Identify the two basic concerns with using a correlation to measure split-half reliability and explain how these concerns are addressed by Spearman-Brown, K-R 20, and Cronbach's alpha
What will be an ideal response?
The two basic concerns with using a correlation to measure split-half reliability are (1) the correlation is based on scores from only half of the test items and therefore tends to underestimate the true reliability of the full test; and (2) many different correlations are possible depending on how you decide to split the test items in half. Spearman-Brown addresses the first problem by adjusting the correlation upward to correct for the underestimation. K-R 20 and Cronbach's alpha both compute a measure that estimates the average correlation that would be obtained from all of the possible ways to split the test items in half.
You might also like to view...
Cognitive learning is best described by all of the following EXCEPT:
a. became the dominant perspective at the end of the 20th century b. learning involves insight c. big names include Tolman and Bandura d. the learner ignores useful information from stimuli e. inferences are made about mental processes that are not directly observable
According to the chapter The Nose Knowsabout the scent of symmetry, what did the researchers find was
the relationship between FA and lifetime number of sex partners? (a) it was significant for both men and women (b) it was weak for men and significant for women (c) it was significant for men and weak for women (d) it was weak for both men and women