IRT
Lakatosian
Regression Coefficients
Construct Building
Moderator vs. Mediator
100
IRT? What does it refer to?
What is a set of mathematical models that describe the relationship between a person's response to a survey question and his/her level of the latent variable being measure by the scale.
100
What is Lakatos' view of scientific progress from a historical perspective?
What is that theories are rarely abandoned rather that they are tenaciously held on to. Also anomalies are rarely falsifications of a theory. Progress doesn't happen in leaps and bounds rather it is a slow accumulative process.
100
Intercept, slope and score on the predictor.
What is b0,b1, and x1 in the line equation for regression.
100
The first important step to building a construct
What is a literature review. It helps us to understand past attempts at measuring the constructs and to have a clear and knowledgeable picture of the construct. It also helps to know how well previous (if they exist) measure measure that construct.
100
Explains a relationship between a predictor and an outcome variable.
What is a mediator
200
Define latent trait
What is a hypothetical construct, trait, domain or ability. It is thought to exist but cant be directly measured by a single observable variable or item. It is indirectly measured using multiple items or questions on a multi-item scale.
200
Describe the positive heuristic.
What is a decision of what way the theory is going to go. IT generates new facts and consists of models or suggestions for the auxiliary belt and how to handle anomalies. In an essence it helps the theory to continue to exist and to grow.
200
R square change
What is the amount of change from the previous model to the next model. It tells how much more variance is accounted for by the additional predictors.
200
Useful to do this before you go on to creating your scale/measure because it enables you to clearly describe the construct showing its breadth and scope.
What is writing a formal definition
200
Explain a moderator
What is an interaction where the effect of one variable depends on the level of another.
300
Define ICC
What is models the probabilistic relationship between a persons response to each category for an item and their level on the underlying construct. (Category Response Curve)
300
Define negative heuristic.
What is a way to decide which way the research program will not go. It protects the core of the theory from refutation.
300
Confidence Interval for the b coefficient.
What is 95% chance that the b values will be within the confidence interval.
300
Over-exclusivity and why we do it?
What is it is easier to get rid of something that we know doesnt fit than to add something that is missing. It also helps in finding out where the boundaries of the construct lie.
300
What is the difference in testing (in terms of when they are testing and how) for moderators and mediators.
Moderators are typically tested for when there are known inconsistencies within the data or between multiple studies. Mediators are typically tested for when there is a strong known relationship between to variables and the researcher wishes to know why/the reason for the relationship. They are both tested using regression. Moderators have the product term inserted and then a hierarchical regression is run. Mediators are tested using regression as well, however it is run three times to test the different relationships between all three variables. It is important to also test for the significance of the product of pathways a*b.
400
Define Item Discrimination.
What is the a parameter in the IRT equation. It describes the strength of an item's ability to differentiate among people at different levels along the trait continuum.
400
Define degenerative research program and tell me when or if it is ok to let go of such a program.
What is a research program which no longer capable of anticipating new facts, is no longer considered progressive (both in the theoretical sense and in the empirical sense). It is ok to have a degenerative program if there are no other programs to replace it. Should a program come along that describes everything that the degenerative is capable of plus more and is capable of incorporating new facts, then it is ok to let go of the degenerative program.
400
The definition of raw slope/standardized slope.
What is the change in the outcome associated with a unit change in the predictor.
400
When getting rid of items why should items with high endorsement probabilities be kept.
What is they can be indicators of extreme outliers and they are valuable as tools to find trait levels of individuals.
400
Partial Mediation?
What is more likely to be found in a mediator regression. It is when a mediator accounts for some of the variance but not all of the variance between the IV and the DV. It is very hard to find a full mediation between two variables due to the nature of the phenomenon that we are studying.
500
Define Threshold.
What is item difficulty. The b parameter in the IRT equation and it is the boint on the latent scale where a person has a 50% chance of responding 'no' to a item. Items with low threshold parameters are optimal for differentiating among persons with poor mental health functioning and vice versa.
500
These are found by looking into the past, one cant decide to just make one. In some ways they do not exist. Einstein is said to have done one.
What is a crucial test. Which is the turning stone from one theory to another IF it turns out to have provided a spectacular corroborating instance for the victorious program and a failure for a defeated one.
500
SE (standard error)
What is the amount of variance of our b-values. When the standard error is small it tells us that different b's from different samples are likely to be very similar due to a lack of variance.
500
Good and bad candidate items
What is a good item : it loads on the primary factor and minimally on the others. Bad item: cross-loads on one or more factors and loads weakly on all.
500
You that there is a relationship if there is an increase in the amount of variance accounted for in the IV.
What is the form of a significant moderated relationship.
M
e
n
u