Devising data Flashcards
what is a data collection tool
a means of collecting data (qualitative and quantitative)
assessor/expert rated instruments
what is the use of ___
pt notes, records review
self-report surveys
qualitative interviews, focus groups
pt notes, records review (no pt identifiers to not breach confidentiality):
used for audit, service evaluation, quality improvement
limited types of info available
self-report surveys:
used for primary research, service evaluation (broader health improvement)
qualitative interview, focus group:
used for primary research, health improvement project (audit, service evaluation, quality improvement)
what is used to test internal consistency/reliability
cronbach’s alpha
using SPSS - data analysis software
what considerations are taken for data collection
logistics, cost
anonymity, data protection rules
consent
what are 3 types of data
categorical (binary = 2 groups, nominal = 2+ groups)
if ordered = likert
continuous (data on scale with equal intervals between points)
qualitative = non-numeric
when are open + closed Qs used
open = qualitative data
closed = categorical data (binary, nominal, likert)
why should questionnaires include many items (large spread of Qs)
to represent every aspect of the concept measured
reduce error of misunderstandings - use simple direct language, be specific, avoid leading Qs
this reduces skewing of results
what are the categories of statistical tests
test of difference with continuous data = t-test if 2 groups, ANOVA if 2+ groups
test of association/correlation with categorical data = chi-squared
test of difference with ordinal (ordered categorical data) = mann-whitney
how do t-test + chi-squared differ
t-test:
assess difference with continuous data
compares means between 2 groups
parametric test - continuous data so assumes variables have normal distribution
chi-square:
assess association with categorical data
compares observed + expected frequency
non-parametric test
what study calculates RR + odd ratio
RR = cohort study
odd ratio = case-control study
what measurement for precision + probability
precision = confidence interval
probability = p value