Psyc 201 Test 2 Study Questions
Professor Smith correlated the attendance of his students and their performance on the final exam. She found a strong positive correlation showing that the students who attended more class lectures were more likely to score higher on the final exam than students who attended fewer lectures. She concluded that attending class produces better test performance. 4a.) Identify the two problems with Professor Smith's conclusion and explain why her reasoning is faulty in terms of correlation and causation. 4b.) How could you re-design this study so that it tests the causal effects of class attendance on final exam performance? What is your independent variable and its levels? What is your dependent variable? Draw a graph of your predicted results
4a) Confounding variable problem. There is a possibility that students who attend more class lectures spend more time studying than students who don't attend as many lectures, and this causes them to perform better on tests Directionality problem: Perhaps people who have a better test performance are more likely to go to class, and those who score low think it's not worth it and don't attend 4b) You could take the group of students that don't score well and don't attend class, make half of them attend all the classes and half of them only attend half of the classes, then test performance. ● IV: attending class ● DV: test performance
Define reliability. Why is it important in scientific experiment? What are the three types of reliability and give an example of each kind. For observation studies, why is inter-rater reliability important? How do you compute the interrater reliability of two observers?
Consistency in a measurement ● Reliability is the extent to which a test can be done multiple times by the same person, resulting in the same results. Important because it allows us to make sure that any change in performance or behavior is a result of the independent variable, not from the testing methods. ● *Internal reliability* - reliability between each item, all should reflect the same general construct ● *Inter-rater reliability* - reliability between raters ● *test-retest reliability* - if done again, produces same results -consistency over time Inter-rater reliability is important because when observing behavior, you want to make sure that one experimenter would mark the same acts in the same way as another experimenter, or you will have flaws in your experiment Compute interrater reliability by dividing number of items both rated the same by total number of items
What is an experimenter effect? What is a participant effect? How would you control for experimenter and participant expectations?
Experimenter effect: effect that experimenter has on each group of participants. for example, if they know that one group is predicted to do better than another, they may treat them differently. Participant effect: if a participant understands the study and knows which treatment they are going to be given, they may act differently to prove the hypothesis Control with double blind experiment
Identify the "threats" to internal validity and give an example of each kind
History Maturation, Fatigue Practice effects Faulty instrumentation Regression to the mean Non-random selection Participation mortality and attrition
What are the types of observational studies? What are the advantages and disadvantages of naturalistic observations? What are the advantages and disadvantages of structured observations?
Naturalistic observation: observe people in natural setting with no intervening ● advantage: people act how they normally would in day-to-day, useful when looking at a certain group ● disadvantage: can't control groups, not representative sample, no internal validity, low external validity, replication difficult Structured observation: researcher partly or fully configures the setting in which behavior is being observed ● advantage: can control groups ● disadvantage: people may act different, less external validity
What is a placebo and a placebo effect? Why does a placebo effect pose a threat to the internal validity of a study? How would you control for a placebo effect?
Placebo: a sugar pill designed to look like a drug -simulated treatment that lacks any active ingredient or element that should make it effective Placebo effect: when you think you're getting a drug so it makes you behave differently than if you thought you weren't getting it -positive effect of a placebo Threat: people who receive drug may improve because they think its going to help them, not because it actually does. Control: give people a placebo drug and see how much they improve from it, if they improve less than people given actual drug, actual drug works
How would you define a variable?
a changing quality of a construct factor or attribute that can take two or more values or levels and is used to measure psychological constructs
Independent variable
the variable that is manipulated by the experimenters, and it thought to be the reason for change in another variable
dependent variable
the variable that may change in response to manipulation of the independent variable
What is the difference between a "between-groups" design and a "withingroups" (repeated measures) design? What is counterbalancing? Does it pertain to a "between-groups" design or "within-groups" design or both? What are the pro's and con's of a "between-groups" study and a "within-groups" study? If you were designing an experiment examing the effects of marijuana on the ability to do a fine motor task, would you choose a "between-groups" or a "within-groups" design? Why?
● Between-groups design: group of participants randomly assigned to 2 or more conditions ● within-groups: 1 group of participants under same condition -can use counterbalancing to prevent carryover effects -each participant views stimuli in different order ● Pros of between groups: no practice effects, maturation, fatigue, ect ● Pros of within groups: no participant variability ● I would choose a within-groups design because there may be a lot of individual variation in the effects that marijuana has on people, and others may be more used to the drug than others
What is the relationship between reliability and validity? Can an experiment be valid, but not reliable? Can an experiment be reliable, but not valid? Explain the difference and give an example
● Reliability is when something can be measured the same way every time, for example, a clock dings every hour ● Validity is when something is accurate, for example, the clock dinging on the hour is when the actual hour is ● Experiments can be reliable but not valid. for example, if you are weighing people and you find that your scale is off by exactly 5 kg, it is not valid, but it is still reliable. ● For an experiment to be valid, it must also be reliable because it must be reliable to be correct
What is external validity? Please define and explain why testing Psych 100 students may pose a problem to a study's external validity. What is the trade-off between the need for internal validity and the generalization of external validity?
● Testing psych 100 students may be a problem because these students are mostly the same age and have similar educations, and are all interested in taking a psych class suggesting they may know some psychological phenomena. This doesn't generalize well to the rest of the population who could have higher or lower education, vary widely in age, and may have no interest in psychology. ● The need for internal validity is generally more important when trying to find causal explanations, but external validity is useful when looking for correlations in the world
Define validity. What is the difference between construct validity, internal validity, statistical validity and external validity.
● Validity is the extent to which an experiment is accurate and predicts results correctly ● Construct validity: extent to which experiment measures what it is supposed to ● Internal validity: extent to which the independent variable caused a change in the dependent variable ● Statistical validity: extent to which the result did not occur by chance ● External validity: extent to which the experiment can be generalized to the real world
