Wk 6 - Devising data

Question 1

TYPES OF DATA COLLECTION

i) name two types of data collection that can be used in primary research and broader health improvement projects?
ii) name three ways to identify data collection tools

Answer 1

i) questionnaires and interview/focus groups
ii) look at similar studies conducted previously, look what data is appropriate/high quality, look what data is freely available

Question 2

TYPES OF DATA

i) what is yes no?
ii) what is strongly agree - strongly disagree
iii) give an example of a question where the answers are multiple unordered catgories / nominal?
iv) what is continuous with likert data?

Answer 2

i) categorical - binary
ii) ordered categorical with likert data
iii) what is your ethnic background?
iv) on how many days in the past 10 have you eaten meat? 0,1,2,3,4,5,6,7,8,9

Question 3

LIKERT SCALES

i) give a question where it may be used?
ii) what does unipolar and bipolar scaling mean?
iii) name three things it can be used for
iv) what does it assume? (2)
v) how many points on the scale are usually used? why?

Answer 3

i) how often does something happen?

ii) unipolar - nothing to one extreme
bipolar - negative to positive extreme

iii) can be used for knowledge, belief, attitude, symptoms, traits
iv) assumes the distance between each response is equal and distance between each value is the same
v) 7 point scale - good because it can be treated as continuous but not too overwhelming for the respondent

Question 4

STATSTICAL TESTS

i) which type of test is used if participants participate only once and all data points are indep of eachother?
ii) which type of test if variables are normally distributed?
iii) which test if data is skewed? if data is transformed - which test can be used?
iv) which test is used if participants participate more than once? give an example
v) what is correlation/association tests used for? (2)

Answer 4

i) independent test
ii) parametric

iii) non parametric
- if transformed may be suitable for parametric test

iv) repeated
- eg pre and post intervention or follow up (data points are non independent eg nested within people)

v) correlation/associatoon - test within grooups
- Whether the level of one variable is related to level of another
- look at whether one variable influences another eg relationship between weight and calorie intake

Question 5

KEY STATISTICAL TESTS - DIFFERENCE BET GROUPS

i) which test is used to look at difference between groups when data is continuous if there are a) two groups, b) more than two groups?
ii) what do these tests compare?
iii) give an eg of when this test may be used
iv) what is provided as an end result (2)

Answer 5

i) a) t test
b) ANOVA

ii) compare the means between the two groups
iii) eg - weight loss for an intervention vs a control group in lifestyle intervention RCT
iv) mean difference and P value

Question 6

KEY STAT TESTS - ASSOCIATION

i) which test is used to look at association between groups in categorical/binary data
ii) what does this test compare?
iii) give an example of when this test may be used
iv) what does the test provide?

Answer 6

i) chi squared
ii) compares the observed frequency of a variable between the expected frequency (by chance)
iii) eg when looking at association between diabetes (pres/absent) and weight (over/normal)
iv) provides a P value

Question 7

SUMMARY

i) name three types of categorical data
ii) what do stat tests allow?
iii) what does a T test compare?
iv) what does chi squared compare?

Answer 7

i) categorical - binary (two groups, may be ordered), nominal (two or more groups, unordered), ordered categorical eg likert data
ii) test whether we can reject the null hypoth and make inference of samples to the broader population
iii) T test - compares the mean of one group to another group to establish if there is a significant difference
iv) chi sq - compare observed vs expected frequencies of one categorical variable at two or more levels of another variable (or between groups)