Study Questions for Quiz #4

Please Note: For many of these questions the number of the link doesn't match the number I say when reading the question before giving the answer. The order-of-coverage of these topics was changed a few years ago, but I liked the audio & didn't try to edit away the numbers.

 

Assumptions

 

1 Describe “degrees of freedom” in general and tell the three different kinds we examined and the computational formula for each.

 

2 Discriminate between a frequency distribution, a sampling distribution and a mathematical sampling distribution.  Which is required for NHST? Explain why.

 

3 Discriminate between descriptive statistics, inferential statistics and inferential statistical tests.  Tell all of each that we’ve studied together in this class.

 

4 Describe the added complexity of NHST when we do not know the population variability.  What additional bit of information is part of a t-test that is not part of a Z-test for this reason.

 

5 Considering all that happens before we begin our NHST and the process of the NHST, what “assumptions” do we make? (Hint: this is really a question about kinds of involved validity you already know, plus some details about NHST assumptions we’ve just talked about together.)

 

 

 

Nonparametric Statistics

 

6 Why does the “level of measurement” matter to data analyst? Briefly describe each of the (4) “levels of measurement” usually applied to behavioral measures. What are two additional “kinds” about which there is considerable controvers, what is that controversy & why does it matter?

 

7 What are the two kinds of statistical models commonly referred to as “nonparametric analyses”? Which one is really nonparametric, and why isn’t the other?

 

8 Tell the three most common reasons that analysts choose to apply nonparametric tests.  What are the counter arguments for each?  What must the data analyst “keep in mind” when applying nonparametric models? How does one decide whether to apply an parametric or nonparametric model for a given analysis?

 

9 Describe the various analyses for qualitative response data we studied and when to apply each.

 

10 Identify the nonparametric model that most closely corresponds to each of the univariate and bivariate parametric models we have studied for analyzing quantitative response data.  What is the fundamental procedural/computational difference between the parametric and nonparametric models.

 

 

Meta Analysis

 

11 How have the goals and purposes of meta analyses changed?  How has this led to changes in the selection of what studies/effects to include in a meta analysis? How has this led to changes in the kinds of information we should include about each study/effect.

 

12 How many effect sizes should be extracted from an empirical study?  How has the answer to this question changed?  Why this change? How has this changed the information that is coded about each effect size?

 

13 Describe the different kinds of effect sizes we examined and when each is likely to be used.  What are the shared characteristics of a good effect size, for the purpose of a meta analysis? What can we do to “prepare” effect sizes before using them in meta analyses?  How do we decide which of these preparations to use?

 

14 What are “inverse variance weights” and why are they superior to the alternatives? What can we do to “prepare” effect sizes before using them in meta analyses?  How do we decide which of these preparations to use?

 

15 Tell the different kinds of meta-analytic “analyses” available and the information obtained from each. Differentiate between “fixed” and “random” effects meta analyses of each type and common difference in their result.  What are the ways one might use to decide which to apply? 

 

 

Back to the Classics

 

1.  Describe the fundamental difference between the intent and the practice of hypothesis testing research. The intent of psychological research is to provide definitive results that prove or disprove causal hypotheses about relationships between psychological constructs, so that the results can be broadly applied.

 

6.  Describe the (four) basic types of validity we want our research conclusions to have (be sure to tell which is dependent on the others and why).

 

34. How does the study of external validity inform our understanding of “sampling”?  What are the “kinds of sampling” that we must intentionally engage when planning our research and data collection?

 

33.  Distinguish between the attributes of a research study that directly influence the causal interpretability of the results and which do not  directly influence the causal interpretability.  What are the attributes of a research study that make it difficult to ensure ongoing equivalence and for what part of internal validity are they a problem?

 

2. What do we mean when we say our data analyses are "inferential"?  How do we know whether our "inferences" are correct?

 

7. Discriminate among the various types of statistical decision errors and tell the likely reasons for each.  How can we be certain whether or not we have committed one of these errors with a particular analysis?

 

A new Classic.  What question are we asking whenever we perform a bivariate NHST? How do we decide which statistical test to use? What do we do after we reach a statistical conclusion?