PSY-0032 Experimental Psychology: Study Guide #1 - Scientific Theory, Hypothesis Testing, Non-experimental Designs, Reliability & Validity, Measurement Errors
PSY-0032 Experimental Psychology: Study Guide #1 - Scientific Theory, Hypothesis Testing, Non-experimental Designs, Reliability & Validity, Measurement Errors PSY-0032
Popular in Experimental Psychology
Popular in Psychlogy
This 8 page Study Guide was uploaded by Amy Bu on Thursday October 8, 2015. The Study Guide belongs to PSY-0032 at Tufts University taught by Dr. Sam Sommers in Summer 2015. Since its upload, it has received 181 views. For similar materials see Experimental Psychology in Psychlogy at Tufts University.
Reviews for PSY-0032 Experimental Psychology: Study Guide #1 - Scientific Theory, Hypothesis Testing, Non-experimental Designs, Reliability & Validity, Measurement Errors
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 10/08/15
PSY 0032 Study Guide Exam 1 Psychology is based on observation and measurement Not everything testable makes it a good science 4 Principles of Psychology 0 1 Determinism Events have logical causes which produce effects Behaviors are shaped by multiple factors cannot lightly say that only X causes Y O 2 Empiricism Observation is key to learning What you see is what you get 0 3 Parsimony When there are competing explanations the simpler the better Simple theories rely on fewer assumptions have fewer weak links and are easy to explain 0 4 Testability Theories must be testable or else they39re moot This is based on logical positivism which states that knowledge must be based on what we can see with certainty Testability implies falsifiability which is more important any theory can have some degree of supporting evidence e g horoscopes the Positive Test Bias Purposes and Types of Research 0 1 Descriptive studies describe the variables 0 2 Correlational studies predict one variable using another 0 3 Experimental studies explain the relationship between variables and why Conceptual Variables general ideas and constructs eg depression mood behavior Operational Variables things that need to be measured one conceptual variable may have several operational definitions Converging agreeing operational definitions make a strong argument Ways of Knowing O 1 Intuition gut feeling I Con often wrong 0 2 Logic reasoning better than intuition I Con assumptions change over time may also be influenced by bias 0 3 Authority the wisdom of experts Con depends on the reliability of said experts May lead to blind following or misplaced trust in people who only appear to be experts Requires verification of their expertise O 4 Observation knowledge through empirical tests preferred I Con some questions are difficult to test due to ethical or practical limits Researcher biases PSY 0032 Study Guide Exam 1 and expectations may color results Laws Universal statements about the nature of things eg the laws of physics However the field of psychology studies subjective experiences and humans are creative and independent beings Hence laws are not as applicable to psychology most psychologists do not seek to find or establish universal laws Theories statements regarding the relationships between variables that have boundary conditions to their situation unlike laws which are universal Good theories should adhere to the 4 Principles of Psych There can be multiple theories about one topic that are all partially true but some theories will be better than others 0 eg Boundary conditions for relationship between similarity 8 attraction for what genders is the relationship true For what type of attraction For what type of gender pairing For what length or quality of relationship Etc Hypotheses specific predictions derived from theories that test theories often under conditions suggested by the theory They operationalize theories to make them specific and testable 2 Types of Reasoning are used to develop theories and hypotheses O 1 Inductive from specific to general They generate theories based on multiple observations I However how many observations are sufficient How can one be sure the theory is true Hence there is no proof for a theory but only support until falsified O 2 Deductive from general to specific They lead to tests of conclusions based on induction I If the hypothesis appears to be true the theory continues to stand If the hypothesis is false revise or discard the theory Reasoning OBSERVATION induction gt THEORY deduction gt HYPOTHESES 3 Strategies for Hypothesis Testing 0 1 Validation attempt to gather information to support the hypothesis 0 2 Falsification attempt to refute the hypothesis 0 3 Qualification using both validation and falsification attempt to identify the limits and boundary conditions for the hypothesis Positive Test Bias confirmation bias tendency to gravitate towards validating evidence Formulating a hypothesis there are 3 possible competing hypotheses for any question For PSY 0032 Study Guide Exam 1 example if someone said Students who use study guides obtain better grades 0 1 Students who use study guides have higher GPAs O 2 Students who use study guides have lower GPAs O 3 There is no difference in GPA between students who do and do not use study guides 2 Types of Hypotheses O 1 Null there is no relationship between variables no difference between groups 0 2 Alternative there is some relationship difference I The null hypothesis is the default assumption until proven otherwise by data that supports the alternative hypothesis The alt hypothesis is the one of interest to the researcher Inferential Statistics mathematical way to test hypotheses e g whether there is a reliable relationship between variables whether effects are meaningful or purely coincidental Significance Testing pvalues where p the probability that findings are by random chance The lower the p the less likely that findings were by chance the better 0 In Psychology p S 005 means the null hypothesis can be rejected If p gt 005 fail to reject null hypothesis reject alternative hypothesis This pvalue of 005 is called the alpha level Alpha Level alevel is the distance between means of conditions e g mean GPA of students who use study guides versus students who don39t use study guides Lowering the pValue O The greater the distance between means the stronger the relationship is between variables 0 The more participants or number of observations the more convincing the relationship Type I Error incorrectly reject null null is true incorrectly accepting alternative 0 e g accidentally eating poisoned food food your alternative hypothesis I Causes researcher to report a finding that does not eXist This can be fixed by lowering the alphalevel however that increases the risk of performing a Type 11 Error Type II Error incorrectly reject alternative alternative is true incorrectly accepting null 0 e g accidentally throwing away good food food your alternative hypothesis I While there is a reliable effect the researcher does not know it Effect Size Asks whether the effect is important and what size the effect is Not a measure of statistical significant but of practical importance Recently added as a requirement in the field PSY 0032 Study Guide Exam 1 Descriptive Studies singlevariable describe nature of variablecondition typically used at beginning stages of research Provides snapshot not analysis 0 eg case studies census population surveys political polls market research the Milgram Study baseball statistics Case studies observe experience of particular person or group very helpful for clinical psychologists and neuroscientists and can be informative if conducted scientifically 0 Pros insight to hardtostudy phenomena analyze real world data 0 Cons very few participants often only one No operational definitions comparison groups or analyses 4 Sampling Methods methods important to increase accuracy about population 0 1 Random everyone in the population has an equal chance of being selected requires complete access to entire population 0 2 Cluster create list of locations randomly sample within clusters e g list of dorms or cities to randomly sample residents 0 3 Stratified when interested in specific subgroups randomly sample within said subgroups e g randomly sample within majors race political orientations O 4 Convenience use anyone that is accessible Problematic for descriptive studies but less problematic for correlational or experimental studies Sampling Error discrepancy between sample response and actual population mean also known as the margin of error 0 Notation i 2 means there is 95 certainty the number is within 2 or 2 3 SelfReport Limitations 0 1 Knowledge Limitation participants may not know the answer 0 2 Conscious Concerns participants may not want to tell you the truth 0 3 Nonconscious Concerns participants may be selfdelusional about the topic 3 Ways to Overcome O 1 Bogus Pipeline convince participants you can detect lies they become less likely to lie 0 2 Reassure participants about anonymity O 3 Include other types of measures to triangulate data e g physiology behavior PSY 0032 Study Guide Exam 1 Correlational Studies predict one variable based on another e g is there a relationship between owning pets and having good health DOES NOT EXPLAIN WHY or IMPLY CAUSATION O Null hypothesis no relationship Alternative hypothesis e g there is a positive correlation O 1 Find appropriate sample measure of pets measure happiness 0 2 Find descriptives mean standards deviation range 0 3 Graph data eg scatter plot 0 Sample result As predicted X was positively correlated with Y r 38 049 p 0001 Pros 1 you can predict variables based on other variables 2 you can correlate anything Cons 1 doesn39t address causality or why they39re related 2 always possibility of confounds Pearson39s r ranges from 1 to 1 tells the strength and direction of relationship 0 Only tests for linear relationships not informative for other types e g exponential Positive r indicates positive correlation negative r indicates negative correlation r of 0 indicates no correlation The closer to 1 the stronger the relationship the closer to 0 the weaker Correlation matrix analyzes data Line of Best Fit positive correlation indicated with positive gradient vice versa The key to prediction using the line gradient you can predict one variable based on the other There are at least 3 Possible Reasons for a Correlation O 1 X causes Y O 2 Y causes X 0 3 Z third variable affects both X and Y Confounds unmeasured variables that influence study results Z O 1 Person Confound eg personality habits O 2 Environmental Confound eg income urbanrural environment 0 3 Operational Confound inadvertently measured B instead of A e g survey item I find it difficult to get out of bed measures depression or physical wellnessability Questionnaires 8 Interviews rely on selfreport are cheap and easy Order of questions and response scales may influence responses Archival Analysis using preeXisting data to analyze eg political speech transcriptions birth PSY 0032 Study Guide Exam 1 records website hits High realism but may be subjectively interpreted with positive test bias Observational Studies records real behavior in natural environments best when unobtrusive However there may be an observerinteraction effect act of being observed changes participants39 behaviors Ethical concerns if participants do not know they are being observed Often requires extensive behavioral coding to turn observations into analyzable data Restriction of range r magnitude compromised when responses don39t cover the full range of possible values e g sumo wrestlers for weight Tufts students for IQ may show no correlation when there is one because you cannot see the big picture Validity accuracy does the variable measure what it should Are the operational definitions good e g watch that tells good time has high validity watch that is slowfast has low validity broken watch is valid twice daily Reliability consistency is there consistency and consensus in what the variable measures e g the watch that is consistently slowfast by 20 mins has high reliability 4 Types of Construct Validity O 1 Face validity does it appear to measure variable of interest I Con may be wrong is very subjective obvious variables can lead to reactance O 2 Content validity does variable cover whole range of thing e g low content validity if an exam covers only half of material learned or if mood study only asks about positive mood O 3 Convergent validity does variable correlate with other existing variables that measure the same thing Do they produce similar results 0 4 Discriminant validity is variable unrelated to other variables measuring other similar things Does it discriminate between thing of interest and other similar things Internal Validity extent to which we can confidently conclude causality that one variable causes the other validity in the lab setting where the study was conducted External Validity extent to which the study is generalizable to other situations outside this study or lab environment using other participants environments or stimuli etc 3 Types of Reliability are measurements consistent or repeatable O 1 Internal Reliability consistency within scalemeasure Do the items on the scale correlate with one another as they should Expressed as alpha 0 PSY 0032 Study Guide Exam 1 O O 2 Interrater Reliability consistency of codersjudges39 agreement on subjective ratings High interrater reliability means they consistently agree Used especially for openended measures Expressed as kappa K 3 TestRetest Reliability consistency across time whether participants get similar results on the measure if they take the test again I Problem Sometimes people will change their answers learn from the first test or get bored and pay less attention when taking it subsequent times etc More is better more validity reliability trials items on a survey participantsetc 2 Measurement Errors 0 1 Random Error chance uctuations in measurement tends to cancel out over time e g participants circling the wrong item coming into labs with different moods computer error range or 05 seconds Poses small threat to validity but bigger threat to reliability 2 Systematic Error nonrandom fluctuations in measurement may not affect reliability but severely affects validity e g participants uniformly underreport drug abuse thermometer or clock consistently off participants all in altered mood 4 Measurement Scales commonly used in research the type used determines the statistical analysis and conclusions that can follow In increasing order of sophistication O 1 Nominal categorical number assignments that are meaningless used only for classification eg in a survey 1 college 2 high school 3 other Frequencies and betweengroup comparisons are meaningful descriptives eg means are nonsensical 2 Ordinal order is meaningful but their exact values aren39t intervals between numbers not standardized Does not inform relative distance eg high school class rank 1st2nd3rd place in competitions 3 Interval differencedistance between numbers standardized but no true zero e g not possible to score zero on the SAT Fahrenheit temperature readings zeros mean different things across scales eg 0 C is different from 0 F 2X is not twice of 1X 4 Ratio ordered standardized difference with a true zero e g height weight age 2X is twice of 1X eg ZOin is twice lOin Zero means the same thing across different units eg 0 cm 0 inches Mathematical analyses are possible PSY 0032 Study Guide Exam 1 How to determine which scales are used for e g Temperature in O a Is 58 F warmer than 57 F gt yes numbers are meaningful not nominal O b Is difference between 58 and 57 F the same as between 13 and 12 F gt yes standardized difference not ordinal O c Is 100 F twice as hot as 50 F gt 32 F is freezing not twice as warm not ratio 0 Answer Interval
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'