Reference no: EM132369720

Assignment

QUESTION 1:

Read the journal article “Survey subjects and the quality of health surveys”. This was given to you as a handout in the first lecture. It is also available on CANVAS. Answer the following questions:

a. What is the difference between a probability sample and a non-probability sample? Give examples. What is the advantage of using a probability sample?

b. Using your own example, distinguish between a reference population and a source population.

c. What is a sample frame?

d. Name and explain the sources of error that affect the quality of health surveys indicating which errors are quantifiable and which are not quantifiable.

QUESTION 2

The data set ASSIGNMENT 1 QT2 DATA.savcontains data on a controlled trial on the efficacy of using acupuncture and pain medication to alleviate back pain. The control treatment is sham acupuncture/sugar pill which is placebo treatment. The clinical trial participants are randomly assigned to either acupuncture, pain medication or sham acupuncture/sugar pill treatment. 

The participants are assessed for pain on a 0 to 100 pain rating scale at baseline and then again at follow-up after 6 weeks of treatment. Hypotheses tests are not required if not specifically requested but APA reports on the analyses results obtained in answering stated appropriate research questions are required using α = 0.01.

In all APA reports in answering the questions below, please quote t or F values, degrees of freedom (df), p-values, 95% or 99% confidence intervals.

a. Explain the purpose underlying the random allocation of participants to treatments.

b. Comment on the strengths and weaknesses in the research design of this study.

c. Are there significant differences in pain scores between the three treatment groups at baseline? What does this result imply about assessing differences between the treatment groups on follow-up pain scores? (APA Report required).

d. Are there significant differences in pain scores between the three treatment groups at follow-up? What sample sizes are needed to achieve 90% power for this statistical test?  GPower output needed. (APA Report required).

e. Is there a significant difference between baseline and follow-up pain scores for the acupuncture group? What sample size is needed to achieve 75% power for this statistical test? (APA Report and GPower output needed).

f. Is there a significant difference between baseline and follow-up pain scores for the placebo group? What sample size is needed to achieve 75% power for this statistical test? (APA Report and GPower output needed).

g. Determine the correlation between baseline and follow-up scores for each treatment group separately (Use the Split File option in the DATA SPSS menu). Interpret the correlations, that is, what can you say about them?

i. Using the follow-up pain score as the dependent variable, the baseline pain score as the covariate and the grouping variable as the independent variable, conduct a one-way analysis of covariance on the data (APA Report required).

j. What is the advantage of including the baseline score as a covariate in the analysis of follow-up pain scores?

k. Conduct a multiple linear regression analysis using the follow-up pain score as the dependent variable and the baseline pain score and grouping variables as independent variables. State the equation of the regression line from the output. Interpret the regression coefficients of the grouping and baseline score variables in the output. What in the output would you use to assess impact on the dependent variable and why?.

l. What sample sizes are needed to achieve 80% power at a significance level of 1% using the one-way analysis of covariance conducted in (i) above? Evidence of GPower output should be provided.

QUESTION 3

Hyperactive children are often treated with a stimulant such as Ritalin to improve their attention spans. In one test of this drug treatment, hyperactive children were given a boring task to work on. A psychologist recorded the amount of time in seconds each time the child worked on the task before becoming distracted. Each child was measured before receiving the drug and again after the drug was administered. The data are tabled below.

CHILD surname first letter	BEFORE DRUG	AFTER DRUG
A	28	135
B	15	309
C	183	150
D	48	224
E	30	25
F	233	345
G	21	43
H	188	110
J	12	15

a. Use a parametric test to assess if there are a significant mean difference of 65 in the attention times in the two treatment groups. Use α=0.05. Hypotheses and conclusions in appropriate format are essential.

b. Repeat (a) above by using a non-parametric test via SPSS.

c. What is the value of the Type 1 error in this study? Use GPower to estimate required sample size to achieve 90% power. What is the Type 2 error which achieves 90% power?

d. Give an account using the Table of Random Numbers on how you would obtain a simple random sample of 6 children from the 9 who participated in the treatment. List the children selected in your simple random sample by the first letter of their surname.

QUESTION 4

Read the journal article Confidence intervals rather than p-values: Estimation rather than hypothesis testing and then answer the following questions:

 

a. State the difference between standard error and standard deviation.

b. Summarise in your own words what the authors say about confidence intervals.

c. Explain in your own words the drawbacks of presenting p-values.

QUESTION 5

The following table is taken from a journal article that summarises statistical analysis carried out on some data.

Outcome variable: Scores on a depression scale.

Influencing variables	B	se(B)	β	t	p-value
Age	3.2	1.6	0.6		0.025
SES	4.6	2.3	1.3		0.01
Parents average years education	2.0	1.2	1.5		0.07
Anxiety score	6.3	2.0	2.5		<0.001

1-R2= 0.13. SES= Socio Economic Status (Low, Medium, High); Model Fit: p= 0.026

a. Determine the t-values in the table.

b. Explain how the β values are calculated.

c. Write a concise APA report on the statistical analysis results displayed in the table.