Chap. 9: Multivariate Correlation
What are the three criteria for causal claims?
Covariance, temporal precedence, and internal validity
How would you tell if a study is independent-groups or repeated-measures?
If independent-groups, different groups are tested in different experimental conditions. If repeated-measures, the same group is tested across time.
A researcher wants to study the influence of Vitamin D and geographical location on happiness. They give some participants Vitamin D supplements, some a placebo, and some they instruct to eat foods rich in Vitamin D. Some participants live in Seattle and some in Washington, DC. What kind of design is this? (Put your answer in the form “___ × ___.”)
3 x 2
Why isn’t a quasi-experiment a true experiment?
No manipulation of variables and therefore provides weaker evidence for cause and effect
Why do scientists never say they can prove a result, even if p is statistically significant?
The p value is a probability; there is always a chance that the null hypothesis may be true.
Why might a researcher choose a multiple regression design?
To get as close as possible to doing an experiment when an experiment cannot be conducted; to rule out third variables, which bivariate correlation cannot do
Identify the independent and dependent variables in the following study: A group of researchers investigated the effect of mindfulness meditation on mental health workers, 10 weeks after a major hurricane. A sample of 15 mental health workers were pretested on their depression and anxiety symptoms. Then they engaged in meditation training for 8 weeks. After the training was completed, they were tested on their symptoms again, using the same test. The study found that anxiety and depression symptoms were lower at posttest (the 95% CI for the drop in scores excluded zero). The researchers concluded the meditation training helped participants.
Independent = meditation training; dependent = depression and anxiety
A researcher wants to study the influence of Vitamin D and geographical location on happiness. They give some participants Vitamin D supplements, some a placebo, and some they instruct to eat foods rich in Vitamin D. Some participants live in Seattle and some in Washington, DC. What are the independent and dependent variables, and is this study within or between groups?
Independent = method of Vitamin D intake and geographical location; dependent = happiness; between groups
Explain two differences between a quasi-experiment and a true experiment
True experiment: control of IV, random assignment, higher internal validity, intentionally planned, requires ethical consideration; quasi-experiment: no control of IV (quasi-IV), nonrandom assignment, lower internal validity, natural opportunity, ethically less risky
What does a confidence interval represent?
The percentage of confidence that the true population value lies within two values. For example, if reporting a 95% confidence interval, there is a 95% chance the true value falls within the two given.
What type of correlation is shown here, and what does it measure? https://drive.google.com/file/d/1qolcLQLJ_REYQOUB2t0JMtkNZRGMx77q/view?usp=sharing
Autocorrelation; it compares a variable to itself across time
Identify the research design (posttest-only, pretest-posttest, repeated measures, one-group pretest/posttest): A group of researchers investigated the effect of mindfulness meditation on mental health workers, 10 weeks after a major hurricane. A sample of 15 mental health workers were pretested on their depression and anxiety symptoms. Then they engaged in meditation training for 8 weeks. After the training was completed, they were tested on their symptoms again, using the same test. The study found that anxiety and depression symptoms were lower at posttest (the 95% CI for the drop in scores excluded zero). The researchers concluded the meditation training helped participants.
One-group pretest-posttest
What main effects and/or interactions does this line graph show? https://drive.google.com/file/d/14UA6lRdxjhlJLhzf04AiQTnKAdeNOXFA/view?usp=sharing
Main effects of major and study time; no interactions because the lines are parallel
Why would a researcher choose a Small N Design?
Small N designs are important when the goal is to explore a few special cases rather than create broadly generalizable results.
Explain p as a conditional probability.
p(D|H0); the probability of getting the data/result you got if the null hypothesis is true
In this image, where would you look to establish covariance? https://drive.google.com/file/d/1dNDUCNqR_pHndh3Bmp8AOTLUCHv0R8zp/view?usp=sharing
Look at the cross-sectional correlations (vertical) to establish covariance.
Identify an internal validity threat in the following study: A group of researchers investigated the effect of mindfulness meditation on mental health workers, 10 weeks after a major hurricane. A sample of 15 mental health workers were pretested on their depression and anxiety symptoms. Then they engaged in meditation training for 8 weeks. After the training was completed, they were tested on their symptoms again, using the same test. The study found that anxiety and depression symptoms were lower at posttest (the 95% CI for the drop in scores excluded zero). The researchers concluded the meditation training helped participants.
Maturation (coping with stressful events naturally improves over time); regression to the mean (symptoms may have been worse at pretest due to chance; no comparison group)
What main effects and/or interactions does this bar graph show? https://drive.google.com/file/d/1033LmLUrAIuZK1YH9i4SAXIgFvkrI5lx/view?usp=sharing
Main effect of study time; no interaction because the pattern is the same
Describe reversal designs and compare an ABA model to an ABAB model.
An IV is introduced and removed at least once; in an ABA design, a baseline measurement is taken, another measurement is taken after the IV is introduced, and a third is taken when the IV is removed. In an ABAB design, the treatment is introduced a second time.
Why might a researcher choose an ANOVA instead of a t-test?
A t-test compares two groups; an ANOVA compares more than two groups.
In this image, where would you look to assess temporal precedence? https://drive.google.com/file/d/1dNDUCNqR_pHndh3Bmp8AOTLUCHv0R8zp/view?usp=sharing
Compare the cross-lag correlations (diagonals) to assess temporal precedence; are they different or similar?
Why can’t an experiment definitively confirm causation, and what can be done to increase the reliability of the results?
Internal validity can threaten causation; confounds can create alternative explanations for results. Increase reliability by replicating the study!
What main effects and/or interactions does this table show? https://drive.google.com/file/d/1nkniALdqWzHAL60Nt-DJvI6D54KEFe6H/view?usp=sharing
No main effects; there is an interaction
What kind of quasi-experimental design is the following scenario? A teacher wants to see if giving candy to students who answer questions will make them participate more. For a week, she gives candy to one class and teaches the other as normal; she measures participation by the number of students who raise their hands every day.
nonequivalent control group time series design; no random assignment
In plain English, what does Eta squared represent?
the amount of variance in the data that can be explained by the treatment effect; ranges from 0-1, with values closer to 1 indicating a greater strength in the treatment effect