empirical estimates

Question 1

what are the assumptions of cronbach’s alpha?

Answer 1

• items are tau
• each items error term is uncorrelated with the others
• error scores are uncorrelated with true scores
• that the items used to generate are composite score measuring only one attribute

Question 2

what is the standardised coefficient alpha?

Answer 2

applied to scores that have been converted from a raw score to a standardised score

Question 3

what is the kuder-richardson 20 formula?

Answer 3

estimates the internal consistency reliability associated with composite scores based on dichotomously scored items

Question 4

what is the parallel reliability assumption model?

Answer 4

equal true score variance, equal means, equal error variance

Question 5

what is the tau reliability assumption model?

Answer 5

equal true score variance, equal means, unequal error variance

Question 6

what is the essentially tau reliability model?

Answer 6

equal true score variance, unequal means, unequal error variance

Question 7

what is the congeneric reliability assumption model?

Answer 7

unequal true score variance (all greater than 0), unequal means, unequal error variances

Question 8

is the cronbach’s alpha a lower or upper bound estimate?

Answer 8

lower

Question 9

are the reliability scores for cronbach’s alpha good?

Answer 9

it does underestimate reliability but the underestimate is barely negligible

Question 10

what happens to cronbach’s alpha if theres a violation in the assumption of non correlated error terms?

Answer 10

can result in cronbach’s alpha overestimating in the order of 0.5 to 0.6

Question 11

what is the max that cronbach’s alpha underestimates?

Answer 11

0.2

Question 12

what is the standard error of measurement?

Answer 12

represents the amount of error around a point estimate in standard deviation form

Question 13

what is the point estimate?

Answer 13

ones best guess of what a persons score is on the test (ie. the score you get on the test)

Question 14

what is a confidence interval?

Answer 14

reflects a range of values that is often interpreted as a range in which the true score is likely to fall

Question 15

what is observed score correlation?

Answer 15

the correlation you get based on the data you have

Question 16

is the observed score correlation compromised?

Answer 16

to the degree to which there is measurement in your data

Question 17

what is true score correlation?

Answer 17

a hypothetical correlation you can estimate, if you know the reliabilities associated with the scores

Question 18

is true score correlation compromised by measurement error?

Answer 18

no

Question 19

what is the r max?

Answer 19

the maximum possible correlation between two variables is equal to the square root of the product of their respective reliabilites

Question 20

what does r max test?

Answer 20

the theoretical correlation

Question 21

how high must reliability be for expiatory research according to nunnally?

Answer 21

.70

Question 22

how high must reliability be for basic research according to nunnaly?

Answer 22

.80

Question 23

how high must reliability be for applied cases according to nunnally?

Answer 23

.90-.95

Question 24

what needs to happen to an observed correlation before it is subjected to the correction for attenuation?

Answer 24

needs to be significant

Question 25

what is the correction for attenuation formula?

Answer 25

the ratio of the observed correlation to the square root of the product of reliabilities (rmax)

Question 26

what types of effect sizes are attenuated by imperfect levels of reliability?

Answer 26

nearly all types in one or more of the variables included in the analysis