Claims and Validities Consumers of Research More than 2 million youth depressed Do articles in press make sense Is peer reviewed research accurate Must understand variables to answer these questions Variable Something that varies must have at least 2 levels or values Can measure or manipulate variables Measure observe and record Manipulate assign participants to a level Examples of variables that can t be manipulated Mental Illness Childhood abuse Smoking Constant Something that only has one level in a study Ex Gender is a constant if only used men in a study Could potentially vary in another study Conceptual Operational Conceptual definition definition of a variable at an abstract level Operational definition specific way that a concept variable can be measured or Ex Love Stress manipulated in a study variable literally 3 Operational defitions To operationalize ask yourself how did they measure or manipulate Hypothesis Individuals who see an emotionally shocking event will have poorer memory for details than individuals who do not see the event Variables A Seeing the event vs not seeing the event B Memory Hypothesis Individuals will be more likely to exhibit aggression after watching violent media Three Claims Examples 1 Frequency Claim 8 million Americans consider suicide each year 2 Association Claim Eating disorder risk higher in educated families 3 Casual Claim Summer sun may trigger suicidal thoughts Frequency Claim describes a particular rate or level of a variable only includes one variable always in measured variable anecdotal claims are not frequency claims i e clams from your own life stories you ve heard not related to research Association Claim when claim argues that one level of a variable is likely linked with a particular level of another variable claims that variables correlate or covary includes 2 variables always measured variables Correlation Basics Correlation coefficient quantitative index of how well we are able to predict one set of scores using the other set Range 1 00 to 1 00 Sign indicates the direction of the relationship Positive varies together Negative varies opposite Zero no relationship Number indicates strength of the relationship Larger relationship closer to absolute value of 1 Can use associations to make predictions but predictions won t be perfect Causal Claims argues that one of the variables is responsible for the change in the other most powerful includes 2 variables must manipulate variables and randomly assign participants to conditions makes a stronger statement than the other two claims we will discuss experiments as a way to test this claim 4 validities we want to know if these claims are reasonable accurate and justified 1 Construct validity how well the study measured or manipulated a variable 2 External validity how well did the results of the study generalize to other people or situations outside of the study 3 Statistical validity whether the statistical solutions are accurate and 4 reasonable Internal validity the ability to rule out alternative explanations for a causal relationship between two variables a Only use this with causal claims Interrogating Frequency Claims using Construct External Validity example 2 million US teens depressed can apply construct validity how did they measure depressed in this claim Therapist opinion Self report Parent opinion can apply external validity who can they apply this to Does this represent all US teens Interrogating Association Claims example heavy cell phone use tied to poor sperm quality construct validity What is the sample Men Women Ages Race external validity statistical validity Problem find association when there is none false alarm Problem find no association when there is one miss AKA type I and type II error