Exam 2 Review Guide 1 You should be able to identify experimental and non experimental designs and the limitations of each Correlational observe variables is there a relationship Limitations Used when third variable problem directionality problem gathering data in early stages of research when manipulating and independent variable is impossible unethical studying the relationship between two naturally occurring variables for the purpose of prediction Experimental manipulate one or more variables what effect does this manipulation have on behavior we want to determine causality not just whether or not a relationship exists Strengths Limitations identification of causal relationships among variables not possible with correlational research cant use experimental method if can t manipulate variables tight control over extraneous variables limits generality of results 2 You should be able to specify and identify why causal conclusions cannot be drawn from correlational research no variables are manipulated 2 or more variables are measured and a relationship is established Because scientists do not directly manipulate variables no causal conclusions can be drawn you should not infer causality from correlation data for 2 reasons third variable problem there may be an unmeasured variable that actually causes variables to covary change together directionality problem not always possible to specify the direction in which a causal arrow points 3 You should be able to define and identify independent variables dependent variables predictor variables and criterion variables Independent variables experimenter exposes participants to different conditions independent since they are chosen manipulated by the experimenter and don t depend on the behavior of the participant Dependent variables the behavior or the participant measured by the experimenter dependent since value depends on the treatment given Predictor variables a variable that can be used to predict the value of another variable Criterion variables In studies of prediction the variable or score that is predicted from predictor variables the best obtainable measure of the constant that is to be predicted 4 You should know about internal and external validity and the threats to each also called confounds What does it mean for an experimental design to have internal validity What does it mean for a correlational design to have internal validity Internal validity the degree to which your design tests what it was intended to test Experimental show that variation in the dependent variable could only be accounted for by variation in the independent Correlational show that changes in the value of the criterion variable are solely related to changes in the value of the variable predictor variable Threats to internal validity History Maturation Testing Instrumentation Statistical Regression Biased selection sample and research setting Threatened by highly controlled lab setting restricted populations pretests demand characteristics experimenter bias subject selection bias External validity the degree to which results generalize beyond your 5 You should be able to distinguish between experimental non experimental quasi experimental designs Experimental we want to determine causality not whether or not a relationship exists that s Correlational one or more IV is manipulated one or more DV is collected if different levels of the IV treatments change the values of the dependent measure behavior you have established a causal link control over extraneous variables some experiments have experimental and control groups but some might not have a control group training Quasi Experimental involves finding naturally occurring groups and determining if they differ in any way IE men vs women uses multiple groups or multiple waves of measurement NO RANDOM ASSIGNMENT you can t manipulate one of the factors you are interested in gender but you can manipulate others video game or you didn t make sure groups were equal by random assignment Non experimental observe variables is there a relationship does not involve manipulation of the situation circumstances or experience of the participants Correlational Strategy measure values of different variables is there a statistical association between them Descriptive Strategy observe events or phenomenon to detail or categorize them or chart their course correlational non experimental identifies a potential cause whereas experimental design confirms causal influence 6 What are the factors you need to consider when choosing a measurement Information Yielded A nominal scale yields the least information An ordinal scale adds some crude information Interval and ratio scales yield the most information Statistical Tests Available The statistical tests available for nominal and ordinal data nonparametric are less powerful than those available for interval and ratio data parametric Use the scale that allows you to use the most powerful statistical test Ecological Validity Sometimes your research requires you to use a measure that reflects what people do in the real world e g a nominal guilty not guilty verdict Such scales have ECOLOGICAL VALIDITY When necessary choose an ecologically valid measure even if it means loss of information There are things you can do to gain needed information Include an additional scale e g an interval scale Create a composite scale by combining a nominal and interval scale Incorporate features of ecologically valid scale into a more informative scale 7 How can we determine if a measure is reliable Valid What does it mean for a measure to be reliable Accurate Valid Reliable measures produce similar results when repeated measurements are made under identical conditions how much error will be in the measurement to determine reliability use test retest parallel forms or split half an accurate measure produces results that agree with a known standard a valid measure measures what you intended it to measure Face validity it looks like it should work Content validity how well does the content of a test measure what it is intended to measure Criterion Related how well can a test score be used to infer an individual s performance on some criterion Construct validity is it a valid measure of the construct you are trying to get at measure 8 What are the scales of measurements we ve discussed What determines which scale you should choose What kinds of statistics are appropriate for each Nominal categories have
View Full Document
Unlocking...