Statistical tests Flashcards
why do we use stats tests
determine whether or not a set of data differs significantly and if the outcomes were produced by chance
what is the null hypothesis
a statement which states that there will be no significant correlation or difference between the variables you’re investigating
what happens to the null hypothesis at the end of the stat test
you either accept or reject the null hypothesis depending on the result
what does significant correlation/difference mean
unlikely to have occurred due to chance
accepting or rejecting null hypothesis comes down to …
probability and chance
what value determines the probability of the correlation/ difference
probability value or P value
in biology the minimum p value is….
0.05 (5%)
when do we use student t-test
comparing 2 means of the same variable to see if there is a statistical difference
when do we use the chi-squared test
collecting data about no. of individuals in particular categories
when do we use the correlation coefficient (spearman’s rank)
looking for a correlation between DIFFERENT variables (see if there is a significant correlation)
the null hypothesis allows it to be…..
an unbiased position/view
how would you make a null hypothesis for the t-test
there is no significant difference between X1 and X2
how would you make a null hypothesis for correlation coefficient
there is no significant correlation between X and Y
where does the test statistic come from?
from using the equations specific to each test
what must I do with the test statistic
compare with the critical value from the table (degrees of freedom and p value)
how do you get the degrees of freedom in t-test
no. of values in variable 1 minus 1, then adding the no. of values in variable 2 minus 1
(n1 - 1) + (n2 - 1)
how do you get the degrees of freedom in a chi-squared test
no. of categories minus 1
when looking for test statistic which p value do we go to?
0.05 (5%)
how do you get the degrees of freedom in a correlation coefficient test
no. of pairs of groups
if test statistic is less than critical value….
accept the null hypothesis, no significant difference, there is more than a 0.05 probability that difference is due to chance
if the test statistic is greater than the critical value…
reject the null hypothesis, is a significant difference, there is less than a 0.05 probability that the difference is due to chance
the closer you get to -1 in spearman’s rank there is a stronger….
negative correlation
the closer you get to 1 in spearman’s rank there is a stronger…
positive correlation
how do you find the expected frequency in a chi-squared test?
add together the observed frequencies. Then divide that number by how many observed frequencies. i.e. 34+ 66 = 100 100/2 = 50 (expected frequency)