Statistics Flashcards
what are stats tests used for
to determine whether differences between data sets are significant or if it is due to natural variation.
how do u plot the data u collect in stats
in a normal distribution curve
what is the mean
the average value and it is the maximum height of the normal distribution graph
what is the standard deviation
the spread about the mean
to what confidence/percentage do biologists use for the s.d.
95% = 1.96 s.d.
95% of pop lie within this value
how can u use a graph to visually see the variation
if the curve is wide then it means there is a lot of variation in the sample
if the curve is quite narrow there is not much variation
in a normal distribution, what is the percentage of the pop that lies within one s.d. above pr below the mean
68%
how much of the pop lies within 2 s.d. above or below the mean
95%
how is standard deviation a better way to measure the spread than the range
it is less affected by extreme values
how can standard deviation values be used to plot error bars
plot one standard deviation above the point an one below so that the total length of the bar is 2 standard deviations
how to work out s.d on calculator
press mode
then press 2
press 1 ad go on table
put in values
press AC
shidt then no 1
no. 4
and then no. 4
which eq should u always use for bio
use the sample one = n-1
example calc - mean IQ is 100 with s.d. 15
a) bwt what IQ’s do 95% fall
b) what proportion have IQ over 130
mean = 100
s.d. = 15
95% lie with +/- 2 s.d.
95% of pop 100 +/- (2x15)
IQ range 70 —-> 130
b) 2.5%
what is a working hypothesis
idea of what you think might happen
what is a null hypothesis
what you think will not happen
What should u make sure that u do before starting analysis
write down null hypothesis
what probalilty do u work with
0.05 probability level = P=0.05
what results of stats are we looking for
95% chance it is significant and not due to chance
or 95% that result was due to chance
what test do u use if u need to compare two sets of data when u have 30 values in each data set and the data is normally distrubuted
the t-test
what does the t-test find
wether the difference between the means is significant
what are the two types of t-test
unpaired
paired
what is an unpaired t-test
used for comparing the means from two groups of different individuals eg the height of limpets in two different habitats
what is the paired t-test
this is used when the data is from the same individuals eg measurement taken from the same people before and after a drug test
will the t value be larger or smaller if the results are significant
larger
will the t value be larger or smaller if the results are not significant
smaller
how to carry out an unpaired t-Test
-start with null hypothesis - there is no significant difference between the means no of …
-calculate the mean for each data set
-calculate the standard deviation for each data set using the formula/ calculator
-put the mean and standard deviation values into the t-Test formula to calculate t value
-calculate the degrees of freedom
-look up critical value at P=0.05 and compare with calculated p value
-conclusion
what is the degree of freedom
number of samples - 1
what is the conclusion if the tcalc >/ tcrit
-null hypothesis is rejected
- there is a significant difference between the two mean values
-with less than a 5% probability of the results being due to chance
what is the conclusion if the tcalc /< tcrit
-the null hypothesis is accepted
-there is no significant difference between the two mean values
-there is a 95% probability of results being due to chance
what is the spearmans rank correlation coefficent
used to look for associations between different measurements from the same sample
-use to test the strength and direction of the correlation
how to set up spearmans rank
-set up a table
-place the values from each site into columns
-rank each - the most being 1
-is there are two values add next two ranks together and divide by 2
-minus the two ranks from each other -A-B
-square this value
-add these tg
-sub into eq
1- (6x sum of D2/ n^3 x n)
next steps after calc spearmans rank
find on critical value of coefficent
interpret stats
calc is bigger then reject null - sig correlation
crit bigger then accept nul - no sig correlation
what is the chi-squared test
looking at frequencies of individuals in different categories - difference bwt predicted and observed is significant
what are chi-squared tests often used
genetic crosses
what is the chi-squared formula
sum of (O-E)^2/ E
how to set up chi-sqaured table
obsereved no and total
expected no and total
O-E
(O-E)^2
(O-E)^2/E and then total
how to find expected values
divide total of observed by no of locations
how to find expected values for more complicated chi-squared
times total of site 1 by total column over pverall total
chi-sqaured step 2
find degrees of freedom
for compliacted ones - ( number of rows-1 ) x (number of columns -1 )
then find critcal values
and use same conclusion for others
