Research Methods Difference between observational correlational and experimental studies conclusions to draw from each what approaches provide better external and internal validity and vice versa Observational correlational experimental nObservational research nInvolved systematic qualitative and or quantitative descriptions of behavior nNo manipulation just describe n3 types naturalistic observation participant observation archival research nArchival is big data like facebook public does not know it s happening Observational scientific method involving systematic qualitative and or quantitative descriptions of behavior no manipulation just recording three types naturalistic participant archival naturalistic type of descriptive method that investigates behavior in its natural environment over prolonged period of time participant researcher records behavior as it occurs in in its natural environment but does so as a participant in the group being studied Archival examine already existing records of an individual group or culture Adv Disadv Ecological Validity no control observer bias still not natural environment invasion or privacy issues no control over third variable observer bias presence alters behaviors or those being studied ethical problems like invasion of privacy correlation research relationship between one thing and another two different variables measuring changes in one variable that are associated with another ex is there a relationship between TV violence and aggressive behavior measure two variables and see if the correlate with one another measured by correlation coefficient r ranged from 1 0 and 1 0 strength predictive power accuracy positive r both variables go up and down together negative r as one variable goes up the other goes down zero r no relationship between the variables there could be a curvilinear relationships U shaped or parabolic Adv Disadv strength when variables can not be manipulated when many variables are examined cheap and easy to do weakness causal direction of relationship is unclear third variable problem cannot determine the cause of the relationship between 2 variables third variable problem which causes which problem assess the nature of the relationship between two or more variables tat are not controlled by the researcher use surveys sets of questions or statements given to a group of people to measure their attitudes beliefs values or behavioral tendencies face to face phone written and computer use random selection to represent the population social desirability bias try to make your answers to make you sound good Experimental Research used to examine the cause and effect relationship between variables causation Types of experiments Field experiments similar to lab experiments but run in a natural setting good external validity not so good internal validity Laboratory experiments most used by far relies on random assignment flip coin and sampling interaction effects are important good internal validity not so good with external validity Meta analysis Disadv with Experiments Demand characteristics subjects expectation can influence subjects behavior Experimenter Effects Experimenters expectations can influence subjects behaviors can use double blind studies Internal Validity is the study measuring what it is supposed to can cause and effect conclusions be validly made be aware of confounding variables or poor measurement External Validity is the study a good analog to the real world do results generalize to other groups or situations is it going to relate back to the real world if you keep controlling balance internal and external validity in experiments Identify independent and dependent variables why experiments are experiments Independent variable IV the variable that is being manipulated cause of an event watching violent TV manipulation of an experiment Dependent variable DV dependent on the independent variable what you measure effect outcome of the event aggressive behavior measured in an experiment random sampling assignment act Advantages and disadvantages of the different types of ways we collect data in social psychology studies self report vs psychophysiology measures or observational vs implicit measures self report measures great for assessing how people feel bad for assessing how people feel Implicit measures IAT work fragments priming measures marked differences between implicit and explicit people can finish different words flashing cards think differently to prime Virtual environment the internet MTURK website in amazon social psychologist Physiological EEG fMRI PET EMG cardiovascular GSR startle EMG GSR cardio Direct physiological measures of indirect physiological processes Electroencephalography EEG measure of brain voltage from scalp event related potentials ERPs great temporal measure bad localization hard to see where it is happening what part of the brain Functional magnetic resonance imaging fMRI great for localization not good for temporal when placed in a magnetic field a hydrogen atom will precess or spin around the axis of that field B0 apply energy like a radio frequency at the same frequency of precession and the atom with go into a high energy state Functional MRI signal Local neural activity increases local metabolic rate which increased blood flow increased oxygenated hemoglobin uptake of oxygen less than supply surplus oxygenated hemoglobin decreased concentrations of deoxyhemoglobin increased local fMRI T2 signal DOES NOT MEASURE NEURAL ACTIVITY compare from when brain is at rest to when it is cognitively active We almost always collect data from sample BOLD stimulated state HbO2 crosses over in Hb venous section Difference between different analytical approaches like correlational analysis regression and analysis of variance what does it mean to have statistically significant findings Descriptive describing characteristics mode mean range variance vs inferential stats many ways to do inferential stats generalizing the sample to population Depends on what you want to look at and what assumptions you data fits Analysis of variance Pure experiments research regression MR correlational used for predictions general linear model combo ultimately want to generalize finding in your sample to populates as a whole is effect statistically significant statistical significance shows that you are not making a Type 1 Error False Alarm 95 p 05 what physiological phenomenon does fMRI asses BOLD Local neural
View Full Document