Research in Biomedical Sciences
summary (exam 1+2)
1
,Lecture 1 | Recap IBMS 3
Lecture 2 | Probability models 7
Lecture 3 | Hypothesis testing 10
Lecture 4 | Con dence Intervals 15
Lecture 5 | Statistical techniques I: Z- and Chi-square tests 18
Lecture 6 | Statistical techniques II: T-tests 21
Lecture 7 | Statistical techniques III: Correlation tests 24
Lecture 8 | Statistical techniques IV: Non-parametric tests 28
2
fi
, Lecture 1 | Recap IBMS
- Research process: research question -> hypotheses -> study design + data collection ->
descriptive statistics -> inferential statistics -> conclusion -> research question -> …
- Research question — the question you want to solve
- Hypotheses — predictions about the e ects you expect
- Study design — investigate RQ and whether there is support for hypothesis or not
- Select sample (representative subgroup of population)
- Methodology: study design + data collection
- Once data is collected, summarize them using descriptive statistics
- Inferential statistics — statistics built on probability models to assess to what extend
our observation in our sample is applicable to the population we’re interested in
- Statistics: inferential+descriptive statistics
- IBMS: methodology + descriptive statistics + probability theory
- RBMS: inferential statistics
- Variables: observable/hypothetical events that can change and whose changes can be
measured in some way
- Independent vs dependent variables:
- Independent - explanatory
- Dependent - response
- Study the e ect of X (independent) on Y (dependent)
- Independent = predictor; manipulated variable controlled by experimenter
- Independent is predictor when variable is not manipulated, but
observed/measured
- Dependent = outcome; observed e ect
- Causal e ects: directional e ects in which 1 variable (X) causes another
variable (Y)
- Extraneous and confounding variables:
- Extraneous/lurking variable: variables that are not of interest to the researcher,
but that might in uence the variables of interest if not controlled; variables that
provide an alternative explanation
- If controlled (e.g. kept constant / manipulated) -> OK
- If not controlled -> extraneous = confounding variable
- Levels of measurement:
- Categorical/qualitative: no meaningful interpretation of di erences; places an
individual into 1 of several groups/categories
- Nominal — variable represents a category without logical order
- If 2 categories -> dichotomous
- Ordinal — ranked variable: represents a category with a speci c order /
rank position
- Quantitative: meaningful interpretation of di erences; usually recorded in a unit
of measurement
- Discrete — counts; nite values
- Continuous — scale variable with in nite values
- Note on rating scales: ordinal vs continuous
- Rank order / on scale of 1-5 is ordinal
- Rank on scale of 1-100 is strictly ordinal (no meaningful interpretation of
di erences); yet better discrimination than mere order -> most researchers
take it as continuous
3
ff
ff ff fl fi ff ff fi ff ff ff fi
summary (exam 1+2)
1
,Lecture 1 | Recap IBMS 3
Lecture 2 | Probability models 7
Lecture 3 | Hypothesis testing 10
Lecture 4 | Con dence Intervals 15
Lecture 5 | Statistical techniques I: Z- and Chi-square tests 18
Lecture 6 | Statistical techniques II: T-tests 21
Lecture 7 | Statistical techniques III: Correlation tests 24
Lecture 8 | Statistical techniques IV: Non-parametric tests 28
2
fi
, Lecture 1 | Recap IBMS
- Research process: research question -> hypotheses -> study design + data collection ->
descriptive statistics -> inferential statistics -> conclusion -> research question -> …
- Research question — the question you want to solve
- Hypotheses — predictions about the e ects you expect
- Study design — investigate RQ and whether there is support for hypothesis or not
- Select sample (representative subgroup of population)
- Methodology: study design + data collection
- Once data is collected, summarize them using descriptive statistics
- Inferential statistics — statistics built on probability models to assess to what extend
our observation in our sample is applicable to the population we’re interested in
- Statistics: inferential+descriptive statistics
- IBMS: methodology + descriptive statistics + probability theory
- RBMS: inferential statistics
- Variables: observable/hypothetical events that can change and whose changes can be
measured in some way
- Independent vs dependent variables:
- Independent - explanatory
- Dependent - response
- Study the e ect of X (independent) on Y (dependent)
- Independent = predictor; manipulated variable controlled by experimenter
- Independent is predictor when variable is not manipulated, but
observed/measured
- Dependent = outcome; observed e ect
- Causal e ects: directional e ects in which 1 variable (X) causes another
variable (Y)
- Extraneous and confounding variables:
- Extraneous/lurking variable: variables that are not of interest to the researcher,
but that might in uence the variables of interest if not controlled; variables that
provide an alternative explanation
- If controlled (e.g. kept constant / manipulated) -> OK
- If not controlled -> extraneous = confounding variable
- Levels of measurement:
- Categorical/qualitative: no meaningful interpretation of di erences; places an
individual into 1 of several groups/categories
- Nominal — variable represents a category without logical order
- If 2 categories -> dichotomous
- Ordinal — ranked variable: represents a category with a speci c order /
rank position
- Quantitative: meaningful interpretation of di erences; usually recorded in a unit
of measurement
- Discrete — counts; nite values
- Continuous — scale variable with in nite values
- Note on rating scales: ordinal vs continuous
- Rank order / on scale of 1-5 is ordinal
- Rank on scale of 1-100 is strictly ordinal (no meaningful interpretation of
di erences); yet better discrimination than mere order -> most researchers
take it as continuous
3
ff
ff ff fl fi ff ff fi ff ff ff fi
