Question

Application Exercise:
Experimenter bias refers to the phenomenon that data tends to comes out in the desired direction even for the most conscientious experimenters. A social psychologist wants to confirm this phenomenon. The psychologist tells a sample of students that they will be the experimenters in a study, and are then told that all subjects in the study will be given sugar one hour before solving arithmetic problems; in reality none of the subjects were given sugar. However, half of the experimenters are told that sugar will lead to better performance and the other half are told nothing. The experimenters are then asked to score the arithmetic problems from the subjects. Below are the scores they gave. What can be concluded with α = 0.01?

told nothing	told sugar
12 14 13 8 14 17 20 10	18 15 21 12 17 21 21 24

a) What is the appropriate test statistic?
---Select--- na z-test One-Sample t-test Independent-Samples t-test Related-Samples t-test

b)
Condition 1:
---Select--- experimenter bias scores told sugar told nothing one hour
Condition 2:
---Select--- experimenter bias scores told sugar told nothing one hour

c) Obtain/compute the appropriate values to make a decision about H₀.
(Hint: Make sure to write down the null and alternative hypotheses to help solve the problem.)
critical value =  ; test statistic =  
Decision:  ---Select--- Reject H0 Fail to reject H0

d) If appropriate, compute the CI. If not appropriate, input "na" for both spaces below.
[  ,  ]

e) Compute the corresponding effect size(s) and indicate magnitude(s).
If not appropriate, input and/or select "na" below.
d =  ;  ---Select--- na trivial effect small effect medium effect large effect
r² =  ;  ---Select--- na trivial effect small effect medium effect large effect

f) Make an interpretation based on the results.

Experimenters that were told nothing gave significantly higher scores than experimenters that expected a good performance.

Experimenters that expected a good performance gave significantly higher scores than experimenters that were told nothing.     

There was no significant score difference between experimenters that expected a good performance and those that were told nothing.

Answer 1

a). Since sample sizes are small and they were divided into half before starting experiment, samples are independent.

Hence, we use t-test for independent samples,

b). Yes, experimenter bias the score by telling sugar has effect vs it has no effect to two groups

c). Ho: there is no significant difference between the two samples or

vs H1: the ability of getting data from subject told effect of sugar is greater than that of other
critical value = t₍₁₄₎ at α = 0.01, t_(14,0.01) = 2.624

test statistic =

For sample, told nothing, = 13.5, s₁² = 14.28

For sample, told sugar , = 18.625, s₂² = 15.125

Under Ho, , = = 3.83

then, = -2.675

Then, since |t_cal| > t_(14,0.01) (= 2.624)

Decision : We reject Ho.

d. For, α = 0.01, CI will be 99%

= (18.625 - 13.5) -/+ 2.977*3.83*0.5

=5.125 -/+ 5.7

(-0.575, 10.825) is the 99% CI for (told sugar - told nothing) at df = 14.

e. d = 5.125/3.83 = 1.338 (large effect)

f. Experimenters that expected a good performance gave significantly higher scores than experimenters that were told nothing.     

Please rate my answer and comment for doubt. There is a limitation of 4 sub-parts, but I have done all, hope u like it.