Psyc 2950, Experimental Methods, Exam 3 Study Guide
Psyc 2950, Experimental Methods, Exam 3 Study Guide PSYC 2950
Popular in Experimental Psychology
Popular in Psychlogy
This 14 page Study Guide was uploaded by Marcela Leon on Monday March 7, 2016. The Study Guide belongs to PSYC 2950 at University of North Texas taught by Alexander Yu in Spring 2016. Since its upload, it has received 66 views. For similar materials see Experimental Psychology in Psychlogy at University of North Texas.
Reviews for Psyc 2950, Experimental Methods, Exam 3 Study Guide
Clutch. So clutch. Thank you sooo much Marcela!!! Thanks so much for your help! Needed it bad lol
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: 03/07/16
Exam 3 Study Guide Experimental Research Design Chapter 8 Experimental Research: Looking for cause and effect by changing the IV Research Design: The outline to look for the cause and effect Weak experimental design vs. Strong design: Weakness or Strength is determined by Cause and effect Internal Validity Weak Experimental Research Designs: 1. One-group Posttest-Only 2. One-group Pretest-Posttest 3. Nonequivalent Posttest-Only One-group Posttest-Only Design: Only one group of participants is given the treatment The participants are only given a test after the treatment Ex: A group of native Spanish speaking students are given an English course to improve their English skills and take an exam by the end of the course. Problems with this design: The students did not take an exam before the course to evaluate their individual starting skill level. Therefore, we would not be able to compare the scores of the posttest and truly determine improvement. Use for this design: This design could be used, however, if background information on the participants’ English level (the DV) and past studies on the manner by which IV (type of English course) affects DV. One-group Pretest-Posttest Design: Only one group of participants is given the treatment There is a test before the treatment is given There is a test after the treatment is given Ex: A group of native Spanish speaking students take an exam assessing their English skill level before starting an English course. When the course is over, participants take an exam assessing their English skill level again to determine change. Problems with this design: Threats to Internal Validity: o History o Maturation o Regression artifact o Attrition Use for this design: This design can be used if control over these threats is possible and if the focus is mostly on the treatment effect. Nonequivalent Posttest-Only Design: Control group is used (not equal in variables or demographics) Experimental group is given the treatment Both the Experimental and the Control group are given the exam after the experimental group completes the treatment Ex: The native Spanish speaking students that take the English course had been in American English speaking classes since elementary, while the control group is a group of Spanish speaking students that recently transferred from Costa Rica. Problems with this design: The control and the experimental group are not equivalent, therefore extraneous variables are possible. There is also no test before the treatment to determine change. Use for this design: Equivalent comparison groups can be attained through random assignment. Control group: The group to which the effects of the IV on the experimental group will be compared to. Will not receive a treatment Natural/standard experience The 2 main functions: Estimate counterfactual: tells us what would have happened to the participants in the experimental group if they had not been given the treatment Control for other explanations for causation Experimental group: Will receive a treatment (IV) to determine its effects on the DV. Strong Experimental Research Designs: Experimental Designs that control for extraneous variables effectively and strongly assure cause and effect. 1. Posttest-Only Control Group (between-participants) 2. Pretest-Posttest Control Group Basic Designs: One IV and one DV Between-participants Within-participants Factorial Designs: Multiple IVs Between-Participants: different groups of participants receive different levels of the treatment Within-Participants: all participants receive all the levels of the treatment Posttest-Only Control Group Design (between-participants): In between-participant designs participants are exposed to different experimental conditions and then are randomly assigned to their groups. Could have more than one experimental group Participants are only given test after treatment Ex: The native Spanish speaking students are gathered from different schools, but are all randomly assigned to either the experimental (take the English course) or control group, and take a test after the experimental group has completed the treatment. Advantages to this design: Equivalent groups Problems with this design: No pretest Pretest-Posttest Control Group (between-participants): • Just like Posttest-Only but participants are given test before and after treatment 4 comparisons will be made: o Experimental group- before to Control group- before o Experimental group- before to Experimental group-after o Control group-before to Control group-after o Experimental group after to Control group-after Advantages of Pretest: Equivalent groups because of random assignment Ceiling and floor effects can be detected because of the comparison of the tests before the treatment Analysis of Covariance- tests the differences between the experimental and control group scores after the treatment Assurance that the IV causes changes in the DV because of the comparison of the pretests Within-Participant Design (repeated measures designs): Most common design: Posttest-only Participants receive every level of the IV No control group Posttest completed after each level Ex: Impact of different exercising techniques on stress. Each participant will try an exercising technique then take a posttest measuring stress levels. Then each participant will try another exercising technique and take a posttest and so on. Advantages: Each participant serves as his/her own control group Less people required Disadvantages: The order in which the different levels are presented to the participants may affect the DV Requires more time from participants o May cause fatigue o May wear out participants FACTORIAL DESIGNS There are 2 or more Independent Variables and can be within-subjects, between-subjects, or a combination of both (mixed design). 2 x 3 Factorial Design: Independent Variable A A1 A2 A3 B1 marginal Independent A1 B1 mean A2 B1 mean A3 B1 mean mean Varibable B B1 A1 B2 mean A2 B2 mean A3 B2 mean B2 marginal B2 mean A1 marginal mean A2 marginal mean A3 marginal mean Row means The average score of all participants receiving one Column means level of The average score of all participants receiving Independent one level of Independent Variable A. Variable B. In a 2x3x3x4x6 Factorial Design, there is 5 independent variables. IV A has 2 levels. IV B has 3 levels. IV C has 3 levels. IV D has 4 levels. IV E has 6 levels. There are 2 types of effects in Factorial Designs. Main Effect Interaction Effect Main Effect- The influence of one of the IV’s on the DV. Interaction Effect- The combined influence of all of the IV’s on the DV. Each IV’s effect depends on the other IV’s Identifying a main effect and an interaction effect when looking at a line graph: 1. 2. B1 B1 B2 B2 A1 A2 A3 A1 A2 A3 When looking for the main effect, we In this graph, the points for B1 are trying to find are higher than the points in B2, 1. the difference between the IV showing a significant main effect for the B1 average and B2 average IV B. And For the IV A, the low points and 2. the difference between A1 high points for one line are different average, A2 average, and A3 from the other. This shows a significant average effect for the IV A. For the difference between B1 and B2 average, look at the blue and red dots. As you can see, the red dots (B1) are neither higher or lower than the blue An interaction effect is determined dots (B2). Therefore, there is no main by the position of the two lines. If effect. the lines cross or touch, there is an For the difference between the IV A interaction effect. If the lines are levels, look at the B1 and B2’s lowest parallel, there is no interaction points and their higher points. When effect. Graph 1 shows an the low and high points of a line is the same compared to the other, it means interaction effect while Graph 2 shows no interaction effect. there is no main effect for the IV A levels. Advantages of Factorial Designs: More than 1 hypothesis can be tested at a time Extraneous variables are handled Increases precision Shows interactive effects Disadvantages: More difficulty manipulating all IVs when there is more than 2 More participants required Chapter 7 Control techniques Before a study: Random assignment Matching o Used when random assignment is not possible o Participants are matched on a given variable o Match by holding variables constant, building extraneous variable into the experimental design, yoked control, equating participants During a study: Counterbalancing o Sometimes used for repeated measures designs o Used to eliminate sequencing effects- order effects and/or carryover effects Control of Participants Effects o Double-blind placebo method o Deception o Control of participant interpretation Control of Experimenter Effects o Recording errors o Attribute errors o Expectancy error After Experiment: Statistical control o Analysis of covariance Matching Methods Holding Variables Constant: Participants with extraneous variables or those who differ on a variable are removed Ex: Conducting an experiment in which IQ is an extraneous variable. We would keep this variable constant by only using participants within the 110-120 range. Building Extraneous variables into the design (aka “Blocking”): Categories are used to match both the experimental and control group on the extraneous variable Ex: Same experiment as above, but instead of only using the 110-120 range, we make the IQ an IV with the levels: 90-99, 100-109, and 110-119. Yoked Control: Participant from control group is paired with a participant in the experimental group in which the outcome of the IV is also applied to the control participant Ex: If in your experiment a monkey is trained to press a button every 10 second interval or else it would receive an electric shock, every time it missed the interval not only would that monkey (experimental) would receive the electric shock, but also a monkey from the control (who would have no influence on the pressing of the button). Matching by Equating Participants: Does not build the extraneous variable into the design Matches individual participants on the controlled variable 1. Participants are listed by order of their score on an extraneous variable 2. Participants are paired (matched) by order 3. One is placed into each treatment group 4. Steps 2 and 3 are repeated Counterbalancing: technique used to avoid sequencing effects such as order and carryover. Applies only to repeated measures designs (aka within-participants design) Treatments are administered in different sequences Order effects: the order in which the participants are receiving the different levels of the IV have an influence on the DV Ex: You conduct an experiment testing how the number of hours of sleep affect mood with 3 levels: 3 hours, 6 hours, and 9 hours. All of your participants will participate in all 3 levels but in different sequences. Those who receive 6 hours of sleep the first day, then 3 hours the second day will most likely report more negative moods during the 2 posttest than those who received 9 hours the first day, and then 6 hours the second day. Carryover effects: the participant’s performance in one treatment condition influences his/her performance in another treatment condition Ex: You conduct an experiment in which you are testing if students write better essays when listening to music versus in a quiet room. Your participants first write an essay listening to music then write an essay in a quiet room. You find that the second essay on average was better than the first. It is a possibility, however that it was not the quiet room that influenced the participants’ writing, but the fact that they had just written an essay before and were still in that mindset. That mindset would be a carryover effect. The main features of different counterbalancing methods: Randomized: For each participant the order of the IV is randomized Ex: P1= 1 2 3, P2= 3 2 1, 3 = 2 1 3 Intrasubject: Participants are given the conditions in one order and then in the reverse order Ex: 1 2 and then 2 1 or 1 2 3 and then 3 2 1 Complete: Each possible treatment condition sequence has an equal number of randomly assigned participants Ex: 123 321 132 213 231 312 ---- treatment condition sequences 4 4 4 4 4 4 ----- randomly assigned participants Incomplete: Based on equality- does not use all possible sequences of conditions Steps: 1. Each order position must have an equal number of times each condition appears 2. Each condition must come and go an equal number of times before and after each of the other conditions 1, 2, n, 3, (n-1), 4, (n-2), 5 ------------------------- n= number of conditions Participant Sequence of conditions 1 A B D C 2 B C A D 3 C D B A 4 D A C B For odd number of conditions, in this instance 5 (1 more than the example above), sequence must then be followed by a reverse sequence. 1 A B E C D Sequence 2: D C E B A 2 B C A D E E D A C B 3 C D B E A A E B D C 4 D E C A B etc… 5 E A D B C 2 types of participants effects: 1. Double-blind placebo method- neither the participant nor the experimenter are aware of which group they are in 2. Deception- either purposefully lying to participants or not providing information 5 ways to control participant interpretation/response: Retrospective verbal: the participant is asked to recall the experiment after he/she has completed it Concurrent verbal: as the experiment is being performed, the participant verbally reports on it Sacrifice groups: participants are stopped at different stages of an experiment to be interviewed (their data is not used) Concurrent probing: participants are interviewed after each trial Think aloud: participants verbalize their thoughts while going through the experiment The 2 experimenter effects: Recording errors Attribute errors- the experimenter’s personality characteristics influence participants 3 types of expectancy error: Blind technique- experimenter does not know what conditions participants are in Partial blind technique- experimenter does not know what conditions participants are in for most of the study Automation- no interaction between experimenter and participant; experiment is done online, through recordings, etc. Chapter 14&15 The 2 main branches of statistics: Descriptive- mean, range, variance, etc. -frequencies, averages -measures of central tendency -variability Inferential- hypothesis testing, significance, etc. Variability: how spread out the scores are We look at variability through: Range Variance and Standard Deviation Standard deviation is the square root of the variance. Normal distribution: 68% , 95% , 99.7% 68% of the cases fall within 1 SD of the mean, 95% fall within 2 SD of the mean, and 99.7% fall within 3 SD of the mean. Z-score: score transformed into SD units. Z-scores are used to interpret how far the scores are from the mean. Statistical Analysis: Independent samples t-test: Used to compare the difference between the means of 2 groups Ex: Is there a difference between the stress level scores for group 1 who performed exercise #1 and group 2 who performed exercise #2? (Used for Posttest-only Control Group and Pretest-posttest Control group) One-way ANOVA: Used to examine the relationship of: one quantitative DV and one categorical IV with multiple levels. Ex: The type of exercise (exercise #1, exercise #2, and exercise #3) will lower stress levels. (Used for Posttest-only Control group- with 2 experimental groups) 2-way ANOVA: Used to examine the relationship of: one DV and two IVs with multiple levels Ex: The type of exercise (exercise #1, #2, or #3) and working hours (0-10, 10-20, and 20-30) of the participant have an influence on his/her stress levels. Simple Regression: One IV is used to predict one DV outcome Ex: An individual’s stress scores will predict their mood scores. Multiple Regression: Two or more IVs are used to predict one DV outcome Ex: An individual’s stress scores and blood pressure scores will predict his/her physical health. The 2 main differences between ANOVA and Regression: ANOVA Regression Looks for significant relationships Predicts relationships Has levels Does not typically have levels
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'