Lab Final Review - Data Presentation Guidelines Flashcards
When would you graph your data?
When you want to emphasize or illustrate a point you want to make
Should you always present individual data points?
No, when possible it is best to report averages
When should you insert error bars?
If multiple measurements of each condition were made and averages calculated. Error bars should be explained in the figure legend
When is it appropriate to use a bar graph?
When the data is discontinuous or qualitative
When is it necessary to use a line graph?
When the data is continuous - e.g. fractions are possible
How should you connect data points if there are erratic jumps due to sampling limitations?
You should not connect them literally, instead you should draw a smooth approximate line
On which axis should both the independent and dependent variable be plotted?
horizontal - independent
vertical - dependent
How should the axes be labeled?
With what variables they measure as well as the units in which they are expressed
Why should the scales on your graph begin at zero when possible?
So as not to over-dramatize the difference between the results being compared
What is considered an outlier in data?
data that decreases greater than 100%
Why is best to graph the most important comparisons on the same graph?
In order to facilitate comparison of it
When is it okay to use the word ‘significant’ when referring to differences on a graph?
Only when statistical significance has been determined via statistical analysis
Why is it important to plot values for negative controls?
It reveals the ease or difficulty with which hypothesized, or tested
What is the purpose of a figure legend?
To succinctly inform a reader that is not familiar with the experiment described in the graph, what its purpose was, what you concluded about that purpose and how you arrived at that conclusion
Why is good to calculate the percent change in data rather than the absolute change?
Because it corrects for individuals with widely disparate starting values
Why is it important to indicate the type of statistical analysis?
Because there are many different types to choose from
What is N
The number of subjects whose mean is being reported; used in statistical analysis
What is a P value?
The degree of confidence you have in your statistical conclusions
How can you be mindful of significant digits?
By rounding, so as not to report more significant digits than were actually measured
Can statics and error bars be used for identical replicates of data?
No, it should only be used for independent data
In order for the results of an experiment to be justified multiple independent samples or ___________________________________
repeated observations or experiments must be done
If a paper has error bars, the author is required to ?
describe what they represent
Statistics do not apply when?
N=1-3
What should be done if the data in an experiment are equivocal or the effect size is small?
Come up with an extra mechanistically different experiment to test the hypothesis, rather than repeating the experiment until it is statistically significant
When are descriptive statistics necessary?
Only when there are too many data point to easily visualize
When should inferential statistics be shown?
only if they make it easier to interpret the results
Why should numerical data be made available?
So that readers can perform and confirm statistical analysis themselves
