HBS Statistics Mean Values Being Affected by Outliers Questions

Culture 1 Culture 2 Culture 3 Culture 4 Culture 5
Time (hr) (µg/ml)
(µg/ml)
(µg/ml)
(µg/ml)
(µg/ml)
0
15.04
16.22
10.03
10.2
2.41
3
15.37
17.6
22.56
23.69
18.53
6
34.86
35.09
32.99
28.79
28.41
9
86.24
57.91
46.93
46.61
58.31
12
54.38
54.07
57.96
51.58
56.57
20
68.77
82.21
81
79.12
75.95
24
94.04
94.09
88.94
81.46
94.5
Student Name
Parts
a. Calculate means, standard deviations, and standard error for each row of data and label this colum
b. Follow relevant online instructions to create a scatter plot (chart) with error bars and trendline usin

https://support.office.com/en-us/article/Add-error-bars-to-a-chart-2072a6d5-0b44-418bhttps://www.youtube.com/watch?v=JeqCl_aD_8Y
Change the chart title to “Protein X Concentration”.
Add “Culture Growth Time (hr)” as the x (horizontal) axis title.
Add “Mean µg/ml protein” as the y (vertical) axis title.
Use your standard error column for the vertical error bars.
Add a textbox saying “error bars = 1SE. n = 5 for each time point.”
Add a trendline.
Calculate the correlation to and the slope of the regression line and add them as a textbox
You can use the Analysis ToolPak Regression tool or worksheet functions (CORREL, LINEST,
https://support.office.com/en-us/article/Use-the-Analysis-ToolPak-to-perform-complex-da
Highlight the most aberrant point with a callout.
Identify the outlier and calculate the means without the one outlier in the data. Label this c
Add “Mean without outlier” as another series to the graph.
Add another trendline for this new series.
Copy the graph and change the vertical errors bars to one standard deviation unit in the ne
c. In this instance, why is one standard error unit not a good choice for the error bars that relate to the
d. What would be a better alternative to using standard error and standard deviation as the error bars
this alternative. But in this statistical situation, there is an even superior solution to using error bars. h
e. Why is the original trendline NOT significantly deviated by the outlier? Under what circumstances w
make a copy of the spreadsheet and try changing data values and testing their effect on the trendline)
f. Why are the mean values more affected by outliers than the trendline slope and y-intercept?
h row of data and label this columns accordingly.
with error bars and trendline using mean and standard error calculated for datacollection1:
to-a-chart-2072a6d5-0b44-418b-9234-7e683798e41b
on line and add them as a textbox to the graph.
ksheet functions (CORREL, LINEST, etc.)
s-ToolPak-to-perform-complex-data-analysis-6c67ccf0-f4a9-487c-8dec-bdb5a2cefab6
one outlier in the data. Label this column as “Mean without outlier”.
e standard deviation unit in the new graph. Update the textbox to indicate error bars are 1StD.
or the error bars that relate to the trendline?
andard deviation as the error bars in this statistical situation and why? Create the resulting graph with
rior solution to using error bars. https://doi.org/10.1083/jcb.200611141
lier? Under what circumstances would the trendline be significantly deviated by an outlier? (Hint:
sting their effect on the trendline).
line slope and y-intercept?