Fichas
Statistics 3 Data analysis
Fichas del conjunto (54)
- 1What are the 2 types of categorical variables?Ordinal
Nominal - 2What are dependant and independent variables determined by?Hypothesis
- 3What form does a scientific hypothesis take?'I think that hibernating a lot prolongs lifespan in hamsters'
- 4In the previous scientific hypothesis what is the dependant and independent variables?D: Lifespan
I: Time spent hibernating
(Lifespan depends on time spent hibernating) - 5What has to be considered in an observational study?Something else that may actually 'cause'/affect both varibales
- 6What do we know in an experimental study if there is an effect on the independent variable?It is the causal pathway for the dependant variable
- 7What is the equation for the t-statistic test?t = mean of group 1 - mean of group 2 / common standard deviation of the two groups
- 8What is the t-statistic?The ratio between the size of the effect and the amount of variation between individuals
- 9What is the term given to the difference between two means in the data?Effect size
- 10What is the variation around the means for the 2 sub data sets on a graph?Inter-individual variation or error
- 11What are degrees of freedom?Are a correction necessary for small sample sizes, where estimates of the true mean of the population are likely to be inaccurate
- 12What does the value of P have to be for there to be a significant difference?P<0.05
- 13What is a hypothesis that suggests that there are no differences or correlation between 2 variables?Null hypothesis
- 14What 2 things are studies normally based on?1. Scientific hypothesis
2. Study hypothesis - 15How can the p value be used to determine whether samples are from the same population?A p-value of less than 0.5 indicates that there is less than 1 in 20 chance that the samples are from the same population
- 16At P<0.05 what normally happens to the null hypothesis?Reject null hypothesis and accept study hypothesis
- 17What error are we making if we decide that two samples are from different population when they aren't?Type 1
- 18Does a significant result alwaysNo
- 19What type of error is being made when we accept the null hypothesis when it is not true?Type II
- 20What is teh most common reason for incorrectly accepting the null hypothesis?Too small a sample size of study
- 21What can be used to work out how big the sample size of the study has to be?Power calculations
- 22What do non-parametric tests make more likely?Type II error
- 23What is Wilcoxon Rank test and example of?Non-parametric
- 24What is the layout of a two-by-two factorial experiment?Has two factors each having two levels
- 25What can you make predictions from in a two-by-two factorial experiment?Scientific hypothesis
- 26Does a factorial experiment have to be two-by-two?No, can be more factors with more levels
- 27What allows more comparisons to be made in a factorial experiment?More factors
- 28What is needed if there is an increased number of factors to reduce the chance of making type II error?Increased sample size in the experiment
- 29What happens when you use more than one t-test to analyse a data set?Increases the chance of seeing a difference between the two of the groups
(1/20 / number of t-tests) - 30What is the 4 components that makes a linear model (or general linear model)?1. Constant
2. +the effect size of supplementation
3. + the effect size of infection
4. + the error term - 31What is a good thing about the general linear model?Very flexible
- 32Where the independent variable is categorical and the dependant variable is continuous what 2 data analysis can be used?1. For 2 categories use a t-test or a Wlcoxon/Mann Whitney (non-parametric)
2. For more than 2 categories, use a general linear model (parametric) - 33What is the name given when the same animal is used to make multiple measurements under different conditions?Paired data
- 34What has to be done so then the individual animals results can be compared rather than the means for the different conditions?Join the individual animals results
- 35When looking at the differences between the two observations, what will be seen if the differences are random?Their means should equal zero
- 36What is the basis for the paired t-test?-Differences are random, the mean should be zero
- Differences predominantly in one direction away from zero, the mean should be significantly different from zero - 37What 2 tests can be used where observations are paired?1. Paired t-test for two groups
2. General linear model for more than two - 38What is the issue with a graph of multiple sets of continuous data?Continuous variables on both axes
- 39If there is a relationship between the 2 continuous variables, what can be used to represent this?Straight line
y= mx + c - 40What is it called when a line is drawn to minimise all the variation around it?Regression analysis
- 41What does regression analysis minimise?Error term
(normally minimises error in y-axis) - 42What is important before drawing a regression analysis?Make sure the correct variables are on the right axis as the line can change depending on which way round the graph is.
- 43Which axis does the independent variable go on?x-axis
- 44What axis does the dependant variable go on?y-axis
- 45Give 2 features of the line produced by regression analysis.1. Maximises the amount of variation which is explained by independent variable
2. Minimises the error variation in the dependant variable - 46What two ways of data analysis can be used where both the dependant and independent variables are continuous?1. Linear regression (or general linear model for parametric data
2. Spearman rank test for non-parametric data - 47What situation can the general linear model not be used for?Where both dependant and independent variables are categorical
- 48What can be used to analyse data with 2 categorical variables?- Generalised linear model
- Chi-squared (xsquared) - 49What is needed to be able to carry out the chi squared test?Results we would expect to see according to the null hypothesis
(the expected values) - 50How can you use data presentation to trick people that your 'result' is correct?1. Using standard errors of the mean rather than standard deviations for error bars
2. Use restricted area of the graph
3. Aspect ratio
4. Removing the grid lines
5. Colouring the bars differently - 51What do you have to have in order to present your data in a more persuasive way?A statistical analysis which says the effect is real/supports the data
- 52What is statistical significance?A measure of whether we reject or accept the null hypothesis
- 53What is the result if we reject the null hypothesis?Imply that there is a real, mechanistic link between two or more measured variables
- 54What is EFSA's definition of a biologically relevant effect?"An effect considered by expert judgement as important and meaningful for human, animal, plant or environmental health."
Información
- Descripción
- BVSc (Statistics) 2. Animal management 2 Fichas sobre Statistics 3 Data analysis, creado por Rachael Jones el 25/10/2016.
- Creado por
- Rachael Jones
- Visitas
- 1
- Valoración