introduction + sampling design Flashcards
what are statistics?
the collecting and analysing of numerical data to test a sample.
what are the 2 types of statistics?
quantitative and categorical
what is quantitative statistics?
the measure in quantity, numerical values that represent different magnitudes of variable.
what is categorical statistics?
each observation belongs to one set of categories.
what do the graphs look like in a categorical set of data?
more ordered than quantitative
what do the graphs look like in a quantitative set of results?
dotted around
what are the 2 parts of quantitative statistics?
discrete and continuous
what is discrete statistics?
values from a set of separate numbers
what is continuous statistics?
possible values from an interval
what are the 2 parts to quantitative statistics?
nominal and ordinal
what is nominal statistics?
the values arranged in groups
what is ordinal statistics?
ordered in ranked data
what is descriptive statistics?
to describe and understand the property of your data.
what does descriptive statistics tell you about?
central tendency (mean, median etc), and the measurement of variability
what is inferential statistics?
methods to test research of hypothesis
why do you need statistics?
to see trends in data, to check likelihood, to make research quantifiable, verifiable, defensible etc.
what do you have to do in a scientific method process?
make an observation, define a question, develop a hypothesis, design an experiment, collect data, analysis of data, develop theory, if reject- go back to designing the experiment.
what is an alternative hypothesis?
clearly stated proposition which should be tested and falsifiable
what is the null hypothesis?
statement which is accepted or rejected in favour of the alternative
what are the aims for study design?
examine how changes in one or more treatments can change. does x affect y etc.
what is an experimental study?
manipulation of treatment of interest under controlled conditions, random
what is an observational study?
observations of different levels under interest. occurs in the natural sites using natural controls.
what do you need more than one treatment or control field?
need for replication, the world is noisy and hence there are lots of other factors which your aren’t controlling.
why is random sampling good?
enables you to quantify how much of an effect there is, and how likely it is to be real. its not biased, and the maths behind it is fair
what is simple random statistics?
random selection from the total population
what is stratified random statistics?
splitting up your population according to treatment, and then randomly selecting within each “strata”
what is important about the independence of samples?
no sample should be influenced by another sample.
what is pseudoreplication?
when individual samples are heavily dependent on one another.
what does pseudoreplication cause?
means that your sample size is smaller than what you think you have.
