Chapter 6.docx

13 Pages
53 Views
Unlock Document

Department
Psychology
Course
PSYB01H3
Professor
David Nussbaum
Semester
Fall

Description
Chapter 6 True experiments I True experiment 1. A controlled investigation in which one or more variables are manipulated 2. Variables differ in the degree to which they can be controlled or manipulated 3. Researcher has complete control over manipulation of the independent variable 4. Standard approach to establish causality Basics of experimentation 1. Control over all aspects of the study design 2. Controls what levels of the independent variables are to be manipulated Experimental control 1. Built into study’s design 2. Experimental control by holding as many confounding variables constant as possible a. If a variable is held constant, it cannot account for the results of the experiment b. Variables are held constant through standardization of test administration and test scoring Randomization 1. Used in assigning participants to groups or different levels of an independent variable 2. Process of random assignment ensures that any extraneous influence is just as likely to affect one group as the other group 3. Many studies cannot use random assignment to create experimental and control groups a. The levels of the independent variable are selected rather than directly manipulated by the experimenter (ex post facto) b. This type of experiment is called a quasiexperiment Independent variables 1. Variable that is manipulated in order to determine its effects on a dependent variable 2. Must have at least two levels 3. Often referred to as factors in an experiment 4. Must manipulate at least one independent variable in order for it to be an experiment a. One  single factor experiment b. More than one  multifactorial design 5. Condition a. Particular way in which participants are treated i. Treatment group 1. Receives treatment ii. Comparison group 1. Receives alternative treatment iii. Control group 1. Receives no treatment 2. Variable that is held constant 3. Provides a baseline against which the treatment group can be compared Dependent variable 1. Dependent variable as a post-test a. Measurement of the outcome in both groups after the treatment group has received the treatment 2. Pre-tests may also occur to measure the dependent variable prior to treatment a. Permit a direct measure of how much the experimental and comparison groups changed over time b. However be wary of using identical pre and post tests since it may result in practice effects i. Use of alternate-form reliability methods to avoid error Control variable 1. Control variable is a potential independent variable that is held constant throughout the experiment because its influence may be extra or unimportant to the study 2. Goal is to show that these extra variables have no effect on how the measured independent variable influences the dependent variable 3. It is unrealistic to control for all confounding variables, but must try to control for the most influential ones on the study Research design alternatives 1. Two types of experimental designs a. Between subject design i. Independent groups of participants receive different levels of the independent variable ii. Requires a large number of people to be enrolled in a study iii. Often used in treatment outcome studies where you need to ensure that there is no chance of one treatment contaminating the other b. Within subject design i. Each participant receives all levels of the independent variable ii. Repeated measures is when the same participant can be tested repeatedly over time iii. Each participant is compared against himself across all experimental levels 1. Therefore acts as a control for himself 2. Performance more likely attributable to the different manipulations of the experimental design iv. Fewer research participants are needed because all participants will receive all treatments v. Only use within subject designs when you are sure that the effect of one treatment will not carry over to the next treatment vi. Essential for longitudinal studies 1. The same research participants are followed over a long period of time (years-decades) vii. Concern 1. Carryover effects a. If performance in one condition affects performance in a subsequent condition b. Results will be confounded by the order of presentation of conditions c. Will not be able to attribute the results of a study to a particular treatment or a particular level of an independent variable d. May also occur as a result of routine factors (practice effects) e. Use randomization to combat carryover effects i. Ensures that treatment orders are due to chance ii. However there are often fewer treatment orders than there are participants so it may not be an effective method for equating the order of treatments f. Counterbalancing is a method to combat carryover effects (scroll down) Randomization and matching 2. Between subjects design concern a. Ensure that there are as few differences as possible among participants in the various treatment groups 3. Random assignment allows for a. Treatment groups to be formed b. Levels of the independent variable to be manipulated and assigned c. A true experiment to exist d. Individual differences to be equalized or neutralized 4. How can you be sure that random assignment worked in creating equivalent groups prior to the introduction of the treatment variable a. A pretest measure before treatment is often used to evaluate whether random assignment was successful i. i.e. sampling or selection biases did not lead to an initial difference between the groups 5. A randomized experimental design with a pretest and posttest is termed a randomized comparative change design a. Or pretest-posttest control group design 6. Pretest is not strictly required a. Random assignment is thought to equate experimental and comparison groups before the intervention is administered to the treatment group i. So technically it is not required 7. Disadvantage of pretest a. Cannot do a pretest on some dependent variables i. E.g. testing for reaction to a teacher  cannot pretest unless teacher is actually present b. Pretest may cause participants to change i. Practice effects 8. Matching a. Another procedure used to equate groups b. Poor substitute for randomization (by itself) c. Matching of individuals in a treatment group with those in a control group may involve i. Pairing persons on the basis of similarity of some objective characteristic 1. May be a problem because individuals only have some common characteristics while the rest may be different d. When matching is used as a substitute for random assignment, the research becomes quasiexperimental e. Matching combined with random assignment results in matched-pairs design i. Ensures that groups are equivalent before the test ii. Individuals matched for testable characteristic and then randomly assigned to either treatment condition or control condition (different levels of the independent variable) f. Greater power to detect statistical significance than simple random assignment Counterbalancing the carryover effect 1. Controlling for the order of treatments 2. More effective than randomizing the order of treatment because it ensures that each treatment or level of the independent variable occurs in each time period of the experiment 3. Every treatment has the same chance of being influenced by confounding variables related to a treatment coming before and after it 4. Complete counterbalancing begins by calculating all possible treatment orders a. The possible amount of treatment orders increase as a factorial function with the amount of treatments i. 1 treatment = 1! ii. 2 treatments = 2! (2x1) iii. 3 treatments = 3! (3x2x1) iv. 4 treatments, it becomes impractical to use because it has 24 different treatment orders that you have to control for 1. Latin square design to account for higher treatment order amounts Latin square design (partial counterbalancing) 1. Way to combat confounds of carryover effects in within subjects experiments 2. Each possible order of treatments occurs equally often across research participants a. E.g. treatments A, B, and C i. Participant 1 gets A  B  C ii. Participant 2 gets C  A  B iii. Participant 3 gets B  C  A 3. Disadvantage a. Because the same latin square arrangement can be used more than once, there is a possibility that somehow th
More Less

Related notes for PSYB01H3

Log In


OR

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


OR

By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.


Submit