RM Exam Questions Flashcards
What are the rules for using parametric statistics tests such as the t test
Data must have been measured on a parametric scale (interval/ratio), the data must have come from a population with a normal distribution, the groups must have similar variance and have been randomly chosen
What are repeated measures design experiments? What are their advantages and disadvantages compared to indent groups
Repeated measures is where the same subjects are tested under each condition
However differences may be due to fatigue, learning effect etc
Explain the logic behind Popper’s idea and give an example
Poppers falsification referred to whether theories could be exposed to disagreement i order to be a theory.
Method of deduction
For example you can prove, it will rain today to be false’.
Limitations - theories are usually complex and mange current theories would have been discarded even though they are scientific
Define internal validity
Whether the manipulation of the IV really caused a change in the DV
Explain the terms normal science and paradigm shift
Normal science increases precision of measurement, is conservative, uses different conditions.
Paradigm shift is when evidence builds up too much that current theory wrong so whole theory must change
What are the 4 levels of measurement that data can have
nominal
ordinal
ratio
What is a meta analysis
A statistical method of combining the results of a series of independent , previously published studies carried out for the same general purpose
it is a type of systematic lit review
Give the equation for relative risk
incidence rate in exposed population / incidence rate in non exposed population
Define risk factor
A factor which is associated with prevalence of a disease
E.g inactivity
What does epidemiological research refer to
the study of the distribution of health related status/events in populations and the application of this to control health problems
distinguish prospective longitudinal surveys from retroactive
Prospective longitudinal surveys are where the same group of subjects are followed over time
retroactive is where a group of people are asked to recall matters from the past
describe the process of a meta analysis
research Q define criteria identify all relevant literature read full text and evaluate extract study characteristics calculate effect size apply statistical techniques to analyse significance
give 5 characteristics of research
logical systematic empirical reductive replicable
what did Kuhn name science which is conservative and adds clauses
normal science
what did Kuhn name science that occurs when the scientific community decides that anomalies have become too serious
revolutionary
what is research called which aims to become familiar with basic facts where there is little/no prior knowledge
exploratory
give threats to internal validity
selection
outside events
repeated learning (learning effect and fatigue)
john henry effect
problems related to the calibration of instrumentation
define external validity
whether you can generalise the results to other people/situations
what is a type 1 error
rejection of a true null hypothesis (found an effect when there isn’t one)
what is a type 2 error
rejection a false null hypothesis (believe there isn’t an effect when there is)
what is the difference between a systematic literature review and a meta analysis
a meta analysis is qualitative
a meta analysis combines the results of a series of independent published studies and this is a type of systematic review
Meta-analysis: data from papers are used for statistical analysis
what is the difference between incidence and prevalence
prevalence is the number of existing cases of a disease at a given point of time
incidence is the number of new cases of a disease in a specified time period
how do you calculate relative risk
incidence rate in exposed population/ incidence rate in non exposed population
what is the ability of a study to spot a real effect called
power
how do you calculate coefficient of variation and when is this used
standard deviation / mean x 100
used to compare variables with different units or when a larger mean will naturally have a larger spread and you want to correct for that (normalised measure of spread)
What would british rankings data type be
ordinal
how could you improve the design of a study to improve the internal validity
randomise order/counterbalance
this is a cross over design
what is the non-parametric test for comparing two conditions
wilcoxon signed rank
what is the Mann Whitney U test used for
non parametric for independent samples
what does it mean if p > alpha (0.05)
not significant
what does p stand for
‘p’ stands for the probability that the observed difference (between distribution data and normal distribution in this case) is coincidental.
What do the r and r2 stand for
Correlation coefficient and the coefficient of determination.
What does the value of r2 (= 0.81) mean?
The coefficient of determination tells us that 81% of the variability in jump distance can be explained by the variability in the run-up speed.
What is an informed consent, what does it need to include?
Informed consent is an ethical principle that requires obtaining the consent of the individual to participate in a study based on full prior disclosure of risks and benefits. Informing the participant, states that the participant can withdraw at any time.
if you increase the significance level, does the chance of a type II error then increase or
decrease as a result?
increase
What is meant by the ‘power’ of a test?
the chance of finding an effect, if there really is one
how can you make sure his data will NOT be affected by the John Henry effect
use a placebo or double blinding
What is this mechanism to increase questionnaire reliability called?
internal check
how do you calculate effect size
subtract the mean of one group from the other (M1 – M2) and divide the result by the standard deviation
A p value of 0.02 means that there is a 2% chance that …
the result is coincidental
How should Ann enter all the data in SPSS, considering she chose an independent groups design
one line per person; one column with the grouping variable (lane 1 or 8) and another with the
scores (RT)
what does a shapiro wilk p value of 0.034 mean
Since this is below 0.05, the distribution of the scores at baseline is deviating from the normal distribution.
What does the value of r2 (= 0.81) mean
The coefficient of determination tells us that 81% of the variability in jump distance can be explained by the variability in the run-up speed.
Why is it important to consider the order of the questions in a questionnaire?
Previous questions may direct person to certain answer (questions about crime,
then CCTV question versus questions about privacy, then CCTV question). Or: if more difficult questions first, people may not continue completing the questionnaire (funnel technique: first general/factual/easy questions, gender etc., then more specific about topic).
What is the difference between a review of several studies and a meta- analysis?
Meta-analysis uses data from a number of studies for statistical analysis (has methodology and results). Review contains description of studies, no statistical analysis.
What is inductive reasoning?
Generating theory by generalising from repeated observations. I see white swans => All swans are white.
Why is inductive reasoning problematic?
The main reason why inductive reasoning is problematic is that generalising from
observations is not logical. “The fact that the sun has risen every morning for as long as I can remember, does not guarantee that the sun will rise tomorrow.” (Hume) How many observations are needed to conclude that it will always happen? Arriving at scientific law through repeated observation is not logical therefore. Many of you mentioned that observations are subjective, which is also a problem.
Describe Poppers method
Falsificationism (2 marks). Method of deduction: define problem (usually a problem with current theory), proposed solution (new theory), deduction of testable propositions, tests (to falsify new theory), choose between old and new theory (2 marks).
If the data is NOT normally distributed, what statistical tests should be used
the mean and standard deviation are NOT appropriate for describing the central tendency and the spread, respectively (the standard deviation is linked to the normal distribution). The median and interquartile range are in this case much better.
