StatPowers
Data Entry
?
✖
Enter data separated by comma, semicolon, space, tab or newlines
Repeated values: 4x3 becomes 3 3 3 3
Number of Independent Samples:
12
11
10
9
8
7
6
5
4
3
2
1
Variable Name:
,
units:
Text Areas
Columns
Group By
Groups:
Groups:
Data Value
Group
Group
Calculate Now
Clear All
Share Data
Summary Statistics
Normality
QQ Plot:
choose a variable
Group A
Group B
Group C
95% Bands
Worm Plot | GOF Bins:
Your browser doesn't support canvas. Please update your browser.
Charts
Histogram
?
✖
If your data is discrete (integers) the histogram will probably look better if you select "Discrete Bars" - however if the range of your data is too large you will end up with spikes and gaps. In this case it would be better to not use Discrete Bars. You can use one of the rules of thumb (Square root n, Sturges', etc.) to automatically choose the number of bins, or you can specify from 4 to 20 bins.
Discrete Bars (only for integer data)
Proportions
Overlay normal curve
Number of bins:
Your browser doesn't support canvas. Please update your browser.
Frequency Polygon
Discrete Bars (only for integer data)
Cumulative Number of bins:
Your browser doesn't support canvas. Please update your browser.
Empirical CDF
Your browser doesn't support canvas. Please update your browser.
Boxplot
Your browser doesn't support canvas. Please update your browser.
Dot Plot
?
✖
Neat will attempt to pack dots neatly. By row will stack dots.
Style:
neat
by row
bottom aligned
line
strip
Your browser doesn't support canvas. Please update your browser.
Stem and Leaf Plot
?
✖
The easiest way to make the stem and leaf plot is to specify the target number of stems and let the plot be auto-formatted. The stem and leaf plot is ideal for datasets of size less than 100 or so.
Auto Format
Target # of Stems:
Stem size:
0.00001
0.0001
0.001
0.01
0.1
1
10
100
1,000
10,000
100,000
1,000,000
Split stems:
None
2 Parts
5 Parts
10 Parts
Your browser doesn't support canvas. Please update your browser.
Ridgeline Plot
Kernel Function:
Uniform
Triangle
Epanechnikov
Gaussian
Bandwidth:
Auto
Overlap:
Opaque:
Your browser doesn't support canvas. Please update your browser.
Violin Plot
Interior:
empty
Boxplot
Mean+SD
Beeswarm
Strip Chart
Your browser doesn't support canvas. Please update your browser.
Confidence Intervals
Confidence Level
Mean Estimation
?
✖
Enter the confidence level above (a decimal smaller than 1.0) and a confidence interval calculated using the t distribution will be constructed.
The confidence interval should be interpreted 'We are X% confident that the interval accurately captures the population mean.'.
Use data
Enter summary stats
Group A
Group B
Sample Mean (x̄)
Sample Standard Deviation (s)
Sample Size (n)
Variance Estimation
?
✖
Enter the confidence level above (a decimal smaller than 1.0) and a confidence interval calculated using the chi squared distribution will be constructed.
The confidence interval should be interpreted 'We are X% confident that the interval accurately captures the population variance.'.
Use data
Enter summary stats
Group A
Group B
Sample Standard Deviation (s)
Sample Size (n)
Proportion Estimation
?
✖
Enter the confidence level above (a decimal smaller than 1.0)
The approximate (Wald) confidence interval uses the normal distribution to approximate the sampling distribution of p̂. This should not be used if the sample size is small enough that there are fewer than 10 successes or failures.
The Exact (Clopper-Pearson) interval uses a binomial distribution for the sampling distribution of p̂. It may be wider than it needs to be though.
The Wilson Score Interval is an improvement over the Wald approximation and performs well even for small sample sizes and extreme proportions.
The Agresti-Coull interval is cute approximation too the binomial interval; For a 95% confidence interval it uses the "add 2 success and 2 failures" rule of thumb.
Use data
Enter summary stats
Consider a success: X
<
≤
>
≥
=
Group A
Group B
Number of successes (x)
Sample Size (n)
Hypothesis Testing
Significance Level (α)
T Test for Mean
?
✖
The t test uses the t distribution to model the sampling distribution of the sample mean. \(t=\dfrac{\overline{x}-\mu_0}{s/\sqrt{n}}\).
Use data
Enter summary stats
H
0
:
μ
μ
A
-μ
B
=
H
1
:
μ
μ
A
-μ
B
<
>
≠
0
H
0
: μ
A
=···= μ
C
H
1
: At least one μ differs
Post Hoc Test:
No Adjustment
Bonferroni
Holm-Bonferroni
False Discovery Rate
Least Significant Difference Test
Tukey's HSD Test
Games-Howell Test
Show all
p
-values
Pool Variances:
None
Pairwise
All
Test for Proportion
?
✖
Two tests are done- one using the Wald approximation (normal approximation) to the sampling distribution of the sample proportion.
The Exact test uses the binomial distribution to calculate the
p
-value.
Use data
Enter summary stats
Consider a success: X
<
≤
>
≥
=
Number of successes (x)
Sample Size (n)
H
0
: π
=
H
1
: π
<
>
≠
.5
H
0
: π
A
-π
B
=
H
1
: π
A
-π
B
<
>
≠
0
H
0
: π
A
=···=π
C
H
1
: At least one π differs
Test for Variance
?
✖
(notes to come)
Use data
Enter summary stats
Group A
Group B
Sample Standard Deviation (s)
Sample Size (n)
H
0
: σ
2
=
H
1
: σ
2
<
>
≠
1
H
0
: σ
A
=···= σ
C
H
1
: At least one σ differs
Test for Median
?
✖
(notes to come)
H
0
: θ
=
H
1
: θ
<
>
≠
0
H
0
: Population medians are the same
H
1
: Population medians differ
H
0
: All population medians are equal
H
1
: At least one median differs
Test for Skewness
?
✖
(notes to come)
Kolmogorov-Smirnov Test
H
0
: Population distributions are the same
H
1
: Population distributions differ