drawn.

Research - the process of gaining knowledge and understanding via either theory or empirical data collection.

Research hypothesis - a statement put forward at the beginning of a study stating what you expect to happen, generated by a theory.

Right to privacy - the requirement for ethical research that no participants are observed in situations that would be considered private.

Right to withdraw - the basic right of participants in a research study to stop their involvement at any point, and to withdraw their results if they wish to do so.

Role-playing experiments - studies in which participants are asked to imagine how they would behave in certain situations.

Sample - a part of a population selected such that it is considered to be representative of the population as a whole.

Scatter diagram - two-dimensional representation of all the participants’ scores in a correlational study.

Science - a branch of knowledge conducted on objective principles. It is both an activity and an organised body of knowledge.

Secondary data - data that are used by a researcher even though they were collected by someone else and with a different purpose in mind.

Self-report techniques - participants provide their own account of themselves, usually by means of questionnaires, surveys, or interviews.

Sensitive - in the context of statistics, “sensitive” means more precise, able to reflect small differences or changes.

Significant figures - the number of digits in a reported finding that carry meaning.

Single-blind technique - a procedure in which participants are not informed which condition they are in.

Skewed distributions - distributions that differ substantially in shape from the normal distribution.

Social desirability bias - the tendency to describe oneself in a more positive and socially desirable way than is actually the case.

Standard deviation - a measure of the spread of the scores around the mean. It takes account of every measurement and can be used to estimate population values.

Standardisation - the process of ensuring that the procedures, treatment of the participants, and the wording of instructions are all the same for all participants.

Standardised tests - tests on which an individual’s score can be evaluated against those of a large representative sample.

Statistical significance - the level at which the decision is made to reject the null hypothesis in favour of the experimental hypothesis.

Stratified sampling - this involves selecting a sample from several layers or strata in proportion to the size of the layer.

Structured interview - all interviewees are asked the same questions in the same order and the interviewee is required

to produce certain types of

answers (e.g. short vs detailed).

Systematic sampling - a modified version of random sampling in which the participants are selected in a quasi-random way (e.g. every 100th name from a population list).

Test–retest - a technique used to establish reliability, by giving the same test to participants on two separate occasions to see if their scores remain relatively similar.

Theory - a general explanation of a set of findings. It is used to produce an experimental hypothesis.

Time sampling - a technique used in observational studies. Observations are only made during specified time periods (e.g. the first 10 minutes of each hour).

Undisclosed observation - an observational study where the participants have not been informed that it is taking place.

Unstructured interview - the questions asked by the interviewer and the order in which they are asked vary from interviewee to interviewee; in addition, the interviewee chooses how to answer each question.

Validity - the extent to which a test is measuring what it is intended to measure.

Variables - things that vary or change.

Variance - the extent of variation of the scores around the mean.

Volunteer bias - any systematic differences between volunteers and nonvolunteers.

Volunteer sampling - choosing research participants who have volunteered, e.g. by replying to an advertisement. Volunteer samples may not be representative of the general population, which means the research may not be generalisable.

Zero correlation - when there is no systematic relationship between two co-variables.

Weblinks

Research Methods

A programme about research methods from the “Discovering Psychology” series. The programme explores the scientific method and the ways in which data are collected and analysed both in the lab and in the field.

http://www.learner.org/series/discoveringpsychology/02/e02expand.html

https://www.youtube.com/watch?v=l49MzGZa_5MAn excellent narrated slideshow tutorial covering laboratory, field, and quasi experiments.

https://www.youtube.com/watch?v=xSsszDVuGcs A short video showing naturalistic observations being carried out in a retail context.

http://www.bps.org.uk/system/files/Public%20files/code_of_human_research_ethics_dec_2014_inf180_web.pdf The British Psychological Society’s Code of Human Research Ethics.

Scientific Processes

https://www.youtube.com/watch?v=k2MhMsLn9B0Monty Python’s illustration of how to do research.

https://www.youtube.com/watch?v=_7sSuhQ1_24 The Big Bang Theory’s illustration of how to do research.

https://www.youtube.com/watch?v=3TPqD9PZS_0 A video showing a couple of “scientists” demonstrating a carefully controlled experiment combining Diet Coke with Mentos. (23 minutes.)

https://www.youtube.com/watch?v=iwrcPTnBavs A short introduction to the scientific method, using the Diet Coke with Mentos experiment. (7 minutes.)

Discussion Points

Throughout AQA Psychology you may have noticed self-assessment questions in orange text in the margins. These are discussion points, designed to help you think about the key issues and research in the text, and to consider how you can apply psychology to real life. Although in many cases there is no correct answer to these questions, we have provided some example discussions that could arise from thinking about some of these points. They should also help you structure your thinking on the topic in question.

1. How can an experiment tell us more about “why” a behaviour has occurred than an observational study?

We place considerable emphasis on whether a given study’s findings can show cause and effect in psychology. Establishing cause and effect is the key to achieving the most certain kind of “why” answer to a psychological question and this is best shown using the experimental method.

The reason for this is because of the superior control that is involved compared to other methods. An observational study can provide lots of detailed information about natural behaviours in a given environment but the experimenter has no control over the situation and, more importantly, the variables. Cause and effect can never be achieved in this situation.

In contrast, an experimental situation can offer control of the variables. The experiment has a well- defined independent and dependent variable. The independent variable is manipulated by the experimenter, for example the temperature of a room, and the dependent varies throughout the experiment, for example recall of two-syllable words; this we measure.

Furthermore, the researcher will have established extraneous variables, which are not of direct interest to the experimental question but need to be controlled so that they don’t cause an effect on the dependent variable that might accidentally be attributed to the independent variable. Examples of extraneous variables needing to be controlled in this context might be those that can affect memory, such as intake of caffeine, alcohol, or medication.

When the experimenter collects the findings, he or she might notice that recall is better in one condition than another. The fact that the only thing that varied was the temperature (everything else was controlled) allows us to judge that the reason why recall was better was because of the independent variable. A conclusion can then be drawn that temperature affects recall and more specifically, recall is better in warmer temperatures. To find out why this is the case, an experimenter would need to go through the whole process again with both conditions enjoying warmer temperatures and varying other independent variables such as levels of arousal.

See also Eysenck’s AQA Psychology, page 331.

2. Imagine that you are going to conduct an observation of children in a playgroup with two other researchers. To what extent do you think that you will record the same behaviours? How might you cope with any disagreements?

Perception of any event can be understood using an information processing model. There is the input, throughput or storage, and output. At each stage there is room for corruption. Humans will unknowingly attend to individuals who are more attractive to them or behaviour that looks more interesting, and this will vary from person to person. When the brain reads what the eye sees it uses previous knowledge in the form of schemas, for example to interpret the behaviour. This again will vary between people. Finally, the breadth or depth of behaviour that someone chooses to record will, again, vary considerably. If left unmonitored, these individual differences will manifest in hugely unreliable records of behaviours.

To overcome this problem, researchers operationalise and categorise behaviours offering clear definitions to observers before the observation takes place. Provided that the type of data collected allows, the data sets can be correlated to see how similar they are. This provides the researcher with a measure of inter-rater reliability.

When disagreements occur, there are several options. Sometimes a half-way point between the differing viewpoints of two observers can be achieved after discussions. The ideal way of corroborating data, however, is through use of a video recording. A film of the scene is not always a perfect substitute; observing in real time and in person is more accurate and allows several different perspectives that an amateur video recording would not. However, a video recording can offer adjudication of discrepancies found in the data at a later date when memories are starting to fade.

See also Eysenck’s AQA Psychology, page 351.

3. If an experimenter used different wording in the instructions to different participants, how might this affect the results of the study?

The specific words that an experimenter uses when explaining instructions can have an enormous effect on a participant’s subsequent behaviour.

Different words will have different explicit meanings. For example, people may or may not understand the meaning of the word correctly or it may be area or subject specific. There are also implicit meanings behind the use of certain words. Participants may have associations with certain words. Beyond focusing on a single word, the use of one word instead of another may be significant in the semantic interpretation of the whole sentence.

Recall the work of Loftus on eyewitness testimony. She demonstrated that the change from “a” to “the” broken headlight resulted in a 10% increase in participants saying they had seen it even though it didn’t exist. Similarly the change of a verb from “contacted” to “smashed” had dramatic results on the estimated speed of a vehicle and whether there was broken glass at the scene.

Obviously, we are not asking people to recall information in standardised instructions, we are giving it to them. What this shows us however, is how influential very small changes in language use can be in the meaning of a message. In the same way as Loftus gave her participants information by using a leading question, instructions can do the same. A poor design with badly written instructions can at the very least make participants unsure of what is expected of them and can often give away the aim of the experiment for example, which can lead to participant reactivity and invalid results.

See also Eysenck’s AQA Psychology, page 347.

4. Why do you think that views about the kinds of research that are ethically acceptable have changed over the years?

The harm that has been done in certain studies over past years can be built into the BPS guidelines, in the same way as precedents are set in the law. So with time there are bound to be changes in guidelines as a result of various different questionable procedures that have been used by psychologists and then are consequently prohibited. With a larger bank of scientific literature to draw from, we can also judge the potential costs and benefits of research more accurately from previous studies. We learn from experience; however, it is not always that easy to predict the potential outcome of research for participants from previous studies.

The other main reason why the acceptability of different kinds of research has changed is changes in cultural norms and technology. The discipline of psychology has changed with the completion of the human genome project and the advent of more sophisticated brain-imaging techniques. This might have made certain kinds of research more acceptable as these kinds of methods have offered increasing levels of accuracy.

Changes in what is culturally acceptable in terms of society’s values will also affect what is acceptable to investigate. As a nation we are becoming more accepting of openness about sexuality in the media and it may be that psychological ethics committees decide to follow other sciences in their acceptance of the use of animals in research as changes in the acceptability of this prevail.

See also Eysenck’s AQA Psychology, page 356.

5. Watson and Rayner claimed that their experiment with Little Albert (1920) was ethical because the psychological harm inflicted was no greater than what he might experience in real life. Is this acceptable?

Little Albert, an 11-month-old orphan infant, was subjected to loud noises extreme enough to make him cry and ended up with a fear response to white rats, which generalised to most other white fluffy objects. Watson and Rayner produced in Little Albert a phobia to objects he would encounter frequently throughout his life. He was programmed to feel a fear response nearly every day of his life!

There are several factors worth considering here. Watson and Rayner had no way of predicting what level of psychological harm Little Albert was likely to experience in real life. It is also reasonable to assume that this is more extreme than most 11-month-old infants would experience, therefore their defence is flawed. Their defence is challenged further by the fact that they expressed an intention to “recondition” Albert to eliminate his fearful reactions. Their expressed intention to do this suggests that they knew that the harm done was severe enough to require reversal.

Little Albert was an orphan. This raises at least two questions. First, did Watson and Rayner believe he was already damaged in some way from being orphaned? Is it possible that they suspected that his future life in an orphanage would be damaging anyway, therefore there was less of a need to preserve his already tarnished mental health? Second, would this experiment have gone ahead if his parents were alive and available to give consent? On considering this, Watson and Rayner’s experiment becomes even less forgivable because this infant needed more protection than most, not further emotional damage. They affected Little Albert’s life chances and failed to provide remedial help for problems they caused through investigation; so no, it was not acceptable!

See also Eysenck’s AQA Psychology, page 362.

6. Psychology students often use other psychology students as the participants in their research. What problems are likely to arise, for example, in terms of evaluation apprehension and demand characteristics?

Even if psychology students don’t know the exact aim and hypothesis of the study in which they are participating, they are much more informed on the subject matters on which psychology focuses than non-psychology students. This means that demand characteristics are much more likely with a psychology student sample than one consisting of laypeople. This will occur because being informed leaves psychology students more likely to anticipate areas of interest to the investigator and more likely to read cues in the design and interpret parts of the process as meaningful. In this way, they may intentionally or unintentionally behave in ways they believe are helpful (or unhelpful) to the investigator. This, of course, threatens the internal validity of the study because this is not true behaviour, moreover the participant may have guessed wrongly.

Evaluation apprehension refers to the concern felt by research participants that their performance is being judged. This apprehension is often justified as this is often the case! Psychology students are aware of how investigators evaluate participants’ performances because they have probably done the same at some point. What this means for the experiment is that these participants will attempt to manage the impression they give of themselves based on their understanding of the research aims.

For example, psychology research might involve giving participants a mental arithmetic task whilst playing very loud, intrusive music and then asking them to complete a questionnaire about their thoughts and feelings. Psychology students may guess that the research is investigating performance under stress and how this might correlate to resistance to mental illness. This might lead these participants to present themselves in a socially desirable way, i.e. to perform as well as they can on the task whilst remaining calm and to not disclose any indicators of mental illness on the questionnaire. This is likely to be because they don’t want any psychological “weakness” to be identified for their own sake and to preserve the perception the investigator has of them.

See also Eysenck’s AQA Psychology, pages 344–345.