OCTOBER 2011
IN THIS ISSUE
- Visit us at Neuroscience
- We are hiring
- Data analysis insights
- Try LabArchives
- Axis gaps
- Creative layouts
CURRENT VERSIONS
CHECK FOR UPDATES
Check
here
QUICKCALCS
Free
online
calculators for radioactivity calculations, detecting
outliers, t tests, ANOVA post tests, and more.
CONTACT US
GraphPad
Software
2236 Avenida de la Playa
La Jolla, CA 92037
USA
A Capitol Event
Stop by and see us at the Society for Neuroscience meeting November 13-16 in Washington, D.C. We'll be in booth #225. GraphPad founder and CEO Harvey Motulsky will be there. Stop by and let him know what you think of Prism.
Looking for an alternative career in science?
We take great pride in providing our customers not only with software, but also extensive information and support to help scientists learn statistics and curve fitting. Hundreds of thousands of scientists use GraphPad products, and many rely on us to help them properly analyze and present their scientific data. We are looking for an extraordinary person to join our team to play a key role in this extremely important aspect of our business. Critical thinking, enthusiasm, and an inquisitive mind are musts. You will deal with unique questions/challenges/opportunities every day from customers around the world, and help us improve our products and processes accordingly. Your work here will have an impact – you will help many thousands of scientists do better science. Interested? Apply here.
We regularly post new data analysis articles to the GraphPad web site. Here are some of Dr. Motulsky's recent items.
Binding of fluorescent ligands. Fittin
Graphing data expressed as ratios. Many kinds of data are best expressed as a ratio of a measurement in a treated group divided by the measurement in a control group. The problem with ratios is that they are inherently asymmetrical. A ratio of 0.1 is logically, but not arithmetically, symmetrical with a ratio of 10.0 (around 1.0, which means no change). The solution is logarithms. Read more.
How to get more statistical power. If you calculate the statistical power of your experimental design, you will often be disappointed in how low the power is. Read a list of approaches to increase power.
Why post-hoc power analysis is futile. While power analyses are useful when designing experiments, they are sometimes used when an experiment is done to ask "What is the power of this experiment to detect the results we obtained?". Read why these analyses are not helpful.
Is the 95% confidence interval wider or narrower than the range of the data? It depends on sample size.
Binding potential. Some ligand binding experiments (especially with PET ligands) never come close to saturating the receptors, so it is impossible to determine the Bmax and Kd. But you can determine their ratio, which is called the binding potential. Read how this is useful.
Q&A about normality tests. Normality tests are less useful than most people believe or hope for. Read why.
If you are using Prism Version 5.04 or 5.0d (or even if you're not!) and haven't yet looked at the software from our partner company, LabArchives, we would suggest that you give it a try. As you may know, LabArchives is a web-based (can also be run on a local server) product that enables you to store, organize, share, and publish your laboratory data. It integrates with Prism so that you can directly store your Projects into your lab notebook, and through special arrangement with the company, Prism users get a special Free version that is limited only in the amount of available storage. You can upgrade at any time to the reasonably priced Pro Version. Sign up for the free Prism user edition here. Also, if you are teaching any laboratory courses, LabArchives has recently made available a special Student Edition. Designed for undergraduates, this version provides all the power of the Pro Edition, but is designed for use throughout a semester or academic year. Not only does it make it far easier to review the work of your students, but it costs less than a paper notebook. Have a look.
Click on the graphs below to learn how to place a gap in a graph axis, and how to combine grouped bars with XY points and lines in one graph.
Using layouts to make composite graphs -- An individual Prism graph can only have one X axis and either one or two Y axs. You can be creative, however, and combine multiple graphs in a layout to create graphs with more that one X axis or more than two Y axes. (Learn more)
GraphPad Prism, InStat and StatMate are trademarks of GraphPad Software Inc.