EXAM 2 REVIEW

Question 1

bivariate correlation

Answer 1

Association that involves two variables
A correlational study can measure multiple variables but it will only measure 2 at a time

Question 2

what do you use when both variables are categorical

Answer 2

scatterplot

Question 3

what do you use when one variable is categorical and one is numerical

Answer 3

bar graph

Question 4

what 2 validities are most important in association claims

Answer 4

construct and statistical

Question 5

threats to statistical validity

Answer 5

effect size, outliers, restriction of range, curvilinear

Question 6

third variable problem: internal validity

Answer 6

an alternate explanation for the association between two variables

Question 7

moderator

Answer 7

A third variable that affects the strength or direction of the relationship between two other variables, indicating when or under what conditions a particular effect can be expected

Question 8

what should the independent variable be compared to during covariance

Answer 8

comparison group (three kinds)
control
treatment
placebo

Question 9

internal validity threats

Answer 9

design confound and selection effect

Question 10

design confound internal validity

Answer 10

a second variable that varies systematically along with the IV this is the mistake of the experimenter

Question 11

selection effect internal validity

Answer 11

participants in one level of IV systematically differ from other levels. Avoid this by using random assignment or matched groups (pairs of participants are matched in terms of key variables, such as age or socioeconomic status)

Question 12

systematic vs unsystematic selection threats

Answer 12

there’s a difference between 2 groups (one kind of person in one condition and another kind of person in another condition), e.g., only males in the treatment group and only females in the control group. Another example: treatment group people all live in the city, control group all live in the suburbs. Clearly, these people all have variation, but they also share a specific commonality. Threat to internal validity.

Unsystematic: variability within each group is not a threat to internal validity. (Ex: treatment group you have people from the city and country, control group you have people from the city and country) It can decrease statistical power and make it more likely for a null result

Question 13

independent group design (between subjects)

Answer 13

different groups of participants placed at different levels of the IV ex: One group of people sees one set of test signs and another sees a different set of test signs but they are all tested on the same dependent measure
pretest/posttest or posttest only

Question 14

posttest only

Answer 14

randomly assigned to IV groups and tested on the DV once. Not the best design; pretest/posttest gives more information.

Question 15

pretest/posttest

Answer 15

randomly assigned to IV groups, tested on the key dependent variable before and after exposure to IV. Benefit: the groups are equal on variable of interest

Question 16

why is pretest posttest problematic at times

Answer 16

Exposure to the experiment may cause fatigue and therefore worse results or people who are preexposed get better at the experiment and their results improve due to familiarity effects

Question 17

within groups

Answer 17

each participant is presented with all conditions of the IV

Ex: one group sees both sets of signs and then gets tested on the same dependent measure

repeated measures or concurrent measures

Question 18

repeated measured

Answer 18

measure on DV after exposure to each level of the IV

Question 19

concurrent measures

Answer 19

exposed to all levels of the IV at roughly the same time and single attitude/preference is DV

Question 20

threat to internal validity within groups

Answer 20

Order effect: exposure to one level of IV influences reactions to other levels of IV
Practice effects: get better from practice or worse due to fatigue
Carryover effects: contamination carrying over from one condition to another

Ways to counter these threats
Counterbalancing:
Full: all possible conditions orders are used (use all combinations so order effects dont have an impact)
Partial: possible condition orders are used

Question 21

within group design disadvantages

Answer 21

1. Order effects
2. Might not be practical or possible
   3.Demand characteristics: participants act in different ways based on knowledge about the IV

Question 22

within group benefits

Answer 22

No selection effects
Unsystematic variability is less of a problem
Statistical power: better ability to detect between conditions
Need fewer participants

Question 23

how can you check construct validity

Answer 23

Manipulation check: an extra dependent variable that researchers can insert into an experiment to convince them that their experimental manipulations worked

Pilot study: a simple study with a separate group of participants, usually completed before the main study to confirm the effectiveness of a manipulation

Question 24

what is the really bad experiment

Answer 24

One group pretest posttest bad bc no comparison group

Question 25

what are the 6 potential internal validity threats in the really bad experiment

Answer 25

maturation threat, history threat, attrition threat, regression threat, testing threat, instrumentation threat

Question 26

what is maturation threat

Answer 26

change in behavior that emerges spontaneously over time. Long lasting experiment or study. Prevented by adding comparison group.

Question 27

history threat

Answer 27

external factor affects most members of group (systematic); e.g., COVID hits and effects everyone. Prevented: add comparison group

Question 28

regression threat

Answer 28

if the group mean is unusually extreme at time 1, time 2 will likely be less extreme. Only a threat when a group is measured twice and has an extreme score at pretest. Prevention: comparison groups + careful inspection of the pattern of results. For example, a star football player got no touchdowns in a game, he is due for touchdown regression (unlikely this would happen again) he regressed back closer to his mean or average

Question 29

attrition threat

Answer 29

Participants drop out systematically. Prevention: remove score from pretest and posttest or check if pretest is close to average

Question 30

testing threat

Answer 30

order effect, exposed to DV twice so better because of practice or worse because tired/bored. Prevention: posttest only, comparison group, or different forms of DV (ex: different depression questionnaires); participants change

Question 31

instrumentation threat

Answer 31

Measuring instrument changes over time Changed standards of scoring Different pretest/posttest forms of DV Instrument has changed Prevention: if two different tests are used, make sure they’re calibrated; use clear coding manuals to retain coders throughout experiment; counterbalance versions of the test, posttest-only design

Question 32

what are the combined threats

Answer 32

selection history selection attrition

Question 33

selection history threat

Answer 33

an outside event or factor systematically affects participants at one level of the IV Ex: lets say all schools are experiencing budget cuts (normal history threat) a selection history threat would be more budget cuts in the control schools relative to the treatment schools (or vice versa)

Question 34

selection attrition threat

Answer 34

participants in only one experimental group experience attrition Ex: lets say students are dropping out of school in both the control and treatment schools (regular attrition threat); a selection attrition threat would be more students dropping out of school in the control schools relative to the treatment schools (or vice versa)

Question 35

potential internal validity threats to experiments with comparison groups

Answer 35

Observer bias: researchers expectations influence interpretation of results Demand characteristics: participants guess what the study is supposed to be about and then change their behavior to match the desired outcome Controlling for observer bias and demand characteristics Double blind study: neither participants nor researchers know who is in treatment group and who is in comparison group Masked study: researchers are unaware Placebo effects: recipients believe receiving the effective treatment so they improve. It reduces real symptoms, both psychological and physical. Ex: antidepressants vs placebo. Placebo showed 75% of antidepressant effects on depression improvement Control for placebo effects by using a double-blind placebo control study

Question 36

power

Answer 36

the likelihood that a study will yield a statistically significant result when the IV really has an effect Leads to more precise estimates Can be improved with Within groups design A strong IV manipulation A large sample size Less within group variability

Question 37

advantages of large samples

Answer 37

Increase statistical power More precise estimate by narrowing CI Small samples are less precise, increasing the likelihood that you will detect an effect that is not actually there making it unlikely to replicate

Question 38

null effects

Answer 38

the independent variable has not exercised any influence on the dependent variable there seems to be no significant covariance between the two. The independent variable has not impacted the dependent variable OR the research was not accurately set up or performed and there was really an effect. Use replication and do more research to ensure there really is no effect

Question 39

what if the IV does not make a difference there might be

Answer 39

not enough between group variability, too much within group variability, or really no difference

Question 40

not enough between group difference

Answer 40

Weak manipulations: the difference between levels of the IV is too small to matter or be meaningful Insensitive measures: null result emerges because the operationalization of the DV does not have enough sensitivity to detect a difference between levels of the IV Ceiling effect: Participants scores are squeezed together at the top end of the DV scale Floor effect: participants scores are squeezed together at the bottom end of the DV scale Ceiling effects and floor effects can be the result of a problematic independent (if money is manipulated in the range of 0-10 cents the amount is so low it might cause a floor effect on the DV, if anxiety is manipulated by telling participants they will get a 10, 50, or 100 volt shock it may cause a floor effect on the DV because everyone is anxious) or dependent variable (test questions too easy everyone gets them right, test questions too hard no one answers right)

Question 41

What's an additional dependent measure to add to a study that can reveal a weak manipulation resulting in ceiling and floor effects?

Answer 41

manipulation check

Question 42

can confounds impact null results

Answer 42

yes

Question 43

too much within group variability

Answer 43

Measurement error: a human or instrument factor that can randomly inflate or deflate a persons score on the dependent variable. All DVs involve a certain amount of measurement error researchers try to keep those errors as small as possible. When the distortions are random they cancel each other out and dont affect group mean. A lot will have measurement errors that are spread out making it harder to detect a difference between groups Solutions: use precise measurements (reliable tools) and measure more instances (large sample) Individual differences: Difference across participants add variability in DV scores Ex: in an anxiety study, people may be more prone to stress or anxiety Solution: change the design (use within groups) or add more participants (the more people the less individual differences will affect the average) Situation noise: any kind of external distraction that could cause variability within groups that obscures between group differences. It can be minimized by controlling the surroundings of an experiment

Question 44

on a bar graph how do we see if there is significance

Answer 44

if the CI error bars dont overlap

Question 45

if the CI includes 0, is the relationship significant

Answer 45

no