Flashcards
unit 9
Cards in this deck (30)
- 1What is the definition of intercept bias?the test consistently under-predicts or over predicts scores for a group. this test is not fair to use with some groups
- 2What is the definition of slope bias?the test is more predictive (higher validity) in some groups, and less predictive (lower validity) in others. the test is not valid for use with some groups
- 3What is the difference between mediation and moderation?Mediating explains why X predicts Y and Moderating explains when X predicts Y.
- 4Assume that you have a moderator variable that is a dichotomous group variable, and you follow the typical procedure for calculating simple intercepts and simple slopes. Using words, list (or name) all the simple intercepts and simple slopes you would calculate using this procedure.simple intercept for group 1 and group 0 AND simple slope for group 1 and group 0
- 5Assume that you have a moderator variable that is a numerically scaled variable (or a continuous variable), and you follow the typical procedure for calculating simple intercepts and simple slopes. Using words, list (or name) all the simple intercepts and simple slopes you would calculate using this procedure.convert M to a z score, could be arbitrary levels but can choose +/- 1 sd from the mean
- 6An interaction term is created by _________________ .multiplication
- 7Assume you get results from a regression analysis that includes an interaction term. What determines whether or not it would be appropriate to calculate and interpret simple effects?It would not be meaningful to report results from probing an interaction if the interaction was not significant
- 8—the equation does not include the variables that were used to create the interaction term
- 9What is a statistic that you need when conducting a mediation analysis that you need to estimate by using bootstrapping? Be precise in your answer.the standard error for the mediated pathway “ab”
- 10If I was conducting a mediation analysis and I was not allowed to use bootstrapping, it would limit my ability to ______________________ .test significance of mediated effect
- 11What are the six steps for using bootstrapping to calculate a standard error for the mediated pathway, “ab”?1. assume your sample is a model of the population
2. sample(with replacement), create a “bootstrap” sample (n=n)
3. calculate ab using bootstrap sample
4. repeat: create bootstrapped distribution of sd_ab
5. sort bootstrapped values of ab in ascending order
6. significance test at alpha = .05 (95% CI interval)
identify values of ab at the 2.5th and 97.5th percentile - 12Write three regression equations that can be used to estimate all the pathways that you labeled in the previous step. Use the same labels as you used in the previous step to indicate the parameters in your regression equations. In this way, it should be clear which pathways are estimated by which regression parameters.yhat=b_0+(c_1)(x)
mhat=b_0+(a_1)(x)
yhat=b_0+(c’_1)(x)+(b_2)(m) - 13There are 4 arrowed lines in the figure above. Using the traditional system for labeling pathways in a mediation analysis (as depicted in class notes), place a label next to each of the four arrowed lines in the figure above.
- 14Which label did you assign to the pathway that is equal to the total effect?C
- 15Which label did you assign to the pathway that is equal to the direct effect?C'
- 16Using the labels that you assigned to the pathways, indicate two ways that you could calculate the mediated effect.ab=c-c'
- 17for intercept bias:For each x and y variable in this equation, state the name of the variable.y is the criterion variable (attendance)
X1 is the test variable (motivation)
x2 is group variable (Depression vs SUD) - 18for intercept bias: State how each x variable is created (for example, is it transformed or scored in a particular way?)x1 is the score on the test variable
x2 is a group variable
scored 0 and 1 (use multiple contrasts if more than 2 groups) - 19for intercept bias: Define the meaning of the intercept in this equation.b0 (Intercept) predicted y when all x’s are 0 (when group=0 test=0)
- 20for intercept bias: Define the meaning of each beta weight in this equation. Also, state which beta would indicate intercept bias if it were significantb1 (slope 1) predicted increase in y given 1 unit increase in test score (when group=0)
b2 (slope 2) predicted increase in y for group 1 (when test score =0)
b2 is significant, it indicates intercept bias - 21regression equation for intercept biasy=b+0+(b_1)(x_1)+(b_2)(x_2)+e
- 22regression equation for slope biasy=b_0+(b_1)(x_1)+(b_2)(x_m)+(b_3)(x_int)+e
- 23for slope bias: For each x and y variable in this equation, state the name of the variable.y is the outcome variable (attendance)
X1 is the score on test (motivation score)
Xm = Depression vs SUD group membership (moderator)
(x1)(xm)= xint = interaction term - 24for slope bias : State how each x variable is created (for example, is it transformed or scored in a particular way?)X1 is converted to a Z score
Xm s scored as 0 or 1 for a group
xint is used the transformed x1 and xm and multiplying them to get the interaction variable - 25for slope bias: Define the meaning of the intercept in this equation.b0 is the intercept for group 0
- 26for slope bias: Define the meaning of each beta weight in this equation. Also, state which beta would indicate slope bias if it were significant.b1 (slope group 0)
b2 (cjange in intercept for group 1) (intercept bias)
b3 change in slope for group 1
b3 is significant, it indicates slope bias - 27In the example above, the beta for the “constant” is listed as -.11. State the meaning of this value.z score intercept for group zero
- 28In the example above, the beta for “Zmath” is listed as .10. State the meaning of this value.standardized beta for group 0
- 29In the example above, the beta for “group” is listed as .31. State the meaning of this value.sd units of change in intercept for group one
- 30In the example above, the beta for “Zmath_X_group” is listed as .42. State the meaning of this valuechange in standardized beta for group 1
Information
- Description
- graduate measurement Flashcards on unit 9 , created by abbi gutierrez on 23/03/2021.
- Created by
- abbi gutierrez
- Views
- 6
- Rating