Lecture 20 (19)

Question 1

What is regression?

Answer 1

Using correlations between variables to make predictions

Question 1

What is the formula for regression?

Answer 1

Y = a + b1X

Question 2

What are you making a prediction about with regression?

Answer 2

This is what we’re making the prediction about. Aka making predictions about who isgonna win an elexction (percentage of votes the candidate will get.

Question 5

what does the y represent in the regression formula?

Answer 5

dependent variable, aka criterion variable

Question 5

What does the x represent in the regression formula?

Answer 5

X = independent variable, aka predictor variable

Question 5

what does a represent in the regression formula?

Answer 5

a = intercept, aka alpha This is indicating what the y is equal to when x equals zero.

when x = 0 what is Y

Question 5

what does b1 represent in the regression formula?

Answer 5

b1 = slope coefficient, aka beta weight

This implies when x increases by a certain amount, how much will y in turn increase. The larger the beta is, the steaper the relationship between x and y.

if x increases by 1 unit, Y increases by beta times X

Question 6

what does it usually mean when you have a lot of variability around the line of best fit?

Answer 6

it isn’t closely related

Question 7

What was the midtemr and smartbook quizzes example for regression?

Answer 7

He wanted to see if the number of quizzes completed correlates with the performance on exams. The completion of all of the quiz activities and the performance on the exam. predictor variable, how many quizzes did the student complete, criterion varaibel, percentage scored on the exam. The intercept means if you complete none of those smartbook activites, the intercept is the best predictor you have of how you will do on the exam. Then for each activity you complete, you get on average a 1.2 % increase on the score.

Midterm = Intercept + Slope*(# completed SmartBook)
Y = 60 + 1.2x

Question 8

What is the effect size for regression?

What about partial r squared

Answer 8

Effect size = R2 or total variance [in Y] explained Partial R2 = variance explained by single predictor
…given chosen regression model

Question 8

What was the multiple regression example provided? How could you identify which predictor is the strongest?

Answer 8

Y = Midterm score
x1 = SmartBook
x2 = Article Review
x3 = Lab completion

this intercept is the predicted score on the midterm exam score if they got a 0 on all of these predictors.

Y=6.0+.4x1 +.7x2 +.6*x3

They predict that for every smart book activity you complete you should expect a .4 increaso on the exam and so non and so forth.

Translation = getting more points on xj correlates with higher performance on midterm!

Y = Midterm score
x1 = SmartBook
x2 = Article Review
x3 = Lab completion

Y=6.0+.4x1 +.7x2 +.6*x3
This formula uses raw units

To compare different predictors, we need to standardize

Y=6.0+.04x1 +.04x2 +.22*x3
This formula uses standardized units

when each of these are standardized we can compare then against each other

peirson’s r is a standardized unit. Coen’s D is a standardized unit. When we standardize something we are divdiing that value by the estimate of the standard deviation.

you want to standardize to account for the variability of scores

Question 8

What is multiple regression?

Answer 8

to extend the example we could include mre predictors. There is no limit on how many predictors you can have. This then becomes mu;tipe regression. Each of the predictor variables get their own beta. Each will have a different relationship with the criterion variable. As we get more predictor variables, our prediction will improve. however, a lot of those correlations can be noise and that’s a problem. Random variability in predictor variables can correlate with random variability on the criterion variable. We want to find a tool to correct for this. The one that he really wants us to know is R squared. This is the square of the peirsons r (peirsons r multiplied by itself). the reason we use this is because ti’s standardized and the standardized variation allows us to compare this r squared with other r squareds. By standardized we mean it includes standard deviation. The higher the r sqaured the better your predictions about what the students midtemr score will be. The better iti will be clustered around the line of best fit.

Question 8

What is the formula for multiple regression?

Answer 8

Y=a+b1X1 +b2X2
X1 = predictor variable 1
X2 = predictor variable 2
b1 = beta weight 1
b2 = beta weight 2

Question 9

What is the logic to qualitative methods?
What is the assumption?

Example?

Answer 9

Logic: Assigning numeric values is not sufficient to capture multifaceted human behavior

Assumption: Quantitative methods are overly simplistic

Example: Quantify the way that a mother interacts with her child
What behaviors would be important?
How could we measure them?
Will we capture all of the important behaviors?

Question 9

Compare and contrast qualitative vs quantitative methods.

Answer 9

Quantitative:
Easily quantified behavior
Data is relatively simplistic
Can quantify uncertainty (aka significance)
Uses mathematical models (like mean, median)
Often artificial
Hypotheses are clear and specified in advance, a priori

Qualitative:
Includes complex behaviors
Data is rich, generative
Cannot quantify uncertainty
Uses verbal descriptions as models (also, could use mode)
Can be very realistic
Hypotheses continually revised

Question 10

Give a summary of naturalistic observation.

Answer 10

Method: Observe everything!
Includes video recording, audio recording, taking notes in a journal § Exploratory: Hypotheses developed as observation proceeds
Contrast with confirmatory
Confirmatory testing sets hypotheses & predictions in advance, a priori

Participant observation: Researcher is like a ‘participant’, occupying the same setting and experiencing the same stimuli as target

you may pretend to be an undergraduate student and enroll in a classroom if you’re interested in the undergrad experience.

Concealed observation: Researcher does not participate, but instead hides to prevent Ss reactivity
if we think us being present is too much of an interference.

Question 11

What was the backlash of the rosenthal experiment?

Answer 11

Rosenthal heavily criticized by psychiatrists

Follow-up study: Send us more confederates, and we’ll find them
Rosenthal: Challenge accepted!
250 patients arrive over the next months
Psychiatrists ID 41 patients as Rosenhan’s confederates
But…

Rosenthal actually didn’t send any confederates.

Question 11

What was the goal, method, procedure, and results of the experiment “on being sane in insane places”?

Answer 11

Goal: Describe a social system
Method: Participant observation

Procedure: 8 confederates
Faked auditory hallucinations
Admitted to 12 hospitals around U.S. § Then; symptoms fully resolve…

Results:
All required to sign admission of insanity (this is stigmatizing and is especially problematic to people who did not actually have a disorder.)

All given antipsychotic medications (were required to take these medications. )

Release was conditional (they were assigned a person who would check in on them on a regular basis)
Mean = 3 weeks following request to leave

Question 11

What was the When Prophecy fails study goal and method?

Answer 11

Festinger (1956): When Prophecy Fails
Goals: Describe situation, events, and group members Method: Participant observation

Question 12

What were the goals and method of the Grassian (1974): Psychological Effects of Solitary Confinement study?

Answer 12

Grassian (1974): Psychological Effects of Solitary Confinement § Goals: Describe and categorize individual’s reactions
Method: Qualitative but not naturalistic

to date we don’t have randomized control data about what solitary confinement does to human beings because it isn’t ethical.

Question 12

What is an example of a systematic observation?

Answer 12

Example: Milgram’s Obedience studies
Structure a situation where a participant believes they are hurting another § Observe participant behavior

How much electricity did they use?

What emotional expressions did they display? § How did they justify their actions?

Question 12

What are the 3 challenges with observational research?

Answer 12

What behavior is meaningful?
How should we measure behavior?
Are raters producing equivalent ratings?

Question 12

What were the goals and method of the Bauer (2016): Four Months as a Private Prison Guard study?

Answer 12

Bauer (2016): Four Months as a Private Prison Guard
Goals: Describe hiring process, training, and prison conditions § Method: Qualitative and naturalistic

Question 12

What is systematic observation?

Answer 12

Researcher structures a situation, then observes and/or quantifies Ss behavior

Question 13

What happens to external validity the more you control?

Answer 13

the more you control, the less external validity you have.

Question 13

Describe the strange situation:

Answer 13

7-stage test to investigate patterns of child’s attachment to their caregiver:
1. Mother + Baby B+M
2. Stranger enters: B+M+S
3. Mother leaves: B+S
4. Mother returns: B+M
5. Mother leaves: B
6. Stranger returns: B+S
7. Mother returns: B+