HP1100 Readings PDF

WM Chapter 1 Psychology and Science Ways of knowing about behaviour Non-empirical methods Empirical methods Authority Logic...

WM Chapter 1 Psychology and Science Ways of knowing about behaviour Non-empirical methods Empirical methods Authority Logic Intuition Science - we believe - Logic can tell - a way of knowing based on spontaneous, instinctive See below something because you that a processes rather than on logic or reasoning. some respected statement is false - Common sense is a kind of intuition because of its person told us it is because it draws dependence on informal methods true (e.g parents, an improper - It has the additional characteristic of emphasising the religious authorities, conclusion agreement of a person’s judgement with the shared attitudes government) - But a statement and experiences of a larger group of people. - Because these can be logically - standards of common sense differ from time to time and from authorities often valid and still not place to place according to the attitudes and experiences of the disagree among be true because it culture themselves, we are assumes -the only criterion common sense recognises for judging the inclined to reject something to be truth of a belief or practice is whether it works authority as a way of the case that is not - According to the common-sense method, no systematic knowing E.g it is raining attempt is made to see why a practice works, or to test its - Authorities often there is no game theoretical explanation to see if it is true. As long as a certain are wrong, even → valid but only practice works, that practice is maintained and the theory when they assert true if it in fact is behind it is considered true their beliefs most raining - We may speak of a scientific result as being counterintuitive; forcefully that is, it goes against our notions of common sense. In fact, we consider a scientific theory to be fruitful if it predicts something that we did not expect *there is no substitute for empirical evidence Science - The most important characteristic of science is that it is a way of obtaining knowledge based on objective observations //characteristics of science// 1. Empirical The scientific attitude is to rely on experience more than on authority, common sense, or even logic 2. Objective The most important characteristic of science is that it is a way of obtaining knowledge based on objective observations Objective observations are those made in such a way that any person having normal perception and being in the same place at the same time would arrive at the same observation when observations are objectively made and carefully reported, they serve as a sort of recipe for others to follow The opposite of objective observations are subjective observations→ observations that a person makes that another person is not required to accept as true 3. Self-correcting new evidence is constantly being discovered that contradicts previous knowledge Science is characterised by a willingness to let new evidence correct previous beliefs 4. Progressive Science moves forward toward truth, adding more and more information to what was previously known 5. Tentative science never claims to have the whole truth on any question because new information may make current knowledge obsolete at any time 6. Parsimonious the principle of parsimony holds that we should use the simplest explanation possible to account for a given phenomenon A good scientist will always prefer a simpler explanation to a more complex one, other things being equal  7. Concerned with theory The importance of theory to science can be illustrated by contrasting science and technology technology is far ahead of the science //relation between science and nonscience// Authority has a reduced role in science Although a power structure is by no means absent from science, authority plays a different role from the one it plays in other human activities Working assumptions of science 1. The reality of the world Philosophers call this assumption the doctrine of realism: the notion that the objects of scientific study in the world exist apart from their being perceived by us In general, scientists have little interest in philosophical debates about the reality of the world. They assume that the world is there, and they go about studying it as best they can They do avoid one variety of realism, which is known as common-sense realism, or naive realism. Common-sense realism is the philosophy of the person on the street who never wondered why coal looks black because anybody knows that coal is black. Common-sense realism says that things are just the way they seem: Coal looks black because it is black Although the scientist and the layperson both believe in the existence of a real world, the world that the scientist believes in is different from the one the layperson believes in. The layperson’s world may contain people who are lazy or hardworking, good or evil. The scientist’s world, by contrast, is more likely to consist of people who are influenced by stimuli, cognitions, and emotions 2. Rationality the world is understandable by way of logical thinking 3. Regularity we assume that the world follows the same laws at all times and in all places The reason regularity is so important for science is that it says that the laws of science are the same today as they were yesterday or a thousand years ago, and they will be the same tomorrow and a thousand years from now 4. Discoverability The scientist assumes that we can discover the way the world works without having a higher being or book reveal it to us 5. Causality assume that events do not just happen by themselves or for no reason Criterias for establishing cause-effect r/s: ○ temporal precedence: something that occurs prior to another thing ○ covariation of cause and effect: when the cause is introduced, the effect occurs probabilistic ○ co-variation: statistical association of a cause with an effect ○ elimination of alternative explanations: no explanation for an effect other than the purported cause is possible Goals of science Discovery of Regularities ○ Description describe the phenomena considered to be important for the science to deal with define events and entities such as stimuli and responses, cognitions and beliefs, or neuroses and psychoses have some agreement, though, on just what it is that we are going to study Description of phenomena is crucially important to a science because it defines the subject matter for which laws are sought and theories are developed creating the subject matter ○ Discovering laws law: a statement that certain events are regularly associated with each other in an orderly way It is not necessary to have a perfect relation between the two variables to have a law some laws are probabilistic; that is, there is regularity between two variables, but the regularity is not such that every time one variable is present, the other is too ○ Search for causes search for the causes of the events that we observe Development of theories ○ Theory a theory is a statement or set of statements explaining one or more laws,usually including one in direct concept needed to explain the relationship If the statements concern only a single relationship between variables, we are speaking of a law. However, sometimes a number of laws are tied together into a more general set of statements,which is called a theory. ○ Theories must be falsifiable  falsifiability: the property of a good theory that it is capable of disproof A good theory must be capable of being tested in an unambiguous way It must make a definite prediction that can be proven right or wrong Science is based on empirical evidence, hence its theories must be capable of being empirically tested A theory must not only be testable, it must also be capable of being proven wrong Role of theories ○ Organising knowledge and explaining laws Theory pulls together a collection of descriptions and some laws into a unified framework → relating individual events to laws and laws to theories constitutes scientific explanation The better the theory, the more events and laws it can explain ○ Predicting new laws Help suggests places to look for new laws ○ Guiding research Most researchers say that they work within a certain theoretical framework A good theory suggests new experiments and helps researchers choose alternative ways of performing them Hypothesis in science ○ hypothesis: a statement assumed to be true for the purpose of testing its validity ○ A hypothesis can be put in the form of an if-then statement: If A is true, then B should follow ○ A scientific hypothesis must be capable of empirical testing and, as a result, empirical confirmation or disconfirmation Defining theoretical concepts ○ Operational definition: a statement of the the precise meaning of a procedure or concept within an experiment – i.e. what we are measuring The principal misuse is taking a trivial definition of a concept and attempting to build a theory on it Another misuse of the concept of operational definition is considering every measure of a concept as independent of every other measure  Chapter 3 Ethics in Research American Psychological Association(APA) Ethics Code Ethical Principles of Psychologists and Code of Conduct Responsibility and Protection from Informed Consent & Coercion Debriefing and deception Scientific writing Harm Conflict is between (1) the - instructions given are - Experiments may Responsibility of commitment of the psychologist to understandable require participants be publishing only data that expanding our knowledge of - Participants are informed of naïve to the research have been legitimately behaviour and the potential benefit research procedures e.g. potential hypothesis obtained and with giving the research may have for society risks, incentive, right to withdraw at -Inform participants as appropriate credit to all and (2) the cost of the research to any time soon as possible about those who contributed to the participants - Willing participation nature of study, answer the project - Researchers who do not review - Participant should be in the right questions and remove ethical problems carefully are state of mind when giving consent misconceptions - No plagiarism → cite negligent toward society. → e.g not on drugs, need guardian - Ascertain whether the all relevant resources - Researcher who refrains from consent if underage, etc → The deception i.e. the - If significant errors are doing an important study because participant’s informed consent experimental discovered, take of an excessively tender should be documented in writing manipulation was reasonable steps to conscience is also failing to keep a - not coerced (compelling or successful correct them commitment to the same society influencing a decision to participate - Sometimes delay until - do not fabricate data → that supports behavioural research in a study) by researchers → the end of the entire Future researchers who with the hope that it will provide Issues of coercion become more study attempt to base their important social benefits serious if substantial sums of money - Participants must be experimental ideas on are offered for participation or if told the purpose and someone else’s false - competence→ pay attention to people are induced to participate expected results of the results will fail → many groups like minors or patients with with promises to “improve your experiment so that their years of futile work and mental disorders → design an relations with the opposite sex” or experience has as much many wasted research experiment suited to the group “gain valuable insights into your educational and personal dollars for other - not to overdo such that the personality.” value for them as scientists to find truth experiment put subjects at risk possible Others: - Federal government requires institutions that receive federal funds to establish an Institutional Review Board (IRB) Ensure that studies present as little risk as possible and have scientific merit approve virtually all research on human participants committees must have a minimum of five members, at least one of whom is not a scientist, and one member must be unaffiliated with the institution many proposals require a full review from the IRB - Singapore: Human Biomedical Research Act (HBRA) Fraud in Research Scientific data are creations that can easily be concocted→ This means that often we are utterly dependent on the scientist’s honesty for the truthfulness of the data Many thousands of research dollars can be wasted following up on a research result that doesn’t exist Likewise, there can be public health consequences, such as when Andrew Wakefield and colleagues published a paper that fraudulently claimed an association between autism and vaccines  ○ many research dollars have been wasted following up on the result and many children have gone unvaccinated, leaving them at risk for damage from measles, mumps, or rubella Ethics and Animal Experimentation APA Guidelines on Animal Experimentation: ○ Minimise discomfort, infection, illness, and pain of animals ○ Treated with humane consideration of their well-being in conjunction with research goals psychologists should have a reasonable expectation that the results of an experiment involving animals will yield results that increase scientific knowledge of people or the species involved in the research Psychologists should assume that stimuli that are painful to people are also painful to animals, so care should be taken to minimise the number of animals involved in the research or to consider non animal research alternatives Psychological research on animals should be carried out by trained personnel under the supervision of an institutional animal care and use committee which usually include a veterinarian Chapter 8 and 9 Non-experimental research The distinction between experimental research and nonexperimental research is based on the degree of control that the researcher has over the subjects and the conditions of the research manipulation: in an experiment, conditions or variables assigned or presented to a participant assignment: in an experiment, pairing a subject with a condition or variable, according to the experimenter’s plan observation: the record of a behaviour An experiment is a kind of investigation in which some variable is manipulated. The researcher has enough control over the situation to decide which participants receive which conditions at which times A second characteristic of nonexperimental research is that the data collection procedure often must forfeit some degree of control in return for obtaining the data e.g keep a questionnaire short to help gain the cooperation of subjects A third characteristic of some nonexperimental research, called qualitative research, concerns the questions that are typically asked by the research. The questions that are asked by qualitative research differ from those of experimental research. Qualitative research is much less interested in the cause and effect of behaviour than is research based on experimentation. Instead, qualitative research is interested in how individuals understand themselves and make meaning out of their lives Nonexperimental research is often called correlational research because it seeks causes of behaviour by looking for correlations among variables We often compute correlations among variables in the truest of experiments. What makes research correlational in the common usage is the inability to manipulate some variable independently In correlational research, relationships are studied among variables, none of which may be the actual cause of the other. For example, doctors have noticed that people who drink red wine have better health. This does not necessarily mean that red wine improves health, or that healthy people prefer to drink red wine. The two items are merely related; we do not know which one caused the other. Either one could actually be caused by some third variable correlation does not prove causation→ We demonstrate causation when we can decide which variable caused the other, and this is best done in an experiment. Nonexperimental research is often a first step in starting to answer theoretical questions by empirical methods. Experimental research frequently is done as a follow-up to previous nonexperimental observations  Varieties of nonexperimental research Observational research Archival research Case study Survey - involves recording - existing records - researcher - participants are requested to cooperate by responding to questions ongoing behaviour without are examined to investigates a attempting to influence it test hypotheses particular How a questionnaire is designed about the causes situation that - Frequently, researchers use existing questionnaires, rather than designing their own instruments → they avoid Naturalistic observation of behaviour has come to redesigning the wheel and they can compare their results with those of previous studies using the same instrument - research conducted in - The term his or her such a way that the archival research attention. The 1) determine purpose of the questionnaire subject’s behaviour is refers to situation may - What do I expect to accomplish? disturbed as little as research be a practical - try to anticipate questions of interpretation that may arise when you have the data possible by the observation conducted using problem that 2) determine the types of qns process data that the must be Open-ended qns Closed-ended qns - The observation is made researcher had solved as in the environment where no part in soon as one that the respondents answer in their own words one that limits the respondents to certain alternatives the behaviour naturally collecting possible, or it occurs - Data that bear may be an >more useful for smaller and preliminary studies (Coding > closed-ended questions makes them more suitable for - Physical trace on the event, a small number of open-ended surveys may be large studies measures→ make use of hypothesis may person, or manageable, whereas hundreds would not) + easier to code and analyse physical evidence of some already exist, animal that >trying out a preliminary version of a survey with + fewer off-the-wall responses behaviour→ e.g and collecting intrigues a open-ended questions can determine the range of likely + alternatives are presented to the respondents, so they researchers have studied new data would researcher answers, permitting you to standardise the alternatives do not have to think as hard→ respondents do not need graffiti in school restrooms be wasteful. Or - Case into a closed-ended format that will be easier to deal to be as articulate to formulate their answers as they do to discover attitudes toward ethics or logistics studies are with in the larger administration with an open-ended question racial integration of the may make it typified by the - issues being studied may be too complex to reduce to school, smudges on pages infeasible to varied nature +more likely that the questionnaire will discover a small set of alternatives of library books to see conduct an of the something not anticipated by its designers. - respondent may not agree with any of them, resulting which pages are most read, experiment methods - harder to code→ necessary to categorise responses in in simplistic answers and grease prints on relating the used to study some way to summarise the data after the survey is - Closed-ended questions tend to put words into the display cases in museums variables of the problems complete→ makes data analysis a messy job and mouths of respondents, suggesting alternatives that to see which displays are interest and intensive makes it likely that you will have to break a cardinal rule respondents might never come up with themselves most interesting to children - however. First, description of of research by not deciding in advance how you are - If a respondent misinterprets the question or a clerical - laboratory observation: most archival a single going to analyse your data error is made in coding the data, there may be no way to type of observation that data are individual or a - open-ended questions require more effort from the discover the fact → To reduce errors, many occurs in the laboratory collected for single group respondents and are more difficult for less articulate questionnaires require that each response be recorded rather than in the field nonscientific of individuals respondents to answer in two places, so the responses can be tested for - rules: careful record reasons. - case studies consistency keeping, the use of a Governments often include variety of types of and private the use of * two types of questions are mixed in a single study, when respondents may be offered the opportunity to expand on measures, and care for agencies collect observation the answers to a closed-ended question. This permits the data to be coded and analysed easily but gives some privacy of the participants the data for their and archival insights into respondents’ reasons for choosing the alternative they did -Careful record keeping is own purposes, methodologie what separates naturalistic and such data s, the 3) Write the item observation from casual often do not suit distinctions Address a single issue per item impression formation. The the purposes of among them ○ No double-barreled qns e.g students should be graded because they can be prepared for the competitive observer should keep a the scientist are not world → contains both an opinion about grading and reason for grading→ can agree with opinion but record of all behaviours of - Archival always clear disagree with reason interest and the times at research is by - Many case Avoid bias which they occur. A check nature carried studies result ○ E.g of bias qns: Do you believe in killing unborn babies? , Should women be forced to bear unwanted sheet may be used when out after the fact, from children? all or most of the categories ruling out problems that ○ E.g of no bian qn: Are you in favour of the death penalty for a person convicted of murder? of behaviour under alternative present Make alternatives clear observation are known in hypotheses for themselves to ○ write closed-ended questions in such a way that the options are distinctly different from one another and that advance. The recording of particular researchers they cover all possibilities→ answers must be mutually exclusive and exhaustive information is facilitated by observed as Beware of social desirability tendency (same e.g as bias) using cameras, video or correlations may opportunities ○ Bias often enters when respondents perceive one alternative as more socially acceptable than the other—a audio recorders, or other be difficult. that must be phenomenon called social desirability devices. Many times, using - A researcher grasped ○ Researchers avoid this problem by wording questions so that each alternative appears equally socially slow-motion or who relies on quickly or lost desirable speeded-motion recording archival data is - Little time ○ include a set of questions designed to detect if a person has a tendency to be overly influenced by social is helpful to make at the mercy of may be desirability → verification key behaviour patterns easier any biases that available for Beware of acquiescence to see; stop-action may have planning, and ○ acquiescence: the tendency to agree with a statement on a questionnaire, regardless of its content recording can freeze critical occurred in the study ○ binary closed-ended questions such as Agree/Disagree or True/False are highly susceptible to bias caused moments collecting the often must be by acquiescence data conducted Determine format of item Participant-observer - One of the under difficult ○ true/false, multiple choice, ratings,etc. research challenges of conditions ○ visual analogue scale (VAS) - investigators participate in performing - narrative ○ Likert scales groups and record their archival research case study: a observations is finding the viewpoint - careful records and specific archives expressed by diaries are crucial in that have the telling and evaluating information listening to ○ participant-observer studies relevant to your stories that ○ Branching → permit the respondent to skip inappropriate items and move through a questionnaire more because of the increased research needs communicate efficiently (e.g yes-skip to q25, no-continue to q2) possibility of subjectivity in meaning Sequence the items these situations - A narrative ○ Answers to some questions may be biassed if they were to follow some others - most useful in studying a is essentially 4) Determine How the Data will be analysed small group that is a story told how the questionnaire is to be scored and analysed separated from the firsthand, and decide what statistics will be used. Will you be able to draw proper conclusions from the data? → devise the population as a whole, it reflects the form on which you will code your data as a way of checking to see if your questionnaire is well designed when little is known about a meaning group, or when the group’s experienced Administering the Questionnaire activities are not generally by the teller. 1) Determine the method of administration available to public view In other Face to face - participating in a group words, one ✔ interviewers can establish rapport with the people being interviewed leads to problems of hears directly ✔ Interviewers can direct the attention of the respondents to the material and motivate them to answer the objectivity. The researcher from the questions carefully must strike a balance participant in ✔ Interviewers can guarantee the order in which questions are administered, thus making sure that people between taking the a narrative answer the survey in the order intended viewpoint of the group ✔ interviewers may be able to notice when respondents seem to misunderstand a question and explain its members and maintaining meaning and can probe for more complete answers when a respondent gives a brief answer or one that scientific objectivity does not respond to the question - certain groups are hostile ✔ Visual aids can also be presented to clarify a survey question in a face-to-face testing situation that may be to the larger society and impractical to present with other methods of survey administration. suspicious of anyone who 🗙 presence of the interviewer creates a social situation that may result in biassed responses→ Respondents shows an interest in them. say what they think interviewers want to hear Admitting that you are there 🗙 more expensive due to need to travel to the respondents’ locations to study them might result 🗙 personal interviewing has the problem that it is more difficult to supervise the interviewers → can fake data in your being kicked out or to save time worse. In that situation, the Written responses researcher would use ✔ questionnaires may be administered to a group, they may be dropped off at a particular location, or they disguised participant may be mailed to the respondents observation to hide his or ✔ Group administration is a very efficient use of time and money and can have a very high response rate if her true purposes from the attendance by group members is high people being observed ✔ Drop-off administration is often done by an organisation, such as a church, that has many members who - by entering the group, the attend a particular location over some period of time but may not all be present at one time observer (by definition) ✔ low cost changes it to some extent. ✔ Except for group administration, respondents can complete the questionnaire at their leisure, and they Therefore, the act of have greater anonymity in their responses, reducing interviewer bias observing the behaviour ✘ Drop-off and mail administration may have very low rates of responding, often less than 50% changes the behaviour to ✘ no possibility of clarifying questions that might be misunderstood be observed. A large group ✘ Illiterate participants are frequently embarrassed to admit their problem and sometimes respond to may not be influenced questions they do not understand, compromising the reliability and validity of the data. much by an observer’s ✘ It is impossible to determine how seriously the respondent took the survey e.g child did form for the fun of it presence, whereas a small instead of parent group may be influenced Computerised administration considerably ✔ questionnaire may be distributed via e-mail, posted as a Web-based survey that is open to the general - Participant observers public, or administered in a laboratory setting cannot always obtain ✔ impersonal, so social desirability may be reduced informed consent from the ✔ absolutely consistent people they study. Some ✔ investigator can be sure that all the questions were asked in order, and none were skipped researchers hold that ✔ can check for invalid responses and prompt the interviewer to recheck implausible answers participant-observer ✔ computer can control the sequencing and branching of questions so that, for example, people who do not research is therefore drive will not need to be asked about how many miles they drive to work always unethical ✔ Internet surveys are available to people 24 hours a day, making a high level of participation possible without the costs of photocopying or postage ✘ Illiterate or uncooperative participants will provide meaningless data (same as written) ✘ totally unmonitored, people taking an anonymous Web-based survey might not be honest about their ages or genders  ✘ A truly random sample of respondents may also be difficult to obtain because people participating in a Web-based survey are self-selected→ unlikely that people who do not own a computer will even be aware of the Web-based study, much less able to participate, so a participant selection bias may be present ✘ storage of survey information obtained electronically must be ensured (via backups, etc.) without violating the privacy of the person responding to the questionnaire Telephone administration ➡ Nowadays the percentage of people who can be reached by telephone is about as high as the percentage that can be reached by other means. ➡ random-digit dialling→ choosing an exchange and then randomly selecting the last four digits from a random-number table→ may take as many as five calls on the average to reach a working number that is a residence instead of a business but may still involve less effort than some other methods of administration ✔ low cost ✔ can be conducted rapidly, without having to wait for interviewers to travel to many locations or for respondents to mail back their completed surveys ✔ computer-assisted interview→ interviewer reads the questions from a computer screen and types the answers onto the keyboard ✔ can be conducted from a central location where the interviewers can be supervised→ can reduce slippage and ensure that they administer the survey as it was designed ✗ external validity of telephone surveys is reduced by the fact that only people who have a home telephone and are willing to put up with the intrusion can be sampled→ more and more people are giving up their home phones in favour of cell phones ✗ cell phone users pay for their minutes, they are somewhat disinclined to spend them on answering a questionnaire ✗ calculation of response rate is more complicated and less reliable with cell phone–based surveys ✗ Some states prohibit using auto-dialers (a way of rapidly calling telephone numbers) with cell phones, so calling these individuals can be more expensive for the agency conducting the survey ✗ a person with a cell phone can be physically located almost anywhere, sampling appropriately from within a specific location, such as everyone who may live within a particular town, is extremely difficult ✗ telephone surveys are less anonymous than mail or Web-based surveys and introduce the possibility of interviewer bias ✗ more difficult to ask complicated or open-ended questions over the telephone than with a written questionnaire ✗ Compared with face-to-face interviews, it is harder to establish rapport or to judge the degree of seriousness with which the respondent is taking the interview, and it is impossible to use visual aids ✗ Telephone surveys must also be relatively short to get participants to finish the survey 2) The problem of response rate More than a third of the American population may refuse to participate in surveys The response rate varies significantly among methods of administration. Surveys printed in magazines may have a 1% or 2% response rate. Mail surveys often have return rates between 10% and 50%, telephone surveys 80%, and face-to-face surveys 90% All of the data come from people who are motivated to respond. Most people feel only moderately one way or the other about an issue such as gun control, but a few are strongly, perhaps violently, opposed→ A survey with a low response rate will be biassed in the direction of the more vocal people  The Hermeneutic Approach ○ hermeneutics: the principles of interpretation of a text’s meaning ○ describe methodology that looks more at interpretation than causation ○ in some places the traditional cause–effect approach seems more appropriate, and in others the hermeneutic approach is more natural Sampling Types of sampling Haphazard samples Purposive samples Convenience samples Probability samples population subgroup a nonrandom sample that is chosen a nonrandom sample that obtain their respondents in some for whose selection for some characteristic that it is chosen for practical manner such that the researcher the researcher uses possesses reasons knows the probability that any hit-or-miss methods - considered to constitute a population - it selects a desirable given individual will appear in the - almost worthless - error in judgement by the researcher group of people but differs sample. Whereas the other three in selecting the sample may influence in that it may not come types of samples permit only the results close to sampling all of a subjective evaluation of the - chosen representative may not know population validity of the results, probability what is most desirable answer for most samples permit one to apply of population they are representing various statistics Probability Samples and Random Selection Random sample: a sample in which every member of the population has an equal and independent chance of being selected The sampling frame: a population as it is defined for the purposes of selecting subjects for a study ○ define the population for the purposes of the survey (e.g students who attend the class even if their name not on class list), and this may be different from the actual population (e.g students on class list) ○ element: individual member of a sampling frame Systematic Samples: a probability sample that is not randomly selected ○ involves selecting elements from an ordered sampling frame ○ Choose every fourth name from the class roster → not random→ fails the equal-probability part of the definition of random selection (1,4,8 have 100% of being chosen, the rest 0%--> not equal probability) ○ although systematic samples are not random, taking every nth individual from a roster is much less work than the random method ○ If the list has no structure to it, however, the results will be as good as random in practice (e.g men’s name after his partner→ have structure, alphabetical order→ no structure) Simple Random Samples: group chosen from an entire population such that every member of the population has an equal and independent chance of being selected in a single sample ○ used when we believe that the population is relatively homogeneous with respect to the questions of interest ○ E.g random-number tables Stratified random samples: a random sample in which two or more sub samples are represented according to some predetermined proportion, generally in the same proportion as they exist in the population ○ By stratified random sampling you can ensure that the proportion of men and women in the sample matches that in the college population ○ treats the population as two or more separate subpopulations and creates a separate random sample of each→ procedure is still random, however, because every member of the population had an equal and independent chance of being selected ○ Sometimes stratified random sampling is used to oversample some sub group of the population—that is, to purposely include some group at a greater frequency than it is represented in the population E.g You would want to include the same number of black people and white people in the survey to get as reliable an estimate of the attitudes of black people as of white people, even though black people may constitute only 10% of the population. You would stratify on race and include 50% black people and 50% white people in your sample. The sampling would still be random within the subpopulations Cluster samples: group selected by using clusters or groupings from a larger population ○ Many populations would be impossible or impractical to number. For instance, making a list of every person in the United States would be impossible  ○ the ease of obtaining the sample would permit you to study more individuals and therefore offset the disadvantages of not having a purely random sample ○ If you wanted to make sure your sample contained the same proportion of students in particular categories as the college as a whole, you could stratify your clusters. ○ multistage sampling: a form of cluster sampling in which clusters are further broken down by taking samples from each cluster Random Sampling 1) Define and identify the sampling frame. 2) Determine the desired size of the sample. 3) Compile a list of all members of the population, and assign each member on the list a number from zero to the required number. 4) Group the columns of digits according to the required number of digits— for example, three digits for numbers up to 999. 5) Arbitrarily select a number in the random-number table by closing your eyes and pointing. 6) If the selected number corresponds to the number assigned to any member of the identified population, that member is in the sample. 7) Repeat Step 6 by running down the table until the desired number of subjects has been selected. Stratified Random Sampling 1) Identify the sampling frame. 2) Determine the desired size of the sample. 3) Determine the subgroups, or strata, for which you want equal or proportional representation. 4) Identify each member of the population as a member of one of the subgroups or strata. 5) For each of the population subgroups or strata, assign each member a number from zero to the required number. 6) Use a random-number table to select the appropriate number of subjects from each of the subgroups or strata. Cluster Sampling 1) Identify the sampling frame. 2) Determine the desired sample size. 3) Identify and list all appropriate clusters. 4) Assign all clusters on the list a consecutive number from zero to the required number. 5) Determine the average number of subjects in each cluster of the population. 6) Determine the number of appropriate clusters by dividing the desired sample size by the estimated size of a cluster. 7) Use a random-number table to select the appropriate number of clusters. 8) Either select randomly from the clusters or use the entire cluster. - A of odd = A of even - but % different → shows how phrasing of question makes a difference  Chapter 5 Variables: aspect of a testing condition that can change or take on different characteristics with different conditions Type of Variables Dependent and Independent Confounded Quantitative Continuous and discrete & categorical (quantitative variable) dependent variable: a measure of the subject’s confounded quantitative continuous variable: one that falls behaviour that reflects the independent variable’s variable: one variable: one along a continuum and is not limited effects [outcome variable] whose effect that varies in to a certain number of values - response that the person or animal makes→ may cannot be amount e.g latency, duration, or force of a be a score on some sort of test, or it may be a separated from e.g speed of bar press → can be measured with behavioural response that can be measured using at the supposed response any desired precision → fineness of least one of several different dimensions independent and loudness measure limited by ability of the - e.g frequency (the number of times that a variable measuring instrument behaviour is performed), rate (the number of times - Confounding is categorical that a behaviour is performed relative to time), a large problem variable: one discrete variable: one that falls into duration (the amount of time that a behaviour lasts), in research using that varies in separate bins with no intermediate latency (amount of time between an instruction and subject variables, kind values possible when the behaviour is actually performed), such as gender, e.g College e.g number of marriages contracted, topography (the shape or style of the behaviour), because males major and murders committed, or books written force ( the intensity or strength of a behaviour), and females vary gender locus (where the behaviour occurs in the on many Although a variable may be environment) dimensions continuous, its measurement is often discontinuous Independent variable: the condition manipulated or Not All Details of e.g height is a continuous variable, selected by the experimenter to determine its effect a Study Are we generally measure it to the on behaviour [predictor variable] Independent nearest inch - what you do to the subject Variables - Every independent variable has at least two values; - variables: have Because continuous variables are otherwise, it wouldn’t be a variable. These values to vary commonly measured in a are commonly called levels (e.g 2 levels: frustrated - To make sure discontinuous fashion, it is necessary and not frustrated) that unintended to distinguish the real limits of a variables do not measure from the apparent limits variable of interest: a variable for which its role in the confound a study, - real limits: the interval defined by cause and effect of an observed relationship is not many potential the number plus or minus half the clear variables are distance to the next number - usually dependent (cause) and independent (effect) kept constant, or - apparent limits: the point indicated the same, across by a number subject variable: a difference between subjects that conditions cannot be controlled but can only be selected - independent variables that researchers do not manipulate (e.g sex, age, IQ) Measurement: the process of assigning numbers to events or objects according to rules Types of measurement scales Nominal scales Ordinal scales Interval scales Ratio scales nominal scale: a measure ordinal scale: a measure interval scale: a measure in which ratio scale: a measure having that simply divides objects that both assigns objects the differences between numbers a meaningful zero point as or events into categories or events a name and are meaningful; includes both well as all of the nominal, according to their arranges them in order of nominal and ordinal information ordinal, and interval properties similarities or differences their magnitude - rule for assigning numbers to - You may use any number - simplest kind of scale - reflect degree of events or objects on an interval that seems appropriate; there - Categorical magnitude→ rule for scale is that equal differences is no upper or lower limit to - no numerical value assigning numbers on an between the numbers on the scale the numbers you may use (e.g - Number assignment is ordinal scale is that the must represent equal psychological -10) arbitrary → objects or ordinal position (rank differences between the events or - rule for assigning numbers to events of the same kind get order) of numbers on the objects events or objects on a ratio the same number and scale must represent the - no true zero point → zero point is scale is that the ratios objects or events of a rank order of the arbitrary between the numbers on the different kind get a different psychological attributes of - cannot express difference in terms scale must represent the number (no. just the objects or events of ratio psychological ratios between representation, no - interval may not be equal E.g Difference between 20°C and the events or objects value/meaning) → scale does not say how 10°C is 10 degrees. Difference - can express in terms of ratio - E.g. Sex, favourite colour, much more the person between 40°C and 30°C is also 10 - true zero point → suggests Citizenship status prefers A to B → scale degrees. BUT, 40°Cis NOT twice as absence of the property - e.g which you prefer gives only the order of hot as 20°C. E.g speed, weight most.. preference, not the E.g how much do you like (rate 1-5) E.g how many times you difference in preference - can obtain a mean value to bought.. among items compare → mean meaningful since difference equal Likert Scale Not nominal → have numerical value Not ratio → no true zero value (absence of property) → zero is arbitrary Not interval → psychological difference between numbers not equal Not ordinal → as mean meaningful By practice is interval as we calculate mean but by theory should be ordinal Reliability and Validity of Measurements - For a measurement to be of any use in science, it must have both reliability and validity reliability: the property of consistency of a measurement that gives the same result on different occasions [consistency] validity: (of a measurement) the property of a measurement that tests what it is supposed to test [accuracy] - Variability and error error variance (random error): variability in the dependent variable that is not associated with the independent variable the task of research is to find how the dependent variable changes with changes in the independent variable → would be affected by error variance - Validity of measurements  Construct validity: a test that the measurements actually measure the constructs they are designed to measure, but no others ○ test should actually measure whatever theoretical construct it supposedly tests, and not something else (e.g test of leadership ability should not actually test extraversion) ○ test should measure what it intends to measure but not measure theoretically unrelated constructs (e.g test of musical aptitude should not require too much reading ability) ○ test should prove useful in predicting results related to the theoretical concept it is measuring (e.g A test of musical ability should predict who will benefit from taking music lessons, should differentiate groups who have chosen music as a career from those who haven’t, should relate to other tests of musical ability, and so on) Face validity: idea that a test should appear superficially to test what it is supposed to test ○ Using common sense to see if test is reasonable Content validity: idea that a test should sample the range of behaviour represented by the theoretical concept being tested ○ e.g An intelligence test, for example, should measure general knowledge, verbal ability, spatial ability, and quantitative skills, among others→ an intelligence test that measures only spatial ability would not have sufficient content validity Criterion validity: idea that a test should correlate with other measures of the same theoretical construct ○ E.g A valid test of intelligence should correlate highly with other intelligence tests ○ If the criterion of an intelligence test is whether it correlates with how well a child is doing in school at the time the test is given, it is called concurrent validity ○ If the criterion of an intelligence test is how well the test can predict some future performance of the child, such as graduation from college, then it is called predictive validity Type of measurement error ○ Systematic error: measurement error that is associated with consistent bias Systematic error is never desirable in research, but it may not be such a serious problem if the error is the same for the entire study—that is, all groups or conditions of the study are equally affected by the systematic error Systematic error can be very serious, though, if it is associated with the independent variable, because it can confound your experimental results → The systematic error may be present for one level of the independent variable and not for another level ○ Random error (see definition above) always a serious problem in research because it can reduce the precision with which you assess the effects of the independent variable → threat to the reliability of measurement Type of reliability measures ○ Test-retest reliability: the degree to which the same test score would be obtained on another occasion A good test gives a similar score on two occasions ○ Internal consistency: the degree to which the various items on a test are measures of the same thing Split-half reliability→ determined when the items on a test are divided into two sets as if they were two separate tests→ scores on the two halves are correlated to see how closely the various individuals’ scores agree on the two halves → If the test is a good test, it will have a high split-half correlation  Chapter 6 and 7 Validity : an indication of accuracy in terms of the extent to which a research conclusion corresponds with reality Internal validity Construct validity External validity Statistical conclusion validity extent to which a study provides evidence of a cause-effect extent to which the results support the how well the findings of an extent to which data are shown to be relationship between the independent and dependent theory behind the research experiment generalise to other the result of cause-effect relationships variables situations or populations: different rather than accident - If the measurement used in some subjects, settings, times, - an experiment with high internal validity, it really was the research lacks construct validity, the treatments, observations, and so - did the independent variable truly independent variable that caused the dependent variable to research as a whole will also lack forth cause a change in the dependent change construct validity variables, or was the result accidental, - confounding: error that occurs when the effects of two - Every study is designed to test some E.g experiments in the past may and thus caused by pure chance? variables in an experiment cannot be separated, resulting in hypothesis; yet a hypothesis cannot be not generalise to today’s world → - It also asks how strong the a confused interpretation of the results (ie two independent tested in a vacuum. The particular lack external validity relationship is between the variable during an experiment causing us to not know conditions of a study constitute auxiliary independent and dependent variables dependent variable recorded was due to change in which hypotheses that must also be true so that The idea that experimental results - To establish statistical conclusion independent variable) you can test the main hypothesis (e.g obtained in a laboratory setting validity, appropriate sampling and - subject variable: a difference between subjects that cannot hypothesis: anxiety is conducive to might be different from those measurement techniques must be be controlled but can only be selected→ participants are learning, auxiliary hypothesis: that obtained in a natural setting used, and inferential statistics must be selected according to the presence or absence of a condition fingernail biting is a measure of anxiety reflects a question about ecological used properly, in keeping with their and not selected simply to have a condition assigned to them and that writing with one’s toes is a good validity: extent to which an underlying assumptions (e.g gender→ Participants cannot be assigned to one learning task → If either of these auxiliary experimental situation mimics a gender, but must be selected from preexisting groups) hypothesis is false, you could have found real world situation - A statistical test establishes only that negative results) an outcome has a certain low For internal validity, you may find it possible to redesign the - To improve the validity of your // probability of happening by chance study to control for the source of confounding experiment, you might have used a alone but it does not guarantee that manipulation check: aspect of an THREAT the change in the dependent variable // experiment designed to make certain that Other subjects was the result of a true cause-effect variables have changed in the way that - Human participants should be relationship with the independent THREAT was intended chosen with equal attention to their variable; there is still a chance that it Ambiguous temporal precedence representativeness relative to might instead be the result of random although two variables are related, it is not clear which one is In the case of construct validity, you must some larger population. If you are error in sampling or measurement the cause and which one is the effect design a new study that will permit a doing an experiment with college choice between the two competing students on bargaining and - Were there enough observations Events outside the laboratory (history) theoretical explanations of the results. negotiation, will the results validly made to make it likely that the null history: events that occur outside of the experiment that predict what a secretary of state or hypothesis could have been rejected if could influence the results of the experiment // a general would do? it were false? - power: the probability of rejecting the Maturation THREAT Other times null hypothesis when it is, in fact, false a source of error in an experiment related to the amount of Loose connection between theory and - Would the same experiment - If an experiment suffers from lack of time between measurements method conducted at another time produce power, the experiment may appear to - Subjects may change between conditions of an experiment E.g Nail biting is a poor method of the same results? show that the null hypothesis is because of naturally occurring processes e.g children get measuring anxiety, and writing with the - Many historical trends render supported; when in reality, it should be older between testing sessions→ change in motor toes is likely to be a poor measure of particular research findings invalid, rejected. In that case, the result suffers coordination, knowledge → Influence results learning. whether they concern use of from a lack of statistical conclusion - it is a more critical problem in research involving children - Much psychological research suffers language, attitudes toward foreign validity because they change more rapidly over time than do adults from poor operational definition of countries, or perception of deviant - effect size: strength of the theoretical concepts. groups relationship between the independent Effects of testing and dependent variables → help to effect of repeat testing: performance on a second test is Ambiguous effect of independent establish whether a significant result is influenced by simply having taken a first test variable Other settings of practical importance E.g participants may become sophisticated about the testing - An experimenter may carefully design - how the phenomenon observed in procedure or may learn how to take tests after taking the first an experiment in which all reasonable one laboratory can be related to a // test confounding variables seem to be well similar phenomenon observed in controlled, only to have the results another laboratory or in the real THREAT Regression effect compromised because the participants world - arise from improper use of statistics tendency of subjects with extreme scores on a first measure perceive the situation differently than the - Through laboratory research in analysing the data to score closer to the mean on a second testing experimenter does ensures a higher level of control, it - power → major threat may be the -e.g students who scored highest on the first test usually do - Whenever people are aware that they is sometimes not easy to decide if conclusion that the independent worse on the second, whereas those who did the worst are participating in an experiment, their a certain effect is simply a variable had no effect, but if your study improve behaviour may be different from their laboratory effect or whether it employed too few subjects or made - For example, on a multiple-choice test, students will know everyday behaviour would survive transplantation to the too few observations, your conclusion some answers and make some lucky or unlucky guesses on world outside the laboratory may be erroneous the rest, resulting in a score that is not a perfect indicator of - inaccurate effect size estimation→ what they know. Instead, their score on the test represents the size of the relationship is the student’s level of knowledge plus some level of random measured poorly. For example, outliers error. that change the shape of a distribution -random error: that part of the value of a variable that can be can increase or decrease the attributed to chance estimated effect size Selection a confound that can occur due to assignment of subjects to groups - must exercise care and ingenuity to choose or create groups that can be considered truly comparable Mortality the dropping out of some subjects before an experiment is completed, causing a threat to validity - Even if there is no bias in selecting participants and you are able to constitute groups that are the same in every respect, your study may be invalid if all subjects do not complete all phases of it. Control: any means used to rule out threats to the validity of research (1) What are the threats to the validity of a contemplated piece of research? (2) What means are available to neutralise those threats? Not every experiment must have a control group Concept of control: Control provides a standard of comparison ○ providing a standard against which to compare the effect of a particular independent variable ○ experimental group: subjects in an experiment who receive treatment ○ control group: subjects in a between-subjects design experiment who are like the experimental group in every respect except that they do not receive treatment - control condition: a condition in a within subjects design experiment that does not contain the experimental manipulation - within-subjects experiment: research design in which each subject experiences every condition of the experiment - between-subjects experiment: research design in which each subject experiences only one of the conditions in the experiment Control reduced variability ○ the ability to restrain or guide sources of variability in research ○ Limiting the things that change to those mandated by the experimental design (such as the independent variable) reduces the chances of confounding variables or measurement error and increases our confidence in the experimental results ○ Keeping everything constant except the independent variable gives us confidence that it is the only thing that could have caused the change in the dependent variable Relating the Two Meanings of Control ○ (1) Primary meaning: Allows one to conclude that a dependent variable is associated with an independent variable and not with any other variable ○ (2) Secondary meaning: Facilitates drawing this conclusion by so limiting the number of variables operating in the situation and their range of values that the conclusion is clearer ○ When we have experimental control (secondary meaning), we have a much more sensitive situation in which to rule out alternative explanations of the experimental results (primary meaning) General strategies Control in the laboratory ○ Laboratory work in social psychology requires control over elements such as choice of participants, beginning and end of social interaction, and freedom from distraction ○ in some cases, the effect of a manipulation might not be realistic enough in the laboratory ○ People who advocate field research agree that they must give up a degree of control and that problems of internal validity thus become greater The Research Setting as a Preparation ○ A preparation is an environment that is selected or constructed for a particular purpose ○ Everything that is a part of the research setting is a part of the preparation, which is really a context for data collection ○ Preparation includes the experimental equipment, the method of testing, and the location of testing, as well as the subject used in the study. ○ one of the researcher’s goals is to choose the most suitable preparation for studying a given problem ○ Exactly which situation will provide the most powerful relationship between the variables of interest?  Instrumentation of the Response as Control ○ important means of increasing the sensitivity of the research is to improve the measurement of the behaviour being studied ○ the precision with which you measure the dependent variable can influence your result ○ One characteristic of a good measurement instrument is that it takes the response out of the realm of casual observation and makes it reliable, giving accurate measurements time after time ○ Only in this way can we speak of the measurement of behaviour as objective, thus meeting the requirement of interobserver reliability necessary for science. ○ Therefore, even a measure of a subjective state, such as the pleasantness of an odour, can be considered objective, provided the instrumentation of the response is adequate Specific Strategies Subject as own control (within-subject control) ○ One of the most powerful control techniques is to have each participant experience every condition of the experiment → variation caused by differences between people is greatly reduced ○ The experimental manipulation is not likely to destroy the naiveté of the participant, who is unlikely to guess the purpose of the experiment even after experiencing it ○ In addition, if enough time is allowed between conditions, there is unlikely to be an important carryover between conditions → The participant will recover in a few minutes from the effect of adaptation to salt and will be ready to experience the next condition ○ In many experiments, however, using subjects as their own controls simply is not possible. For example, once the participant has learned something by one method, learning the same problem again by using a different method is impossible ○ For some experiments, contrast effects exist between the conditions of the experiment, so that experiencing one condition may carry over and influence the response to another condition → using subjects as controls not feasible Random Assignment (Between-Subjects Control) ○ random assignment: unbiased assignment process that gives each subject an equal and independent chance of being placed in every condition ○ The advantage of random allocation of subjects to conditions is that once subjects have been randomly assigned, the only way that confounding of subject-related variables with the experimental variable can occur is by chance ○ Computer generated randomizations may or may not be truly random, depending upon the process by which they are calculated. Some programs use the time of request as a “seed” to generate the numbers, while others use different technique Matching (Between-Subjects Control) ○ matching: control procedure to ensure that experimental and control groups are equated on one or more variables before the experiment ○ When the subjects differ among themselves on an independent variable known or suspected to affect the dependent variable of interest, matching may be necessary ○ The first requirement to justify matching is a strong suspicion that there is an important variable on which the subjects differ that can be controlled. ○ Further, you must believe that a substantial correlation will be present between the matching variable and the dependent variable ○ It is possible to weaken your experiment by matching the subjects if the matching variable is not substantially correlated with the dependent variable. This effect results because the statistical test appropriate for a matched-groups design considers the data from pairs of subjects, whereas the randomised groups test considers individual subjects ○ A second condition necessary to justify matching is that it must be feasible to present a pretest to the subjects before assigning them to the conditions. Building Nuisance Variables into the Experiment ○ Nuisance variable: a condition in an experiment that cannot easily be removed and so is made an independent variable as a means of control ○ Nuisance variables are known or suspected to affect the dependent variable, but variables in which you have no theoretical interest  ○ Left uncontrolled, these variables may affect the dependent variable so strongly that they hide the true effects of the independent variable. ○ Building these nuisance variables into your study allows you to measure their effects and to examine the effects of your independent variable. Statistical Control ○ statistical control: mathematical means of comparing subjects on paper when they cannot be equated as they exist in fact Replication replication: repeating an experiment to see if the results will be the same Direct replication ○ occurs when someone repeats essentially the identical experiment in an attempt to obtain the same results Systematic replication ○ occurs when Researcher B says, “If A’s theory is correct, then the following should happen.” Then B performs an experiment different from A’s but based on it. If A’s results and theory are correct, B should find a result that supports the theory Systematic replication tests external validity by using different subjects, species, or situations. Construct validity is tested when different ways are used to measure the theoretical concepts. Statistical validity is tested in all replications, both direct and systematic  Chapter 10 True Experiment Quasi experiment research procedure in which the scientist has complete control over all aspects research - Control over the who of the experiment means that the experimenter can assign subjects to procedure which conditions randomly → preferred because it allows one to conclude that any other variable does not meet the could be confounded with the independent variable only by chance requirements of a - Control over the what, when, where, and how of the experiment means that the experimenter true experiment has complete control over the way the experiment is to be conducted. The conditions of a true experiment allow the researcher to infer causality, or to say that the changes in the dependent variable were caused by the independent variable Factors Levels Conditions Treatments the independent in an experiment, a a group or treatment in an another word for a variables of an particular value of an experiment condition of an experiment independent variable (min Between subject: 2 levels experiment two) With subject: various treatment * An independent variable always has at least two levels—if it didn’t, it wouldn’t be a variable (e.g reading and no reading; 0,1,2,3,4 min) Valid Experimental Design 1. the existence of a control group or a control condition 2. the random allocation of subjects to groups or each subject experiences all conditions Within-Subjects Designs a subject experiences more than one experimental condition the possibility exists that some variable may influence the data as a result of the repeated testing. The outcomes as a result of these variables are called carryover effects: Order effects Sequence effects changes in a subject’s performance (DV) resulting from changes in a subject’s performance (DV) resulting from the position in which a condition appears in an interactions among the conditions themselves experiment E.g experiment on judging the heaviness of lifted E.g Whichever condition is presented first will show weights→ there is likely to be a contrast effect such that poorer performance than later conditions simply because a light weight will feel even lighter if it follows a heavy the subjects had not warmed up to the task one Control of order and sequence effects Block randomisation Reverse counterbalancing control procedure in which the order of conditions is randomised method of control in which conditions are but with each condition being presented once before any condition presented in order the first time and then in is repeated reverse order - If there are four conditions and each one is to be represented - When relatively few subjects will be tested twice, block randomization might give you the following sequence: and you have several conditions that can be BCAD, ADCB presented only a few times - most useful when conditions are presented several times to each - e.g ABCCBA subject The disadvantage of this method of counterbalancing is that as the number of conditions increases, the number of orders required increases geometrically e.g 2 possible orders of 2 conditions (AB,BA), Even for only four conditions, you would need 24 subjects to control for order and sequence using complete counterbalancing  ○ If presenting each condition enough times to randomise the order is not possible, or if counterbalancing within subjects does not seem appropriate, you must leave order and sequence confounded with conditions within subjects. Then you must control for order and sequence between (or across) subjects, essentially within the group Latin square: control procedure in which each subject experiences each condition in a different order from other subjects Incomplete balancing → Instead of every possible sequence , ensure that each condition occurs in every rank order A disadvantage of the Latin square technique is that sequence is not controlled ○ However, you can control for sequence effects of the immediately preceding condition by using particular sets of Latin squares known as balanced squares. In the balanced Latin square, each condition is immediately preceded once by every other condition e.g ○ ○ When you can assume that the carryover effects are primarily between pairs of conditions, the balanced Latin square will be effective The advantage of the Latin square technique over complete counterbalancing is that it permits greater flexibility in choosing the number of subjects to be tested. Instead of needing 24 or 48 subjects in a four-condition experiment Two Conditions, Tested Within Subjects [IV→ DV→ IV→ DV] ○ two-conditions design: the simplest research design, involving only two conditions ○ All subjects experience both conditions in counterbalanced order ○ there is the possibility of carryover effects from one condition to the other Multiple Conditions, Tested Within Subjects [IV→ DV→ IV→ DV → IV→ DV] ○ multiple-conditions design: research design that involves more than two conditions ○ determine the shape of the function that relates the independent and dependent variables ○ presence of more than one rival hypothesis that must be ruled out → To accommodate the three hypotheses, you would need three conditions. Each condition serves as a partial control for the other hypotheses. ○ Many multiple-conditions experiments are between-subjects experiments because it is often impossible or inappropriate to expose all subjects to the various conditions *both tests face problems with order effects→ deal with counter balancing (see above - latin square) Between Subject designs Two Conditions, Tested Between Subjects [1 IV, 2 levels] ○ E.g exercise and no exercise ○ each person served as their own control.  ○ This design may not be desirable when the possibility of large order or sequence effects is present. Multiple Conditions, Tested Between Subjects [1 IV, 3 levels] ○ E.g running, stretching, no exercise ○ Some Designs to Avoid The One-Group Posttest-Only Design The Posttest-Only Design with The One-Group Nonequivalent Control Groups Pretest-Posttest Design research design that measures the nonequivalent control group: a group of research design that behaviour of a single group of subjects subjects that is not randomly selected measures the behaviour of after they are given a treatment from the same population as the a single group of subjects experimental group both before and after - no baseline for comparison treatment - would not know change or whether - A nonequivalent control group is change was due to IV better than no control group, but you - no control group would have to consider this study a - have problems e.g surveying only after without before to quasi experiment at best because the determining what caused compare → leaves many threats to validity subjects were not randomly assigned the change → is change uncontrolled to groups attributed to IV? Chapter 11 factorial design: research design that involves all combinations of at least two values of two or more independent variables Main effects the effect of one independent variable, averaged over all levels of another independent variable Difference in row mean → main effect of B Difference in column mean → main effect of A If difference=0 then no main effects, if less than or more than 0 then main effect present E.g IV 1: attractiveness: attractive and unattractive IV 2: facial expression: neutral and smiling DV: “Guilt” Another e.g  Interactions when the effect of one independent variable depends on the level of another independent variable The two curves in Figure 11.2 are not parallel; therefore, there is an interaction between facial expression and attractiveness → if parallel suggest no interaction → this is true no matter how complicated curve might be Main effects may be present without interactions ○ Medication not affected by exercise It is also entirely possible to have an interaction if one or the other independent variable has no main effect, or even if neither independent variable has a main effect Type of interactions Antagonistic interaction Synergistic interaction Ceiling-effect interaction Interaction in which the two independent interaction in which the two interaction in which one variable has a variables tend to reverse each other’s effects independent variables reinforce smaller effect when paired with higher each other’s effects levels of a second variable - no main effect of either variable - the higher level of B enhances - The higher level of B reduces the the effect of A, and vice versa → differential effect of A on the the slope of the line relating the dependent variable dependent variable (response) to - Variable A has a smaller effect when - however an interaction is present → lines in A is steeper when B is larger it is paired with the higher level of B. one graph not parallel→ A increases the - Likewise, Variable B has less effect response for condition B2, but decreases it for when paired with the higher level of A B1 - the dashed line showing the data averaged over the variable B indicates that there is no main effect of A (same for B over A)  Factorial designs Within subject Between subjects Mixed designs - 2 or more IVs (e.g A and B) - needed when subjects cannot act - when doing otherwise would be - Each IV has two or more levels (e.g., as their own controls in factorial impossible A1 and A2) designs - when using the same participants - e.g The two levels of the two variables - each participant only goes through in all conditions is possible but not give us four possible combinations of the one condition → participants split desirable independent variables randomly into the diff