Chapter 10 Notes
Chapter 10 Notes Psych 2300
Popular in Research Methods in Psychology
Popular in Psychlogy
This 7 page Class Notes was uploaded by Emma Dahlin on Saturday November 14, 2015. The Class Notes belongs to Psych 2300 at Ohio State University taught by Seth Miller in Fall 2015. Since its upload, it has received 23 views. For similar materials see Research Methods in Psychology in Psychlogy at Ohio State University.
Reviews for Chapter 10 Notes
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 11/14/15
Chapter 10 Notes: Intro to Simple Experiments Experimental Variables Main reason that scientists conduct experiments=to make causal claims Experiment-means researchers manipulated at least one variable and measured another Independent Variablemanipulated variable-variable that is controlled, such as when researchers assign participants to a particular level (value) of the variable o Comes first, has temporal precedence o Its levels are referred to as conditions Dependent Variablemeasured variable- takes the form of records of behavior/attitudes, such as self-reports, behavioral observations, or physiological measures o Comes second o Outcome variable Random assignment to different conditions (levels of IV) is designed to help “control” variables by equalizing levels of most third variables across conditions o Easier but not always as effective as directly controlling or measuring these variables Control Variablesany variable that a experimenter holds constant on purpose o Preventing variation rules these variables out as possible third variables o Essential for experiments o Helps to improve internal validity o Random assignment to conditions of IV with help with most variables o Allow researchers to separate one potential cause from another and eliminate alternative explanations Why Experiments Support Causal Claims Three rules for causation: 1. Covariance-is causal variable related to effect variable? Are distinct levels of the independent variable associated different levels of the dependent variable? 2. Temporal precedence-does causal variable come before effect variable in time? 3. Internal validity-are there alternative explanations for the results? CovarianceDoes manipulating the independent variable make a difference? Well…compared to what? Comparison groupA level of the IV that we compare the treatment group(s) to. (Often this is not a control group). Control groupA level of the IV that represents “no treatment” or a neutral condition. o We compare the treatment group to the control group. o Placebo control group: when control group exposed to something like the “treatment” condition but without the critical elements thought to exert an impact on the DV. All experiments need a comparison group, so the researchers can compare on condition to another; but the comparison group does not NEED to be a control group Temporal precedenceThe independent variable is manipulated, so researchers can usually ensure that it has temporal precedence. o Gives experiments advantage over correlational studies Internal validitycan we rule out other variables as possible causal agents? o One of most IMPORTANT validities for causal claims o Must ensure that casual variable is responsible for change in effect variable o Confounds are threats to internal validity. They are variables that confuse or “confound” our understanding of what variable actually changing the DV. o Design confounds can often be thought of as a researcher’s mistake; a variable that happens to vary systematically along with the IV Correction- control variable or redesign study to break the covariation b/t manipulated variable and confounding variable o Systematic Variability-not every potentially problematic variable is a confound…only a problem for internal validity if they show systematic variability with IV o Selection Effect-occurs when, systematically, different types of participants are in various groups Solution- random assignment or matched groups (matching) Well designed studies often use random assignment to avoid selection effects o Ex: Flipping coin, random number generator, rolling die o However, need to have larger sample size b/c it will most likely fail when there is small sample size Matched Groups will often required pretesting o Requires more effort/time o Participants matched on particular variable of interest (e.g IQ) o Especially useful is sample size is small Types of groups used to show that there is covariance b/t the DV and IV=control, placebo control, and comparison groups Between-Groups Design: different groups of participants assigned to different levels of the IV ; also called independent- groups design o Posttest-only designs participants randomly assigned to IV groups and are tested on DV once o Pretest/posttest designsparticipants randomly assigned to at least two groups and are tested on key DV twice- once before and once after the exposure to IV Allow researchers to evaluate whether random assignment made groups equal for variables of interest Allow researchers to examine how people change over time in response to manipulated IV Within-Groups Design: only one group of participants and each person is presented with all levels of IV o Concurrent-measures designsparticipants are exposed to all levels of an IV at roughly the same time; single attitudinal/behavioral preference is the DV o Repeated-measures designsparticipants are assigned to all levels of the IV and participants are measured on DV multiple times o Advantages of Within-Groups Designs- all conditions should be equivalent for non-manipulated variables Same people are in all conditions Participants serve as “their own controls” Generally requires fewer participants overall May increase power (relative to an analogous between subjects design) Power= the ability to detect a statistically significant result (e.g., an association or a difference between means) when that result actually exists in the population Allow for reduction of “noise” or unsystematic variability in data o Potential Disadvantages of Within-Groups Designs 1. Order effectsoccurs when exposure to one level of IV changes how participants react to another condition or level of IV…participant performance may be affected by sequences in which conditions are experienced (confound) Practice effects: (fatigue effects) long sequence might lead participants to perform better as time goes on, or get tired or bored toward end Carryover effects: some form of “contamination: spills over from one condition to next 2. May not be possible or practical 3. Demand characteristics (experimental demand) Being exposed to multiple levels of IV may allow participants to guess hypotheses of study/how they are expected to behave (this diminishes internal validity) o Counterbalancingway to avoid order effects (therefore increasing internal validity) Present levels of the IV to participants in different orders This should cancel out order effects Participants should be randomly assigned to diff. orderings Full counterbalancing: all possible condition orders are presented Useful/feasible when # of levels of IV is limited However, as # of levels of IV increases, many different orderings are possible! Partial counterbalancing: only some of the possible condition orders are presented Easy to accomplish via computer…conditions can be provided in random order for each participant Use of Latin Square may be helpful (counterbalancing system that ensures that each condition appears in each position once) Is Pretest/Posttest a Within Group Design? o No, b/c in a true within-group design participants are exposed to all levels fo IV o In pretest/posttest, participants see only one level of IV, not all levels Interrogating Causal Claims with the Four Validities Construct validity: How well were DV measured and how well were IV manipulated? o Manipulation checks may be used to ensure that experimental manipulation worked correctly (extra DV is inserted into experimental to help quantify how well experimental manipulation worked) o Pilot study is a simple study with a separate group of participants, that is completed before (or sometimes after) conducting study of primary interest o Researchers can show results support their theory by collecting additional data External validity: To whom or to what can the causal claim be generalized? o Researchers often sacrifice some external validity in order to improve internal validity o Make sure there is random sampling from population of interest Statistical validity: How well do the data support your causal conclusion? Statistical significance? o Is difference statistically significant? (look at diff. b/t means) Unlikely to have occurred by chance o If so, how large if effect size? Cohen’s d used to measure (larger d=less overlap b/t scores) In general, larger the effect size, the more important/stronger the causal effect But sometimes, small effect sizes can be important! Internal validity: Are there alternative explanations for the outcome? Internal validity is often the top priority for researchers conducting an experiment. 1. Are there design confounds? (did researchers manipulate more than just the IV) 2. Were selection effects controlled for via random assignment or matching? (in a between-groups design) 3. Were potential order effects (practice, fatigue, or carryover effects) controlled for? (in a within-groups design) T/F STATEMENTS “You always need a control group in an experiment.” o FALSE: Comparison group can also work quite well. In some cases, actual control group may not be possible. “You always need a pretest in an experiment.” o FALSE: they are useful for ensuring equal groups before manipulating IV…however, not needed. Random assignment and matching can also create equivalent groups. For within- groups designs, participants serve as own controls. “You need a random sample to get anything useful from an experiment.” o FALSE: random samples are good b/c they improve external validity…but researchers often sacrifice external validity to maximize internal validity.
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'