PSY3213C Research Methods in Psychology Exam 3 Study Guide Assume that there are about 4 6 points for each question on here equivalent of 2 3 multiple choice questions If a question is labeled with BIG is about 8 10 points I have labeled the questions that are asked about in the short answer section A label of short answer does not mean that there are not also multiple choice questions there could be both multiple choice and short answer for a question BETWEEN PARTICIPANTS DESIGNS AND STATISTICS 1 What is error variance and why is it a problem e g total variance systematic variance error variance What are the four ways to deal with it control it randomize it across conditions increase effectiveness of your IV and stats Error variance all of the things that relate to our dependent variable that are not our independent variable Total variance systematic error systematic is what we can explain from our dependent error variance often comes from individual differences between people age IQ working memory capacity etc To deal with error variance you want this to be small having more error variance is going to make it harder to see our systematic variance Reduce it ex treat all participants the same sample from a group that s already pretty similar to each other match if age is a factor find people with similar ages match them and put them in separate groups Increase effectiveness of IV Making your groups that are different from each other more different Randomize it across conditions main strategy of randomized group design Use statistic t test or ANOVA helps us look at systematic variance in comparison to our error variance 2 BIG short answer What are the characteristics of advantages and disadvantages of Randomized 2 Group designs Randomized Multi Group designs Matched Pair designs Matched Multi Group designs and Within Participant designs Be able to identify and design a study of each kind More info with multi group but more participants needed 2 groups is the Randomized 2 group vs multi group simplest design it s easy Matched vs randomized Matched is nice because you removed some of you error variance but you have to measure whatever you want to match on before the study you d have to know who all of your participants are measure them and then you randomize them you can t use as powerful statistics not as strong outcome Within vs between Factorial vs Non factorial Within 1 participant is giving you data for all of your conditions within is more powerful and fewer participants are required but it s more burdensome on participants Carry over effects something you do find in one condition can change the outcome of the other condition More than one independent variable good because they give us more info tell us how 2 or more variables are related to our IV but also if they are related to each other you can t look at the relationship if you only have 1 IV less participants Bad sometimes have interaction effects can be difficult to interpret 3 BIG short answer What are the different types of analyses you can do with data from an experiment with one IV t test ANOVA a Know null and alternative hypotheses for each b Know how to interpret results of each reject or fail to reject null statistical significance how you describe results in regular words c Why do you need to do posthoc tests in an ANOVA t test looks at the difference between your groups compared to the variance within your groups the stuff your IV has caused vs other stuff if we have a larger difference between groups and little variance within groups that will give us a higher number which is good basically saying we can see a difference that s there being able to see the signal through the noise Null hypothesis IV will not affect DV no difference between groups in the population H0 ui uc alternative hypothesis IV will affect DV there is a difference between groups in the population if p 05 you reject the null hypothesis there s a really small chance that you found a result this big by chance with no influence of the IV if p 05 you fail to reject null you re not confident that it s not 0 ANOVA need to use this when we have more than two groups f test based on the same ideas as a t test only difference is that between group variability is that it s across more groups omnibus there s some difference but we don t know where Null IV will not affect DV no difference between groups in the population Alternative IV will affect DV there s some difference in the groups in the population key work some we don t know what groups have difference or how much there is first group is different from the second AND OR first is different from third AND OR posthoc test tells us where the difference can be found if we fail to reject null then we don t even do posthoc because we know there s no difference Pairwise comparisons Turkey deal with the type 1 error posthoc shows individual difference and that s what interpretations conclusions are based off of only necessary if you reject null 4 What are t and F ratios of between variance within variance How do mean differences and variability impact t and F and therefore whether you reject the null hypothesis WITHIN PARTICIPANTS DESIGNS 5 What are carryover effects what types of carryover effects are there and how can these be addressed In what situations is it difficult to deal with carryover effects Notes look at learning fatigue contrast effects etc Each participant gets every level of the variable every person in every condition Two situations when carryover effects can t be dealt with easily When a treatment produces an irreversible effect if once you ve done it Differential carryover effect if the order of the conditions is different than you can t go back it s hard if you had flipped it 6 What different methods are there of counterbalancing and what are the advantages disadvantages of each You have different people get the different conditions in different orders because not everyone had it done the same way run all potential orders of treatments it s easy when you have 2 groups but if you have like 5 all of the possible orders is like 120 not a great strategy if you have a lot of conditions but great if you only have like 2 Partial counterbalancing fewer orders tested fewer participants but we don t get to see how each treatment potentially impacts other treatments Reverse counterbalancing if there s 5 conditions you d get 1 2 3 4 5 and then 5 4 3 2 1 Random counterbalancing there s no specific
View Full Document
Unlocking...