Why is labeling carefully important? What kinds of problems does it avoid? When is it sufficient to label the values on one but not on every item in a scale? Describe two guidelines for labeling variables and specifying values. You can run a ____________ test of internal consistency when you enter individual items from a scale.
Thinking back to Chapters 3 and 7 where the topic of power was discussed, remember that losing subjects affects the potential power of your research, that is, your ability to avoid a Type II error. The loss of power is one reason Allison (2001) said, “The only really good solution to the missing data problem is not to have any.” But if you do have missing data, what choices do you have to deal with it? In that regard, explain the difference between listwise deletion and pairwise deletion. When you replace data through imputation, what does that mean?
Explain the difference between data that represent out-of-range values and data that represent outliers.
What is opportunistic bias? How does it apply to the concept of an outlier? Are genuine experimental errors a justifiable reason to exclude outliers? If there is no evidence of genuine error and outliers exist, reporting the results with and without the outliers is recommended. Why is transparency in reporting your data “adjustments” so important?
In your own words, describe “going fishing” and p-hacking. How is p-hacking related to the file drawer phenomenon? Opportunistic bias deals with adjustments to data in the context of your original hypotheses, whereas HARKing deals with hypothesizing after the results are known. When HARKing occurs, you explore aspects of your data that were not originally hypothesized, and significant findings may emerge. Why does this process communicate a more positive image of science than is justified?
What does it mean to say that a finding is significant at the .05 level?
Describe the difference between Recode into Different Variables and Recode into Same Variables. Remember that you can use the Recode Function for col-lapsing categories and reverse scoring items. Be sure to keep track of the recoded items and use them when you calculate Cronbach’s alphas and compute scale totals.
