Validity Flashcards

1
Q

When is the correction for attenuation used?

A

It is most useful when a test developer wants to determine which test is the most promising one. Then they will attempt to improve its reliability.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

How do you interpret a correlation coefficient that has been corrected for attenuation?

A

The correlation between the true scores for X and the true scores Y; or the correlation between X and Y if there is no error of measurement in either observed score

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What cautions should be taken when using the correction for attenuation?

A
  1. It is important to make sure that the reliability estimates used in the formula are accurate. Otherwise, the correlation between the true score of X and the true score of Y will be misleadingly overestimated.
  2. The error of measurement that affects pXY should be the same error of measurement that affects *pXX’ *and pYY’ otherwise, the results will also be overestimated.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

How do you interpret a confidence interval for a prediction?

A

Confidence intervals can be interpreted as follows:

E.g., We are 95% confident that the predicted value will fall within this interval. Or if we created 100 of these intervals, then 95 of them should contain the predicted value.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Why is it informative to determine confidence-interval estimates in addition to point estimates?

A

It is more informative because this makes it clear that the predicition is contains error.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What errors are possible when screening people in dchotomous categories?

A

Type I Error = claiming something is true about the person when in fact it is false

Type II Error = claiming something is false about a person when in fact it is true

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Describe how to establish a cut-off score that maximizes the hit rate?

A

To find the cutting score that maximizes the hit rate, it is important to construct a testscore frequency distribution table for successful and unsuccessful people. And then try to calculate the different hit rates by varying the cut scores, after comparing the hit rates, the maximal hit rate and its corresponding cutting score will be obtained.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Describe how to establish a cut-off score that minimizes the misclassification of people who would be successful?

A

Lower your cutoff score

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Describe how to establish a cut-off score that minimizes the misclassification of people who would be unsuccessful?

A

Increase the cut-off score

How well did you know this?
1
Not at all
2
3
4
5
Perfectly