Experiments Flashcards
whats the aim of controlled experiments
isolate a variable (which is diagnostic of a hypothesis) that we are interested in.
hall’s principle 1
dont experiment when we can find information through observation
hall’s principle 2
no experiment should be performed without a distinct and definiate objective
hall’s principle 3
dont needlessly repeat experiments
hall’s principles 4 and 5
must efficient and responsibilty to desgin and execute and report efficient experiments
hall’s principle 6
results be laid before the public in simplest and plainest terms
define an experimental unit
a subject or group of subjects upon which an observation can be made
when is an experimental unit required
both observational and experimental study
why does an experimental unit need to be independent
systematic difference between units will confound the experiment
blocking
the distribution of replicates in experiments to minimize any systematic difference between samples
randomisation
esures the sample is representative of a population
ensures equal chance of a unit being placed in each group
controls
comparison group
can be positive or negative
How effective is this treatment? will require what type of control
negative
is this treatment more effective? will require what type of control
postive
statistical power of 0.95
we should detect an effect in 19/20 experiments.
type II errors and statisical power
type II errors are directly related to statisical power
5% type II error rate is the standard in biology
nominal data
mutually exclusive classes
male/female
ordinal data
ordered categories
race order 1st/2nd/3rd
strategies to increase effect size
choose best measurment
larger stimulus
use more sensitive subjects
strategies to decrease variabiliy
improve uniformity of subjects or procedures
standardize the measured variable
improve experimental design
how to present data
Aim is to convey information as clearly as possible.
Interpretation and commentary describing results should be clearly separated (often supported by report format).
Eg. Legends/Titles should describe content but avoid interpretation.
Methods cross referenced to results.
presentation of raw data
It is good practise to plot out raw data as we have seen in earlier lectures.
We may well use statistical parameters to analyse & summarize these data but we need to show that they are truly representative.
Representation of precision
Use a logical unit and number of significant figures when reporting numerical data.
Error bars! Think about what’s appropriate (e.g. variability vs. precision) and report what was used– see earlier presentation
Imperative to describe precisely how these were derived from the raw data.
for genuine replication which three criteria must be satisfied
experimental units must be randomly and independently assigned to treatment groups.
The treatment(s) should be applied independently to each experimental unit. Injecting animals with a drug is an independent application of a treatment, whereas putting the drug in the drinking water shared by all animals in a cage is not.
The experimental units should not influence each other, especially on the measured outcome variables.