🧬
Bio P5
Null hypothesis- no significant (difference/correlation) between ... and ...
not significant- due to chance, p>0.05, calculated value < critical
significant- not due to chance, p<0.05, calculated value > critcical
t test-
to compare mean/SD of 2 samples
continuous data
normally dist
similar SDs
10 < x < 30 readings for each sets of data
chi squared-
categorical data
find if theres a significant diff between expected and observed
pearsons linear correlation
2 sets of interval data
5 or≥10 pairs of data
normal dist/continuous
spearmans rank correlation
non linear relaitonships
ordinal data type
not normally dist
5 or 10 < x < 30 pairs of data
if correlation coefficient is closer to 1 strong, otherwise it is weak.
Significant- do not overlap, not due to chance. happens when calc value >
critical value
Bio P5 1
, - result caused by another factor
Not significant- overlap, due to chance. happens when calc value < critical
value
Confidence Level- shows accuracy of mean, 95% confident mean is in the
limits of interval
confidence interval- 95% of all data is expected to lie in the range, shows
reliability of mean.
Standard Deviations shows reliability of mean, spread of values from mean,
larger = less reliable.
Standard error (drawn on graph) is spread of sample mean from population
mean.
null hypothesis- there is no significant correlation between x and y. For
crosses/ inheritance- no significant correlation between the observed and
expected value for x.
p<0.05 : 95% confident due to other factor
why anamolous? not follow trend
relationship may be weak/strong positive/negative correlation or no relationship
(when data is scattered)
for control usign dyes, control will have no dye
conclusion not valid:
no replicates/ repeats
correlation does not mean causation
fewer than number of paired observations recc
narrow range of x used
test is not investigating what it's supposed to (generalising)
small sample size.
other factors x
not from same origin
range
different samples
Bio P5 2
Bio P5
Null hypothesis- no significant (difference/correlation) between ... and ...
not significant- due to chance, p>0.05, calculated value < critical
significant- not due to chance, p<0.05, calculated value > critcical
t test-
to compare mean/SD of 2 samples
continuous data
normally dist
similar SDs
10 < x < 30 readings for each sets of data
chi squared-
categorical data
find if theres a significant diff between expected and observed
pearsons linear correlation
2 sets of interval data
5 or≥10 pairs of data
normal dist/continuous
spearmans rank correlation
non linear relaitonships
ordinal data type
not normally dist
5 or 10 < x < 30 pairs of data
if correlation coefficient is closer to 1 strong, otherwise it is weak.
Significant- do not overlap, not due to chance. happens when calc value >
critical value
Bio P5 1
, - result caused by another factor
Not significant- overlap, due to chance. happens when calc value < critical
value
Confidence Level- shows accuracy of mean, 95% confident mean is in the
limits of interval
confidence interval- 95% of all data is expected to lie in the range, shows
reliability of mean.
Standard Deviations shows reliability of mean, spread of values from mean,
larger = less reliable.
Standard error (drawn on graph) is spread of sample mean from population
mean.
null hypothesis- there is no significant correlation between x and y. For
crosses/ inheritance- no significant correlation between the observed and
expected value for x.
p<0.05 : 95% confident due to other factor
why anamolous? not follow trend
relationship may be weak/strong positive/negative correlation or no relationship
(when data is scattered)
for control usign dyes, control will have no dye
conclusion not valid:
no replicates/ repeats
correlation does not mean causation
fewer than number of paired observations recc
narrow range of x used
test is not investigating what it's supposed to (generalising)
small sample size.
other factors x
not from same origin
range
different samples
Bio P5 2
