exam 2k17

Question 1

CONSORT flow diagram

“sampling by convenience”

Answer 1

when subjects are chosen based on convenience/availability rather than random sampling.
Eg the lecturer would of invited all students in his class to participate in the study
consequence - will likely not be representative of a wider population and so results may not be generalised

Question 2

CONSORT flow diagram

blind study

Answer 2

not blind study as participants would of been aware of receiving treatment or not

Question 3

CONSORT flow diagram

end results n=50) when they initially started at 120. gave no clear explanation of why this may be of concern/important

Answer 3

may have excluded outliers or values that did not match their hypothesis, thus distorting the results

Question 4

chi-square test

Answer 4

• What are the chances of seeing difference if the null was true
• allows us to compare observation to what we expect
• in the table p value is value in likelihood ration > asymptomatic significance

Question 5

chi-square test p = value 0.488

Answer 5

= no evidence to suggest a difference in gender distributions between the groups
importance = BEFORE INTERVENTION, WANT THE GROUPS TO BE OTHERWISE HOMOGENOUS - HENSE CONCLUSION SUPPORTS

Question 6

box plots

Answer 6

look at symmetry, spread

Question 7

box plot - why do compare change in grip strength/variable of each participant rather than directly comparing the means of the baseline and post-intervention measurements?

Answer 7

baseline and post-intervention values are not independent

Question 8

box plot - skewed meanings

Answer 8

positively skewed: median is closer to the lower/bottom quartile - meaning that the data consists of more higher frequency of high value score
- mean > median

negatively skewed = median is more towards upper quartiles and and higher frequency/quantity of low value scores (mean < median)

Question 9

definitions box plot

Answer 9

Inter-quartile range: 50% of scores fall within the interquartile range

whiskers = Whiskers
The upper and lower whiskers represent scores outside the middle 50%.

Question 10

p value for when give variety of SPSS output

Answer 10

go to table of independant samples t test

> row : equal variances assumed and colum sig (2-tailed) and that’s it

Question 11

mann-whitney U test

Answer 11

nonparametric alternative to the independant-samples t test

Question 12

independant t test

Answer 12

when distributions is roughly symmetric, t - test is more appropriate as it is generally more powerfull when assumptions are satisfied