COGS14A Final Review Material FA2024 PDF

COGS14A Final Review Material You can use this material to guide your review. Roughly 1/3 of the final will cover old material. The final should be about twice as long as each midterm....

COGS14A Final Review Material You can use this material to guide your review. Roughly 1/3 of the final will cover old material. The final should be about twice as long as each midterm. REMINDER: Final exam is Tuesday, Dec. 10, 2024 at 11:30am. 2024: It will take place in the Recreational Gym to allow for additional space. Current material What is a quasi-experiment and what are multiple reasons why might you want to do a quasi- experiment rather than a true experiment? Be able to identify and describe these types of quasi-experimental designs: pre-experimental design nonequivalent groups design time-series design multiple time-series design What are potential confounds and concerns when doing a quasi-experiment? Which confounds are different from doing actual experiments? ---- How do observational studies differ from people-watching, and when might you want to do an observational study or field experiment? When would you want to do naturalistic observation vs. participant observation (disguised vs. undisguised) vs. a field experiment vs. a lab experiment? What are the tradeoffs (pros/cons) of these different approaches? If you observe participants overtly, what can you do to lessen experimenter effects? What are the confounds to be concerned about in observational/field research? What kinds of observations can you do? Understand: narrative records; recoding; data reduction; checklists; sampling methods --- What is a nomothetic approach? An idiographic approach? How do single-subject experimental designs differ from case studies? Under what kinds of circumstances would you want to do a single-subject design?  What are characteristics of a good baseline, and why are baselines important? Be able to identify and describe these types of single-subject experimental designs. Keep in mind that students in previous classes have mixed up the first two, which are similar. Withdrawal design Reversal design Alternating-treatments design Multiple-baselines design Changing-criterion design What are some threats to internal validity in a single-subject design? --- What are the benefits and dangers of using physical trace studies and archival data sources? Understand: Traces vs. products; accretion vs. erosion; natural trace measures vs. controlled trace measures; selective survival; selective deposit; continuous records vs. discontinuous records; records vs. documents; coding systems What is “big data” and what kinds of datasets are out there? What kinds of questions is big data good at answering? What are some potential difficulties when working with a big dataset? What are some concerns/cautions to keep in mind when using big data resources? Understand some of the types of very large datasets that are being used in cognitive science, such as GWAS, large language corpora and large language models, etc. From previous sections Identify independent variables; dependent variables; subject variables. Understand the various ways that a correlation can occur: X causes Y; Y causes X; Z causes both X and Y (that is, a “third variable”); completely spurious correlation (no real relationship, it just by chance looks like a relationship) Mean, median, mode; standard deviation, range, entropy Understand what confounds are, and when a particular variable is a confound vs. just a nuisance variable. What makes for good internal validity? Good external validity?  Be able to recognize when a Type I error (alpha) vs. Type II error (beta) occurs in an experiment. Remember that Type I errors are false positives (it looks like there’s an effect, but there wasn’t actually) and Type II errors are false negatives (you didn’t detect an effect even though it was really there). Be able to describe each of these AND distinguish between them. Random selection (vs. convenience sampling) Random assignment Know what a factorial design is. Understand and be able to explain how factorial designs can be within-subjects, between-groups or both (mixed or “split-plot” design). Know how to look at a graph or data table and identify whether or not there are main effects and interactions. Be able to compute marginal means and draw the data pattern. Be able to describe the number of factors, number of factor levels, and number of cells in a factorial design. Be able to distinguish the following: true experiments; quasi-experiments; correlational designs; observational studies.