POS3713 Exam 2 Review 1 What is research design and what are its chief goals a Research design is the strategy that is used to answer a research question Research design specifies the procedures that will be used to answer the targeted research question and also ensures the data that we examine provides an answer to the question b The chief goal of research design is to allow for valid inferences as to causal effects of X on Y Simplified it means the goal is to use data observations we infer in order to explain why X is effecting Y 2 What is an experiment how does it differ from an observational study a Experiments and observational studies are the two practices used to answer a research question b The most significant difference between an experiment and in observational study is in an experiment the research takes some sort of action which manipulates controls the X variable In an observational study the research sits back and observes and has no influence on the outcome 3 What are the strengths and weaknesses of experiments Observational studies Experiments Strengths Have a high internal validity which means researchers are confident that effect of X on Y is real Weaknesses 1 Not all IVs are controllable 2 Ethical considerations prevent some experiments 3 Low external validity Observational Studies Strengths Have a high external validity which means the researchers are confident in the relationship between X Y beyond the sample Weaknesses 1 Internal Validity Could Y be causing X 2 Ecological Fallacy Drawing inaccurate conclusions about individuals based upon aggregates 3 Regression of the Mean Are extreme values of Y due to chance 4 What are field experiments and how do they compare to traditional experiments and ob servational studies a A field experiment is conduct in the real world as apposed to a laboratory A field ex periment tries to influence values of X but may not have complete experimental con trol and or may not be able to use random assignment Because they are conducted in the real world they have a high external validity b They are different than experiments because the researcher does not have complete experimental control and he may not be able to use random assignment 5 What is the difference between random assignment and random selection What type of studio do each of them use a Random assignment is the process of randomly selecting values for X while dom selection is the process of randomly selecting a sample group from the lation for the experiment ran popu b Experiments must have a random assignment but don t always have random selection Random assignment helps ensure all groups are identical in regards to Z variables 6 What are two types of observational studies Why would a researcher choose one over a The two types of observational studies are a Cross Sectional study and a time series another study b Cross sectional studies are used when analyzing many units over a single period in time Holding extraneous factors constant Time Series studies are used when analyz ing one single unit over an extended period of time 7 What are internal and external validity How do they differ What are some common threats to the both of them a Internal validity is the confidence that the effect of X on Y is real External validity is confidence that the effect of X on Y is real beyond the sample b The difference is that internal validity is the confidence that within the specific experi ment being conducted X is in fact effecting Y External validity brings up the validity of the experiment outside of the small scale sample and questions the validity in a real world setting Experiments often have high internal validity and low external while observational studies have low internal and high external c Internal Validity is threatened when their are any differences between the control group and the treatment groups besides X any Z variables External validity is threatened when their are differences between participants and non participants or if their are differences between the setting when it s being studied and it in the real world What does it mean to Control for Z Why is controlling for Z so important in research Z variables are external variables that effect the X and Y variables within an experiment Controlling for Z is research design that eliminates alternative explanations for an observed relationship between X and Y Without control of the Z variable the experiment loses its va lidity 8 a 9 What is ecological fallacy What examples did we use in class Why is it a problem for researchers a Ecological fallacy is the error of drawing incorrect conclusions about individuals based upon aggregate a Alternate definition Using the relationship for groups to infer the same relationship among individuals when no true relationship exists b In class the example used was the Jeffery Frankel quote which compared irresponsible activities of the 5 red est states vs the 5 bluest states The study concluded that the 5 most red states committed top irresponsible activities more than blue states As a re sult the fallacy Frankel concluded was Those people with the least responsible be havior were most likely to vote republican This is false based upon the definition of ecological fallacy b Another example in class An election in Washington concluded with the Demo cratic candidate winning 60 of the votes It was discovered that 1 700 votes that had been cast were fraudulent but no one knew which political party each vote counted towards The Re publicans argued that since the Democrats received 60 of the votes 60 of the fraudulent votes should be thrown out The judge ruled against the Republicans using ecological fallacy as his reasoning c Ecological fallacies are troubling to researchers because they are among the major problems with validating an observational study 10 What is regression to the mean When is it most likely to occur Why is it a problem to researchers Ex Think of the Madden curse A player has a fantastic year based upon skill and also above average luck for the year so they are put on the cover The following year they are just as skillful but they have an average amount of luck and it makes it seem like the player had a bad year but really the season before they were just way above average a Regression of the mean An extreme value of a variable is likely to be followed by a less extreme value that is one that is closer to the mean b Regression of the mean is most likely to occur
View Full Document
Unlocking...