Psychology Notes - PSYC 201

PSYC 201 NOTES CHAPTER 1 SCIENCE OF PSYCHOLOGY What is science? -Scientific approach: study a topic of interest 1.Systematic empiricism: -empiricism: learning based on observation -systematic: careful planning, recording, observation of phenomenon 2.Emprical questions: -questions that discov...

PSYC 201 NOTES CHAPTER 1 SCIENCE OF PSYCHOLOGY What is science? -Scientific approach: study a topic of interest 1.Systematic empiricism: -empiricism: learning based on observation -systematic: careful planning, recording, observation of phenomenon 2.Emprical questions: -questions that discover the observable truth 3.Public knowledge: -research is published (scientific article) -share knowledge of methodological improvements -replication of studies are vital Example: bilingual cognitive advantage Science vs Pseudoscience -looks like science yet is not -lacks one or more of the claims of scientific approach -scientific claims must be falsifiable, show otherwise -issue with the pseudoscience (examples, astrology, bio scream therapy) Scientific research in Psychology Methodology: 1.literature review/observations 2.decide on a research question and devise a hypothesis to be tested 3.conduct a study to test the research question or hypothesis 4.anaylse the data 5.make conclusions about the answer 6.publish your work Who conducts research: -PHD and masters degree in psychology -government departments -associations -private sector -universities Purposes of research: -basic research: to understand human behavior (ex: are there sex differences in multitasking) -applied research: to solve a practical problem (ex: law enforcement research on cell phone use and driving) Science and common sense: Folk psychology: -we can understand behavior via our common sense or intuitive belief -can lead to false belief -example: to reduce anger "let it out" shown to lead to greater anger Reasons for error: -mental shortcuts (heuristics): to form our beliefs we use mental shortcuts -systematic observation is difficult -example: general belief is that women talk more than men -example study: count the number of words spoken by men and women, difficult to test in daily life study could be done via recording device -confirmation bias: we focus on cases that confirm our beliefs (like the case that women talk more), we retain beliefs that we would prefer to be true -we all have these beliefs which is why we are skeptical (the need for evidence before accepting claims) -example: sleeping before an exam will improve your retention of information? (is there evidence to support this claim?) -tolerance for uncertainty: avoid judgement when there is lack of evidence, if there is no evidence this could be a good starting point for research -clinical practice of psychology: diagnosis and treatment of psychological disorders and related issues, before claiming that a certain treatment is effective, we need evidence to show that its empirically supported treatments -example: exposure therapy, we need pre-post test, experimental/control study design, if clinical psychologists do not conduct research themselves, they need to read and evaluate the literature and be informed of effective treatments CHAPTER 2 GETTING STARTED IN RESEARCH -Variable: a quantity or quality that varies across individuals (ex: reaction time) -Quantitative variable: quantity (ex: height), numerical variable -Categorical variable: quality (ex: a major like psychology, smoking status, hair color) -Population: all people we can test -Sample: representative group from the population to test Sampling: -Random sampling: each individual in the population has an equal chance of being selected leads to representative sample -Convenience sampling: people who were nearby or willing to participate, sample may not be representative Operational Definition: -how will you measure the variables? -example: to test recall- participant's total correct number of words retrieved from a list of previously encoded words -for an individual the total number of words recalled would be called a score -a set of scores (all the scores of your sample) would be called a data Types of studies Single variable studies: -focuses on a single variable rather than the relationship between variables -ex: how much time do students spend online? Do older adults play video games? Statistical relationships: -two variables, one differs systematically across the levels of the other -ex: is recall of an event higher for people who sleep after the event than those who do not sleep after the event? -two types of statistical relationship research 1.differences between group means: -ex: does exercise improve memory? (exercise group recall mean: 10 out of 15, non-exercise group recall mean: 5 out of 15) -the mean and SD (standard deviation) represented via a graph bar 2.corellation between quantitative variables: -average score on one variable systematically differs across levels of others -ex: does lower levels of social support decrease life satisfaction? (positive-negative relationship) -represented via a scatterplot -strength: Person's r, linear relationship -+1(strongest positive relationship), -1 (strongest negative relationship), 0 (no relationship) -closer to -1, +1 points are closer to straight line -problems: correlation does not mean causation -directionality problem: direction of effect unknown, x variable could be affecting y or y could be affecting x, (ex: children who play violent video games are aggressive or aggressive children play violent video games) -third variable problem: x and y may be related but this could be due to the third factor (z) -ex: exercise increases happiness, third factor: physical health may cause people to exercise and happy Experiments: -to find causation conduct experiment -manipulate the independent variable to see the effect on dependent variable -experimental and control group needed -ex: can playing video games improve reaction time in older adults (a group of adults randomly assigned to control or experimental group), experimental group plays video games control group does not, if video games improved reaction time conclusion is that video games improve reaction time Research questions: -discussion section of a research article-future directions, previous research can be a starting point, observations, practical problems, causes or effects of a behavior Testable research questions: -ex: we recall less from the past, could passage of time affect recall retention? -research question: what are the factors that affect memory performance Which research questions shall we test? 1.interestigness: interest to the public scientific community a\. answer is unclear, there are two opposing views b\. gaps in the literature, no one answered the question before c\. practical implications, will the results improve functioning and public safety ex: cell phone use impairs driving ability, implements laws to prevent phone use while driving 2.feasiblity: time, money, equipment and materials, technical knowledge and skills, access to the participants -ex: are dreams influenced by daily activities? We need sleep laboratories Literature search: -reveals whether your question has been addressed -if not, you have an original research question; if so you could do a replication study or alter the questions in some way and re test -research literatures are publications of research for a particular field -professional journals: Turkish journal of psychology, frontiers in psychology, international journal of psychology -empirical research reports: studies conducted by researchers -review articles: summary of research with a new take on the result -theoretical article: presents a new theory -double blind peer review: experts evaluate your research (reviewer names concealed), three outcomes (accept, revise, reject), some journals are moving towards an open review process (names of the reviewers as well as the author identity kept hidden), helps with accountability and transparency Consider: -how else can you define the variables? (ex: recall vs recognition recall) -are there factors that may make the statistical relationship weaker or stronger? (ex: age) -literature review sites: google scholar, jstor, PsycINFO, PubMed, ERIC -for literature focus on recent research in 5 years, including classical papers CHAPTER 3 RESEARCH ETHICS Respect for person: -autonomy of research participants and informed consent (purpose of research, right to decline or withdraw once the research is begun...) -protect those who cannot give informed consent, (youth, parental consent for children, cognitive impairment, other illnesses) Concern for welfare: -participants should not be exposed to unnecessary risks -consider privacy and confidentiality of participants, provide adequate information about the risks and benefits of research -risk to science: poorly designed study -risk to society: research results misapplied with harmful consequences Justice: -treat people fairly and equally -ensure minority groups are not excluded -in a treatment study, control group should be allowed the same treatment (if it works) after the study -act with integrity and ensure trust, researcher will be honest and keep their promises of confidentiality, ensure that benefits are maximized, and risk is minimized -scholarly integrity (no plagiarism, or fabricated data) Unavoidable ethical conflict: -no research that is %100 risk free -research may be beneficial to society but harmful to research participants -deception is used Research ethics board: -committee reviews research protocol for potential ethical problems -at least 5 people with different backgrounds -at least 2 people in the relevant field, one knowledgeable in ethics and one person not affiliated with the institution -duty: ensure minimal risk, benefits outweigh the risks, research done fairly, informed consent In practice: -know your ethical responsibilities -identify and minimize risks (avoid collecting unnecessary personal information) -identify and minimize deception (should be avoided unless necessary) -risk/benefit trade off (think of the benefits to society and researcher) -ensure you have informed consent and debrief form -gain ethical approval -ensure that you abide by the protocol approved CHAPTER 4 THEORY IN PSYCHOLOGY -phenomenon: reliable general result -replication of findings are usually required (failure to replicate could indicate that initial findings were due to chance or methodological differences) -theory: is an explanation of phenomenon, provides references to variables, processes structures beyond examined (ex: serial position effect explained via long term and short term memory) -model: precise explanation using equations, computer programs, biological processes -hypothesis: generally a prediction about the result and relationship between variables, should be based on theories (if, then relationship) -evaluation of theory: avoid saying we "prove" the theory, there may be cases that do not confirm the predictions -theoretical framework: is the context applied to understanding the result -multiple theories: findings are explained in multiple theories theories have purposes other than providing an explanation: -organization: organize and facilitate understanding of findings -prediction: predicts outcome in new situation -new research: new statistical relationships Variety of theories: -formality: less formal theories are simple verbal descriptions (explanations of phenomenon) vs more formal theories have mathematical descriptions (computational equations for their basis) -scope: number of dimensions that the theory tries to account Functional theories: -explain phenomenon according to their function or purpose (why explanations) -ex: theory of self-repeated injury (people's behavior is to relieve negative emotions) Mechanistic theories: -show how structures and processes interact to produce the phenomenon (how explanations) -ex: levels of processing theory (explains how differential encoding strategies lead to differential retention of information) Stage theories: -explain stages that we go through during developmental or adapt to our surroundings in forward order and do not revert back to the previous stage -ex: Piaget's theory of cognitive development Typologies: -categorize people and behavior into types, no particular order of progress (ex: different personality types, type a vs type b or intelligence types, mathematical vs emotional IQ) Theory testing: deductive method: -start with a theory and make new predictions -if the theory is correct then the predictions should be supported -theory revised based on findings Ex for deductive method: -prediction: second language speakers have faster reaction time compared to first language speakers -theory: language coordination allows better cognitive control of attention -method: examine reaction time under full attention and divided attention for both speakers -evaluation: if theory is correct second language speakers compared to first language speakers should have shorter reaction times under divided attention condition than full attention condition CHAPTER 5 PSYCHOLOGICAL MEASUREMENT -measurement: process that gives a score that represent characteristic of individuals (ex: working memory and digit span task) -construct: variable that requires measurement beyond simple measures like naturalistic observation, involves internal processes (ex: personality traits, emotional states) Definitions: conceptual definition: -are behaviors and characteristics that make up the construct and how they relate to other variables (what the concept means) -ex: construct=psychological well being, conceptual definition= self acceptance operational definition: -description of how to measure a variable (how to measure concept) -ex: self report measures, satisfaction with life scale Types of measures: behavioral measures: participants behavior is observed and measured, observation can occur in lab or in natural setting, to make sense of observation we need a list of behaviors based on definition physiological measures: blood pressure, skin response, hormone levels, electrical and blood activity from brain (EEG recordings) Levels of measures: nominal level: -scores assigned to category labels, no specific order -ex: sex male or female ordinal level: -scores represent a rank order -ex: satisfaction scales, very satisfied, satisfied, very dissatisfied interval level: -numerical, intervals have same difference, no true zero point (0 doesn't mean absence of something) -ex: temperature difference between 20-30 is same as to 40-50, psychology example is IQ levels ratio level: -numerical scale yet has a true zero (if 0 it is absent) -ex: reaction time, memory score Reliability: -consistency of a measure... over time (test-retest reliability): same test measured more than once, used when construct is stable across time such as IQ across items (internal reliability): correlation between answers to items on the same test, split-half method (items on the test are divided in two, measure correlation between) across different researchers (inter-rater reliability): different researchers similarity on judgements Validity: -extent to which the construct is correctly measured face validity: ratings of whether the items appear to measure the construct (weak form of validity as it relies on people's beliefs on construct), does it test what its designed to test content validity: assessment is content appropriate for what is being measured Criterion validity: -extent of correlation between current test and other variable expected to be correlated with the construct, how well the test performs in relation to an external standard concurrent validity: -extent of the agreement between two measures or assessments taken at the same time (existing validated test vs new found test) -whether the new test produces results comparable to an accepted standard -ex: a new test to measure depression should correlate with Beck Depression Inventory convergent validity: -whether the test correlates with measures that it theoretically should correlate with, even if they aren't identical measures -ex: a new test measuring social anxiety should correlate strongly with other constructs related to social anxiety such as shyness or fear of negative evaluation discriminant validity: test does not correlate with unrelated variables (does the test differ from unrelated tests) predictive validity: variables are obtained after the test, (does the test predict later performance on a related criterion) Evaluate measure: -even if a reliable and valid measure was used, make sure that it was valid and reliable for the current sample -if measure is not valid and reliable need to revise the test, if it is an adapted test need to see whether the items are in line with the current population CHAPTER 6 EXPERIMENTAL RESEARCH -causal relationships, iv manipulated, levels of iv called conditions, dv is measured Internal validity (Manipulation of IV) -extent that change in DV is due solely on IV -active intervention from researcher, important direct manipulation to avoid other differences between participants -obtaining self-report data is not a manipulation, some variables such as age cannot be manipulated -manipulation check required to see if you really manipulated the IV, ex: questionnaire after a manipulation -control extraneous variables: other variables that could impact the result, need to be controlled (kept constant), ex: daytime, can become confounding variables -confounding variables: distorts other variables in question External validity -extent that the study conditions could be generalized to the wider community -study conditions similar to the conditions that we want to generalize to Construct validity -quality of the experiment's manipulations -operational definition and assessment conditions that address the research question is important -the extent to which your test or measure accurately assess what it's supposed to -factor analysis Statistical validity -do the statistics used support the results, conclusion -need an adequate sample size (dependent on the study topic) and correct statistical tests -power analysis: determines the number of participants required -number of conditions and participants will determine the size of the effect Types of control conditions: -no treatment control condition: issue with placebo effects, participant expectation to improve cannot be ruled out -placebo control condition (active control condition): receive treatment yet no active ingredient -waitlist control condition: people told they will be taken to the treatment group once space is freed, expectation to improve -current treatment vs past effective treatment condition: improvement in both groups expected Between Subject Design -each participant is tested in only one condition (either a or b) -participants should be matched on extraneous variables so that they don't become confounding variables (ex: both groups participants should have the similar age) -random assignment, each participant has equal chance of ending up in condition a or b -block randomization: all conditions occur once before any of them are repeated, dividing participants to blocks and randomly assigning them to groups within each block, conditions appear in random order RCT (randomized clinical trials): -treatment condition: condition that improves DV -control condition: condition which the treatment is tested against Within Subject Design -each participant tested under all conditions (both a and b) -advantage: groups not needed to be matched on extraneous conditions, smaller sample size -disadvantage: biased reporting (influence of prior conditions may not be feasible due to IV) -participants behavior in later condition influenced due to prior condition -Practice effect: better performance on later condition, participants practice doing it the first time or they figure out a strategy -Fatigue effect: worse performance on later condition, tired or bored doing it in the first time -Context effect: change in perception of task with the change of condition, demand characteristics (cues that indicate the research objective to participants) change behavior \- counterbalancing to overcame carryover effects performed, each participant exposed to the conditions in random order, latin design square can be used (A B C D, B C D A, C D A B, D A C B) -recruiting participants: in field experiments, need to select the participants Standardizing the procedure -experimenter could create the extraneous variables (ex: variations in how the researcher says the instructions can change participant responses) -experimenter expectancy effect: if we expect the experimental group to recall more words than the control group we may encourage them more (smile more) -computerized programs: rather than reading out the words computerized word presentation -double blind: procedure to reduce experimenter expectancy effect, researcher does not know which treatment the participants are receiving -assign id numbers to participants, keep a record of testing Pilot tests: -small scale tests to ensure the procedure works -formal or informal participant recruitment -done to validate the instructions of a study, duration, equipment, can participants guess the hypothesis CHAPTER 7 NONEXPERIMENTAL RESEARCH -lack of manipulation of IV and random assignments to groups or conditions -used when, the researcher does not ask about causal relationship, manipulation of the IV is unethical or impossible (ex: does damage to the brain cause memory loss) Single variable research: -single variable in the study -ex: how many words do students recall -such a study cannot tell us under which conditions students recall better Correlational research: -interest in the relationship between variables -do not use the term IV as there is no manipulation of a variable -ex: is sleep correlated with academic performance Naturalistic observation: -observation of behavior in an environment that is expected to occur -ex: observation in playing grounds -issues: sampling (when, where and conditions that the observations will take place), measurement (what are the behaviors to be observed) -coding: categorize the behaviors as a set of target behaviors Archival data: -use of data that has been collected for another purpose -ex: record of car accidents to determine whether the age of driver matters -content analysis can be done to code previous data Quasi-experimental research (nonequivalent groups design): -there is a manipulation of IV yet no random assignments to groups -ex: morning class on Monday and Tuesday compared (participants are already in groups class 1 or class 2) -groups may differ on factors that could affect the DV Quasi-experimental research (pre-test/post-test design): -DV is measured once before and after the treatment -ex: effect of exercise on stress levels, stress levels assessed before and after exercise Other reasons for change in DV: -history: other events may have happened between pre and post test -maturation: growth and learning overtime could lead to change -regression to the mean: tests have true and an error component (error component may reflect factors such as fatigue), conditions change on second test thus error component changes regressing the score back to mean, if a random variable is extreme at first the next sampling of the same random variable will be likely closer to the mean -spontaneous remission: psychological problems improve over time without treatment Interrupted time-series design: -set of measurements taken at intervals over a period of time -these measurements occur before and after treatment -difference with pre-post test design is that, interrupted time series design have multiple tests for multiple intervals not only pre test post test Combination design (two types of Quasi experiment combined): -experimental-control group/pre-post test -multiple measures could be included -may included additional groups -complex, broader analysis of relationship and variables not just a change due to an intervention -both groups are tested prior and after the treatment -if there is a random assignment to groups this becomes a true experimental design Qualitative research: -less focused research question -important factor is to understand the experience of participants -non numerical data obtained sometimes via interviews, data described using non statistical techniques -ex: can you explain your experiences during adapting to college life -thematic analysis (analyzing qualitative data) -large data from small number of individuals -unstructured interviews: participants talk about what they think is important -structured interview: set of questions to find out specific information -semi structured interview: both structured and unstructured technique -focus groups: small group of participants who partake in the interview together for a particular topic, important to be aware of how people react in a group -participant observation: researchers become active participants in the group, information could be obtained if you go unidentified as the researcher Grounded theory: -start with data and then develop a theory or interpretation about the data -identify ideas that are repeated -ideas categorized into themes -theoretical narrative, interpretation about the data is written (subjective experiences of participants) Qualitative vs Quantitative research: supporters of quantitative research indicate: -qualitative data lacks objectivity, lacks validity and reliability, does not generalize information outside of the participants interviewed supporters of qualitative research indicate: -quantitative data lacks to capture the richness of human behavior and experience -mixed method research: qualitative research is added to the quantitative research to generate more specific hypothesis, comprehensive understanding -triangulation: cross validate findings by comparing and integrating data from multiple sources, methods, a technique -quantitative: identifies behavior -qualitative: understanding meaning behind mechanisms CHAPTER 8 COMPLEX RESEARCH DESIGNS Measurements of constructs (Multiple dependent variables): ex: does sad music lead to recall of sad memories or cause decreased mood (same constructs) -if null supported either no effect or music did not cause decreased mood -we could include a DV to assess mood (manipulation check) ex: does video games play increase aggressive behavior (different constructs) -use multiple measurements (questionnaire and physiological measure like heart rate) to measure aggression Measurements of constructs (multiple independent variables): ex: the effect of age and divided attention on memory processes -allows the investigation of whether the effect of one IV is dependent on the second IV (interaction) Factorial Design: -each level of one IV (factor) is combined with the other IVs levels, producing all combinations -each combination becomes a condition ex: the effect of age and attention on recall -this is a 2x2 factorial design (2 variables with 2 factors) -age: young vs old (2 factors) -attention: full attention vs divided attention (2 factors) -4 possible combinations ex: the effect of exercise and age on stress -3x2 factorial design -exercise: walking running swimming -age: young vs old -6 conditions/combinations ex: many IVs and levels are possible -such as 3x3x2 (first 3 IVs, second 3 IVs, third level has 2 IVs) -18 conditions -more than 3 IVs with more than 3 levels is unusual (number of conditions too many, too many participants needed) Between subject factorial design: -need to decide whether the IVs will be between or within subject designs ex: effect of attention and exercise on recall -one participant takes one condition only -attention: participants in one condition only full attention vs divided attention -exercise: participant in one condition only walking, running, swimming Within subject factorial design: ex: effect of attention and exercise on recall -one participant takes part in all conditions -attention: all participants peform full attention vs divided attention -exercise: all participants perform swimming, walking, running Mixed factorial design: -one IV is a between subject design the other IV a within subject design ex: effect of attention and exercise on recall -attention: within subject design (all participants perform full attention vs divided attention) -exercise: between subject design (participant only one condition from walking swimming running) -so different people who are walking, swimming, running but all tested under full attention and divided attention conditions Nonmanipulated independent variables: -some IVs are not manipulated, participant variables ex: age, bilingualism, personality trait -between subject design -causal relationship for only the manipulated variable -if there is only one manipulated variable still called an experimental method Main effects and interactions: ex: the effect of exercise and motivation on weight management -main effect: the relationship between IV and DV -main effect 1: effect of exercise on weight management irrespective of motivation -main effect 2: effect of motivation on weight management irrespective of exercise -interaction is when the effect of one IV is dependent on the level of the other IV -the effect of exercise is stronger for those who are highly motivated than those who are not motivated Three types of interaction: 1. Variable 1 only affects one level of variable 2 (exercise only effectful for those with high motivation but not those with low motivation) 2. Variable 1 affects one level of variable 2 more than the other (divided attention affects both younger and older adults but more so those who are older) 3. Variable 1 affects the two levels of variable 2 but in opposite directions (bystanders may decrease performance in introverts, it may increase performance in extroverts) Complex correlational designs -when studies only have nonmanipulated variables, it's a between subject design, no longer experiments but a correlational study ex: effect of age and bilingualism on recall -interaction: the effect of bilingualism is greater for older than younger adults because younger adults are already at their peak cognitively -other correlational studies include measurements of several variables that may be categorical or do not assess relationship among them ex: how does level of mental status relate to functionality in older age (physical strength, fall rates, driving ability, financial management, living at home) -results presented in a correlational matrix which shows the correlation between each variable -could be used to explore causal relationships between variables -statistical control of potential third variables, to show that they do not correlate with the DV so ruled out Multiple regression: -several IVs (x) and one DV (y) -DV is the additive combination of the IVs -b1x1 + b2x2 +... = y -b= how much contribution the IV makes on the DV -for each one unit change in the IV how much change there is in the DV -indicates how much contribution there is by each IV when the other IVs taken into account Factor analysis: -statistical analysis: multiple variables are analyzed and clustered according to their level of correlation -factors are constructs they do not indicate whether participants are high or low on that factor and they are independent of each other -underlying variables are displayed, what they constitute or why they are different is not provided, this is up to the researcher CHAPTER 9 SURVEY METHOD -quantitative and qualitative questions -variables measured via self-report, thoughts, feelings, behavior -participants are called respondents -large samples, random sampling -surveys can be long or short -surveys can be conducted in person, over the phone, email, internet -most are non-experimental can be experimental Constructing Survey Questionaries (Cognitive model): -respondents interpret the item -retrieve information from memory -form a judgement -convert the judgement to an option -edit response if necessary Context effects: -context in which the item appears influences the response -item-order effect: order of the items affects the response (certain items may prime or make some influence more accessible in memory therefore affect the retrieval), rotate questions and response items to eliminate order effect -response options: middle responses seen as being normal or typical for most people, so most people respond with middle option Types of Items: Open ended questions: -respondents answer with a question -ex: what are benefits of exercise? -used when research question is vague -no response option so no order issues -not bias responses -time consuming for respondents, more difficult to analyze Close ended questions: -questions and responses to choose from -ex: how happy are you with your community? (0 not at all, 10 very) -more difficult to formulate as they require responses as well -quick and easy to complete for respondents, easier to analyze -more common form -categorical variables: sex, ethnicity, categories listed; participants choose -quantitative variables: ratings scales used, ordered set of responses that participants choose from (3 to 11, 5 to 7 ratings ranges, most popular) Scales: Unipolar scales: -5-point scales used (never, rarely, sometimes, usually, always), goes from never to an extreme Bipolar scales: 7-point scales used (very much, somewhat, slightly, neutral, dislike slightly, dislike somewhat, dislike very much), goes from an extreme to its opposite extreme, 2 poles -branching: first ask do you like or dislike, then provide options for degree of like -use verbal labels rather than numerical, later convert the verbal to numerical for analysis Likert scales: -similar to rating scales but are used to measure peoples attitudes towards others, groups, ideas -includes reverse scoring -measures how much people agree or disagree with a set of statements (strongly disagree, disagree, neutral, agree, strongly agree) Writing effective items: BRUSO= brief, relevant, unambiguous, specific, objective -clear, understandable, not to lengthy questions -if a question such as ethnicity is not relevant to the study, avoid -avoid the use of two ideas requiring one response (are you happy with school and the subjects you have chosen) -don't reveal your own ideas or have leading questions -cannot present all options, use the other option, let participants add their response -if the person can belong to more than one category, indicate this item in the instruction (select all categories that apply) -certain concepts such as attractiveness, pain and likelihood numerical rating can be better understood by 0-10 scale -include a midpoint so that people can choose an appropriate answer when others don't fit Formatting the questionnaire: -have an introduction: state brief purpose of survey and importance, provide information about the sponsor of the survey, acknowledge the importance of their participation, indicate any incentives (gift card maybe) for participation -informed consent: describe to participants the aim of the survey, the topics that will be questioned, amount of time it'll take, withdrawal, confidentiality, written informed consent are not usually used in survey method -should have instructions on how to complete the survey -group items by topic or type -demographic items presented last (most easy to answer when participants may become fatigued or bored) -end by thanking respondents for their time and how much you appreciate their contribution Sampling: -sample size, confidence, budget -non respondents important (if a certain characteristic is not responding?) -ensure a high response rate (incentives, face to face, reminders, short and simple survey) Probability sampling: -specifies the probability that each member of the population will be selected to the sample -must have a specific definition of the population (ex: people with PTSD) -must have a sampling frame (reach or list of the target population, diagnosis required so hospital and clinical records) Simple random sampling: -place all the members into a hat, mix and draw your sample, each person has a random chance of being selected -or computerized draw Stratified random sampling: -population divided into subgroups "strata" (based on demographics usually) and a random sample is taken from each stratum -proportion of respondents in sample match the proportion of people in the population \- ex: survey regarding the satisfaction of academics with facilities at the university. If 10% of academics are from aboard and there are 1000 academics at the university then you would need to include 100 academics from aboard for their views to be represented. Cluster sampling: -large clusters of individuals randomly sampled and then individuals from those clusters are further randomly selected -ex: how happy are students in Istanbul -randomly select some universities in Istanbul (clusters) and choose samples randomly from those universities -makes the sampling process effective when face to face interviews are done, would need to travel to fewer universities CHAPTER 10 SINGLE SUBJECT RESEARCH -behavior of each of a small sample of participants -does not mean only the study of a single participant -could be group research (testing the effects at the group level) -study includes experimental manipulation and control, structured data and quantitative data analysis -differs from case studies which is a detailed description of an individual (both qualitative and quantitative analysis) -case studies do not allow for causal relationships Assumptions of single subject research: -single subject research takes into account each individuals behavior -group research can hide individual differences in the behavior of participants -important to reveal causal relationships, therefore need a DV, manipulation of IV and control of variables (good internal validity) -important study effects that have biological or social importance, implementation to real life contexts (social validity) ex: Pavlov's classical conditioning research -Skinner measured non human objects (rats and pigeons) to see how rewards and punishment influence behavior, experimental analysis behavior -using this approach to conduct applied research with humans, applied behavior analysis -can be applied to apply theoretical principles Features of single subject research: -DV measured at regular intervals over time divided in phases or conditions -change from one phase to the other is not fixed but rather dependent on the participant behavior -once the participant behavior becomes consistent in one condition then there is a move to the next condition (steady state strategy) Reversal designs ABA design -Phase A is the baseline for the DV, level of responding before the treatment -when steady state responding is reached phase B begins -Phase B, the treatment is introduced -once a steady state is reached in terms of behavior a consistent behavior pattern and change from the baseline is reached, the phase ends -treatment removed and DV assessed again (A phase) until behavior comes steady again -if behavior increases or decreases consistently during the treatment phase and levels off during the baseline condition then we can infer the changes in behavior to the treatment -ABA design is important than just a AB design as it clarifies whether changes in the DV is a result of the treatment -important that the treatment does not cause a permanent effect otherwise the behavior wont return to the baseline condition Multiple treatment design (ABCA) -Phase A baseline -Phase B treatment 1 (ex: positive attention for studying) -Phase C treatment 2 (ex: punishment for not studying) -Phase A baseline -ABCA or ACBA reverse the order of the treatment to eliminate carryover effect Alternating treatments design -alternate between the treatments on short intervals to compare the effect of the treatment Multiple baseline design -issues with reversal design (is it ethical to remove a treatment that is working? after a treatment the behavior may not return to baseline condition) -AB design (treatment given at a different time for each participant), this avoids any change in behavior to be a coincidence -multiple baselines for multiple DVs -treatment introduced at a different time for each DV -baselines for both DVs established -treatment for one DV introduced and for other DV at a different time -multiple baselines for same participants but at different settings Ex: baseline for reading at home and at school -treatment is positive attention -treatment provided at different times for reading at home and school -if after the treatment reading increases at both home and school than we can say that the treatment was effective in changing behavior Data analysis -visual inspection: plotting the individuals' data and making a judgement about whether the IV had an effect on DV -inferential statistics (SD, mean, Pearson's r) not always used -amount of change in the DV from one condition to other -latency: the amount of time it takes for the behavior to change within the condition (short change indicates that the treatment was the account) -statistical procedures can be used, t-test or analysis or variance (means and SDs examined for each participant under conditions) -percentage of nonoverlapping data (percentage of all data that do not overlap in the conditions), the greater the percentage the greater the effect of treatment Debate -visual inspection is not an adequate way of determining the effect of treatment -visual inspection is not sensitive to detect weak effects -visual inspection may be unreliable, judgements may vary for each researcher -judgements may be hard to compare across studies -use of the statistical procedures help to compare -external validity is an issue, replication of effects needed -groups means from groups research may be misleading, treatment may have a positive effect for some and negative for others (mean will become null 0 indicating no effect) -ecological validity is important (does the lab conditions generalize to real life) CHAPTER 11 DESCRIPTIVE STATISTICS Single Variables: -statistics that help to summarize and describe the data -distribution: how the scores in data are distributed across the variables -frequency tables: table that shows the frequency for the values of variables (frequency scores go from most to the lowest), for wide range of values grouped frequency tables -histograms: graphic that represent the distribution (x-variable, y-frequency) Distribution Shapes: -bell curve (peak at the middle and lower two ends): unimodal -some distributions are bimodal, two peaks -shape of the distribution can be symmetrical or skewed -symmetrical: has mirror images on left and right -skewed (negatively or positively):negative= majority of data is on the left, positive= majority of data is on the right -outlier: extreme score (higher or lower then other scores) ,can be real or error in data Central tendency: -point where majority of the data is found -mean: sum of all scores divided by number of data points, becomes problematic for skewed data better to use median -median: middle score- organize the scores from lowest to highest -mode: most frequent score in distribution Variability: -extent that scores differ from the average -low variability: scores closer to average -high variability: scores spread across -range: measure of variability, difference between the highest and the lowest scores, may not be a reliable measure when there are outliers (indicating great variability) -standard deviation: common measure of variability, average distance between the scores and the mean to calculate: -find the difference between each score and mean -square each difference -variance: find mean of the squared differences, divide total by the total number of scores -find the square root of the variance Percentile ranks: -percentage of scores that are higher or lower than the pre-determined score -number of scores lower or higher than the pre-determined score converted to the percentage Z score: -indicates how far away the mean from each score is -difference between score and the mean/ SD -z=(X-Mean)/SD Statistical relationships: -important that the SD is similar among conditions and group -effect size: strength of the relationships -Cohen's d: difference between two means/SD , show how much the two groups differ in terms of SD, d=1 indicates that the groups differ by 1 SD, 0.20(low), 0.50(moderate), 0.80 (high) Correlations: -can be presented as line or scatterplot -linear relationships (positive or negative) -nonlinear relationships (curved line or U shape) ex: stress levels over the semester -Pearson's r: describes the strength and direction of the relationship for linear relationships,.10 (low), 0.30 (moderate), 0.50 (high), the sign of the r indicates direction, -1 or +1 To calculate: -transform all scores to z scores (subtract mean of X from score / SD, subtract mean of Y from score / SD) -multiple the two z scores (scores for each individual / N) CHAPTER 12 INFERENTIAL STATISTICS Null hypothesis testing: -sampling error: random variability in sample statistics -statistical relationship in sample does not always mean there will be a statistical relationship in the population -any significant difference between variables could be due to sampling error -null hypothesis testing: helps to determine whether significant differences are due to real differences in the population or sampling error -null hypothesis: there is no relationship in the population, relationship due to sampling error -alternative hypothesis: there is a relationship in the population -if the null is true no relationship in the sample Under the assumption that null is true: -if the sample relationship is unlikely, we reject the null hypothesis and keep the alternative hypothesis -if the sample relationship is likely we retain the null hypothesis P value: -to determine the likelihood of whether null hypothesis is true we use a probability, p value -a low p value means, sample relationship would be unlikely if the null were true, so we reject the null hypothesis and keep the alternative -a high p value means, sample relationship would be likely if the null were true, we retain null hypothesis -to determine whether p value is low or high we use an alpha level of 0.05 -p value lower then 0.05 means that there is less than %5 chance that the extreme result would occur if null were true thus the null is rejected -p value higher then 0.05 means that there is higher than %5 chance that the extreme result would occur if null were true thus the null is retained Sample size and strength of relationship: -p value is dependent on sample size and strength of the relationship between variables -the greater the sample size and relationship strength, more likely that we reject the null Ex: compare a group of 500 men and 500 women on reaction time, effect size d=0.50 -we would say that it is unlikely that this relationship would occur if null were true hence, we reject the null Ex: compare a group of 5 men and 5 women on reaction time effect size d=0.10 -we would say that it is likely that this relationship would occur if null were true hence we retain the null -if there is a strong relationship even a small sample size is adequate, if the relationship is weak a larger sample size needed -statistically significant does not mean a strong relationship, even a weak relationship can be statistically significant -practical significance is the important result for the society T test: -comparison of two means -ex: testing if the average score of a class is different from 70 one sample T test: -sample mean compared to hypothetical population mean -t value important as it shows the amount of deviation of sample mean from hypothesized value in standard error -t= sample mean subtracted by population mean / (SD / Squared N) N= sample size critical values: -decide whether there are differences between sample and population mean -decide on degrees of freedom N-1 -check the critical values for the df and if the t value falls outside the critical values then you can indicate that the score is extreme (p outside the %5), reject the null hypothesis -outside the 5% means: two tailed and one tailed test -for two tailed p value falls in %2.5 of %5 on either side of distribution -for one tailed p value falls in extreme 5% in one direction -how to decide which to use: a specific direction is a one tailed test (if drug A is better then drug B), both direction is a two tailed test (if drug A is better or worse than drug B) -one tailed hypothesis: this fertilizer increases crop yield (testing for an increase) -two tailed hypothesis: this fertilizer affects the crop yield (testing for either an increase or decrease) Dependent samples t test: -comparison of means for two times of testing the same sample -ex: testing if a training program improved employee productivity by comparing scores before and after the program -pre test and post test, within subject design -difference between the two scores compared to hypothesized value (mean=0) -t test formula similar to one sample t test, replace the sample mean with difference score mean -similar to one sample t test, use p value to determine difference or the t value and critical value Independent samples t test: -comparison of means for two different samples -ex: testing if the average scores of two different classes are significantly different -between subject design, or preexisting groups -null hypothesis: means of both populations are the same -alternative hypothesis: means differ -df (degree of freedom): N-2 -t test formula similar to one sample t test, two means, sd, n (sample) The analysis of variance ANOVA: -more than two groups are compared -one way ANOVA between subject design (comparing test scores among students in three different teaching methods) -two way ANOVA can test for interaction effect, effect of two IVs on a DV (analyzing how teaching method and study time affect test scores) -repeated measure ANOVA, within subject design for more then two condition (testing the effectiveness of a drug over three points, baseline-1 month-3 month) -test statistics is called F ratio of two population -f statistics is the ratio of MSB (mean square between group) to MSW (mean square within group), MSB/MSW -between group variability: variability due to differences among group means, within group variability: variability within each group due to random error -Post hoc comparison, follow up to determine where the difference between groups are Basic null hypothesis test: -null: all means in population are equal, alternative: means are not the same -If f statistics are large reject the null