You are on page 1of 5

390270

1.a) In this experiment, the null hypothesis is that agamine’s effectiveness will not be affected by the
presence of blockamine while the alternative hypothesis is that agamine will be less effective in the
presence of blockamine. Dr Wocky analyses his data with unpaired t-test giving p=0.063.
In the Fisherian ‘significance testing’ paradigm, P- value indicates the strength of evidence
against the null hypothesis, the smaller the p-value, the more confidence we are able to reject the
null hypothesis. The evaluation process of rejecting the null hypothesis takes into account of existing
findings related to the experiment. A P-value of 0.063 suggests limited evidence against the null
hypothesis. In this interpretation, it is inappropriate to reject the null hypothesis without further
investigating the drug effects. Further investigation can be comparing his result with existing
experiments or repeating the experiment. Increasing the experimental unit is a good way to more
accurately measure the real mean of both treatment groups; control and Blockamine, and so will be
able to help Dr Wocky to better interpret his data. Increasing experimental unit does not however
help him to infer if the change in effectiveness of agamine is practically significant.
Interpreting with the Neyman-Pearson ‘hypothesis-testing’ paradigm, Dr Wocky will be
unable to reject the null hypothesis with p=0.068, which is higher than the pre-set α=0.05. Neymon-
Pearson approach is used to limit false positive error rate, α. Since Dr Wocky has predetermined his
maximum tolerable α as 0.05(probably by convention), getting a P-value higher than predetermined
α indicate false positive error rate of higher than maximum tolerable and hence inability to reject
the null hypothesis.

b. The choice of whether to make a one tailed or two tailed t test should be determined before
observation is made.
Using one tailed p-value would give a smaller and more significant p value. This would allow him to
reject the null hypothesis in terms of Neyman-pearson hypothesis and increase strength against the
null hypothesis in terms of Fisher’s interpretation. Futhermore, his alternative hypothesis that
agamine will be less effective in the presence of blockamine implies the direction of the tail.
He must however has a strong reason to be sure that blockamine will reduce the effectiveness of
agamine. Otherwise, he risk being forced to accept his null hypothesis even if there is a significant
change in the opposite direction of what his alternative hypothesis says(ie. Agamine increase in
effectiveness in the presence of blockamine) and accept that the extreme in result is due to chance.







2.) A ) The two unlabelled rows should be labelled StDev(standard deviation), representing the
spread of technical replicates of each jar, before and after application of thickening compound.
B) The value for technical replicate 3, Jar 4 is 11.1, resulting in a high means of replicates and
standard deviation of technical replicate, which results in a high p-value. The value is unusually high
as it is higher than the viscosity of mayonnaise after mixing with the thickening compound. James
unneccesarily counted Stdev of technical replicate. In that, he calculates the accuracy of his
measurement instead of the initial or final mean of the viscosity of mayonnaise.
Viscosity unit of 11.1 could be misplacement of decimal point. This can be checked given access to
his notebook where the raw data is written.
C) Addressing the viscosity of technical replicate 3, Jar 4, if we found that it is indeed a misplacement
of decimal point, we can change the value to 1.11 and recalculate the mean of technical replicate of
jar 4. Using the mean of technical replicates to calculate the mean of initial viscosity of all 4 jars.
Using the means of technical replicates to calculate the StDev and using that to calculate P-value.
Technical replicates are only measuring the accuracy of viscometer and the consistency of
mayonnaise, which are not what the aim of experiment is. By measuring the mean of biological
replicates(mean of the mean of technical replicates of 4 jars), and comparing the before and after
treatment’s viscosity value, only then can we measure the effect of the new thickening compound.
If the raw data in his notebook indeed showed 11.1, I would suggest him to repeat his experiment.
D. The 3 forms of T tests are single, paired and unpaired T-test.
Using the values of means of replicates in the table and calculating using graphpad prism, we
obtain a p-value of 0.4434 which if rounded off to 2 decimal point will give 0.44, his value, this test
however is an unpaired t test. He should use paired t test as he is testing the effect of his agent
before and after addition into the same mayonnaises.











E.


The statistical analysis shows that the change in viscosity of mayonnaise after treatment is
statistically significant. The null hypothesis, the new compound does not make a change to the
viscosity of mayonnaise is rejected and the alternative hypothesis where the new compound
increases the viscosity of mayonnaise is accepted.

F. Protocol: Careful with the selection of mayonnaise to prevent unnecessary introduction of more
variables. This includes batch number and stores at which mayonnaise was bought. Some possible
variables include possible compound variation between batches, storage condition such as humidity
and temperature which may affect the action of thickening compound on mayonnaise. It is advisable
to keep in mind that this experiment does not aim to look at the effect of storage condition on
thickening compound but rather if the compound increases the viscosity of commercially prepared
mayonnaise. A different brand can be tested with the same protocol but in only interpreted as a
different set of data.
A control Jar with no treatment. Stirring might increase viscosity(?).
Analysis: Along the changes made in 2(E), (that is analysing the statistics using Mean of reps), I also
suggest to compare the change in viscosity of each jar and compare the mean of change in viscosity.
To use paired t test as we are testing the effect of thickening compound before and after addition
into the same mayonnaise

Amendment:

Before Jar 1 Jar 2 Jar 3 Jar 4
T.R 1.07 0.96 1.60 0.80
T.R 1.15 1.22 1.47 0.99
T.R 1.34 1.23 1.30 1.11
AVERAGE(1) 1.19 1.14 1.46 0.97


Mean(2) StDev SEM


1.19 0.20 0.10

After Jar 1 Jar 2 Jar 3 Jar 4
T.R 2.09 2.69 3.84 1.91
T.R 2.42 3.01 3.47 1.86
T.R
2.65 2.79 3.76 2.12
AVERAGE 2.39 2.83 3.69 1.96


Mean StDev SEM

2.72 0.74 0.37
Variable 1 Variable 2
Mean 1.19 2.72
Variance 0.04 0.55
Observations 4.00 4.00
Pearson Correlation 0.94

Hypothesized Mean
Difference
0.00

Df 3.00

t Stat -5.54

P(T<=t) one-tail 0.006

t Critical one-tail 2.35

P(T<=t) two-tail 0.012

t Critical two-tail
3.18


3.
A)

























Freddo-
frogamine
treated
Control

22 19.9

38.6 31.6

33.2 20.8

50.2 22.4

28.1 36.1

38 22.9

41.2 25.1

Mean 35.9 25.5
StDev 9.18 6.05
SEM 3.47 2.29
Df 6 6
t(a/2), a=0.05 2.4469 2.4469
Upper 95% CI 44.39 31.14
Lower 95% CI 27.41 19.95
Difference between mean 10.36
SE(difference between mean) 4.155
Df 12
t(a/2), a=0.05 2.179
Up 95% CI 19.41
Low 95% CI 1.30
Values calculated using Microsoft excel spreadsheet. t(a/2)
of a=0.05 with a degree of freedom of 6 is 2.4469
according to t distribution
table(http://easycalculation.com/statistics/t-distribution-
critical-value-table.php). CI calculated with Mean ± [t(a/2)
x SEM]. The 95% CI of Freddofrogamine treated is 27.41
to 44.39. The 95% CI of control group is 19.95 to 31.14.
SE(difference between mean) is calculated by the
following formula:
SE
(difference between mean)
= sqrt [ StDev
2
(Freddogamine)
/
n
(Freddogamine)
+ StDev
2
(control)
/ n
(control)
] = 4.155(3d.p.)
95% CI of difference between mean is calculated by
difference between mean ± [t(a/2) x SE(difference
between mean)] with degree of freedom 12 and t =2.179.
The CI is from 1.30 to 19.41.

B) Using t=

, we calculate t value to be 2.493. With a degree of freedom of 12, we obtain
a p value of 0.0283. (x1=Mean of freddogamine treated = 35.9, x2=Mean of control=25.5. SE
x1-
x2
=SE(difference between mean) = 4.155.)

C) The assumptions are random student sampling, unknown population mean and standard
deviation, both groups have normal distribution, equal variance and is unskewed. The bleeding time
for the control group will most likely resemble a normal distribution. The distribution is not likely to
be affected by factors such as age as we sample from a similar age population. Unknown population
mean and standard deviation are reasonable assumption as we did not sample the whole population.
D) The true means of freddogamine, control and the difference in mean falls within their respective
interval 95% of the time. Significance test shows a p-value of 0.0283 which has a strong evidence
against the null hypothesis that there is no difference between the groups. However, it is not
possible to conclude the anti-bleeding property of Freddogamine as the difference in mean has a
positive value, indicating an increase in bleeding time.
E) Conduct a linear comparison of all 14 student, before and after drug, dividing them into two
groups of either taking Freddogamine or placebo(vitamin pill). Subjects should be selected randomly,
within a narrow group range, healthy(not taking medication to prevent drug interference). Study
includes man only as hormonal cycle in woman might affect drug effectiveness.