Vaux Lectures 1, 2 Flashcards
What formula is used to calculate SD?
SD=√((∑(X-M)^2)/(n-1))
What formula is used to calculate SEM?
SEM=SD/(√N)
What would be considered a descriptive statistic?
a. Standard error of the mean b. Standard deviation c. Confidence interval d. Poisson median
b. Standard deviation
What happens as sample size increases?
a. The mean gets further away from the true mean b. The range decreases c. The standard deviation stays much the same d. The standard error increases
c. The standard deviation stays much the same
What must a figure legend contain?
a. The value of N b. A description of the error bar used c. A calculated P value greater than 0.05 d. A and B
d. A and B
What might indicate that an image has been manipulated?
a. The background is clearly visible b. Edges are blurry c. The axes of a graph begin at zero d. Western blot bands have inconsistent “smiles”
d. Western blot bands have inconsistent “smiles”
Which statement about standard deviation is incorrect?
a. Many biochemical journals require that data be graphed as mean +/- SD
b. When N is small, the sample SD often underestimates the SD of the population
c. It can be calculated using the average of the squared differences between each sample and the sample mean
d. It is a robust statistic and mean +/- SD encompasses approximately two thirds of data points in a normal distribution
a. Many biochemical journals require that data be graphed as mean +/- SD
Consider the following experiment: you have triplicate cell cultures and repeat the experiment on four different occasions. What would be the value of N?
a. 4 b. SD x 3 c. 3 d. 12
a. 4
A binomial distribution:
a. Is also referred to as a Poisson distribution
b. Does not consider variability or sampling error
c. Adopts a bell shape when p is close to 0.5 and N increases
d. Is used for continuous variables only
c. Adopts a bell shape when p is close to 0.5 and N increases
d wrong because it is used for discrete variables
What is FALSE about a Poisson distribution?
a. It is seen when measuring discrete variables
b. It involves integer’s that are commonly less than zero
c. When plotted, the tail is longest towards the right
d. The expected SD is the square root of the mean
b. It involves integer’s that are commonly less than zero
Which distribution would be used for a continuous variable such as weight or height?
a. Poisson b. Normal c. Binomial d. Inferential
b. Normal
If a 95% Confidence Interval was generated 100 times by taking groups of samples from a population, how often would you expect to encompass the population mean?
a. 5 times b. 100 times c. 1 time d. 95 times
d. 95 times
A P value…
a. Is the probability of obtaining the difference you observed if you were actually sampling from a single population
b. Is the probability of getting the observed result if the null hypothesis were false
c. Is the probability that the gene or drug your testing does have a real effect
d. Is enough to conclude that a result is significant biologically as well as statistically when p < 0.05
a. Is the probability of obtaining the difference you observed if you were actually sampling from a single population
When are statistics useful in cell biology?
a. When presenting and interpreting data for publication
b. When monitoring the conduct of experiments
c. When deciding which experiments are worth repeating
d. All of the above
d. All of the above
What should NOT be included when preparing a western blot figure for publication?
a. Molecular weights
b. Loading controls
c. Cropped images of the bands with the background removed
d. A note stating that an image has been cut and joined (if applicable)
c. Cropped images of the bands with the background removed
Which statement is false?
a. A conclusion can only apply to the population where you took the sample from
b. Replicates are useless in biological research because they do not add to the N value
c. Experimental design is the art of varying one factor at a time while controlling others
d. A N value of 1 holds little value in statistical relevance
b. Replicates are useless in biological research because they do not add to the N value
When error bars between two conditions overlap, the difference between those conditions is not significant.
True
When p < 0.05, it can be concluded that results are significant.
False
Consider effect size, biological importance, underlying mechanisms
Descriptive error bars describe how the data is distributed whilst inferential error bars indicate the confidence in the mean.
True
A value of P > 0.05 means that no effect exists.
False
Effect may be small, may need to repeat experiment more
The shorter the length of an inferential error bar, the more “uncertain” the data.
False
If N ≥ 10, the 95% confidence interval is approximately M +/- 2 x SE.
True
When N = 1, it is critical to show error bars and the P value.
False
It is not appropriate to show error bars and statistics for replicates.
True
The SE becomes smaller as you repeat your experiment more and more because the SE varies inversely with the square root of N.
True
If a figure shows SE bars, you could mentally half the width of them to get approximate 95% CIs.
False
You could DOUBLE them to achieve this. And N needs to be greater than 10.
CI’s mean the same thing regardless of N.
True
The range and SD are the most commonly used types of inferential statistics.
False
they are descriptive
A high level of variability would lead to a wide range.
True
In Poisson distributions, the expected SD is the square root of the mean.
True
To ensure that a published figure is true and non-manipulated, it is best to print it out and view it on high quality paper.
False
As N increases, SEM also increases.
False
Representative results can be shown for numerical and nonnumerical data.
False
Only nonnumerical such as western blot
If you’re the author of a paper, you don’t need to read it yourself if you know the first author is thorough and diligent.
False
Contrast and brightness may be changed but only if it is done equally across a whole image and doesn’t lead to the removal of bands, spots or the background.
True
When showing a blot in a paper, it is best to cut out the region with the bands of interest to make it easier for the editor and reviewers to read through.
False
You don’t need to acknowledge that an image has been cut and joined as long as the join is neat and not noticeable.
False
