Analyse-it

Analyse-it Method Evaluation Edition 3.5 alpha 2 now available

Fri, 10 May 2013 14:45:46 GMT

Today we released the 2^nd alpha of the Analyse-it Method Evaluation Edition 3.5.

Alpha releases are pre-release versions of the software that are still in active development. We release them to a small group of customers so we can get feedback and quickly identify and fix any problems before the public beta release. If you want to take part in the test phase reply or comment on to this post or contact us. You can use pre-release versions of Analyse-it alongside your existing version of Analyse-it, so it won't disrupt your work. And, if you help during in the test phases you will get a discount on the upgrade (a free upgrade for those who contribute the most) when the product is released later this year.

This latest alpha release includes linearity and precision analysis, plus diagnostic test performance (ROC analysis and binary test performance) and reference intervals from the 1^st alpha.

Some of the new features included so far are:

Precision

Graybill-Wang / Ting-Burdick-Graybill-Jeyaratnam-Lu modified large sample confidence interval estimators for variance components – produces confidence closer to the stated level than Satterthwaire confidence intervals, which can be too liberal.
Support for unbalanced 1-way and 2-way nested random ANOVA models.

Linearity

Hsieh-Liu confidence interval estimator for degree of nonlinearity.
Equality and equivalence hypothesis tests for nonlinearity.

Reference interval

Harrel-Davis, Bootstrap, and Robust Biweight quantiles.
Box-Cox and other transformations.

ROC

Compare up to 10 independent or correlated ROC curves.
Equality, equivalence and non-inferiority hypothesis tests.
Estimate false positive fraction (FPF) at fixed sensitivity, sensitivity at fixed FPF, and sensitivity/FPF at fixed cut-off.
Find optimal decision threshold based on costs.

Binary diagnostic tests

Compare sensitivity and specificity of 2 independent or correlated tests.
Diagnostic odds-ratio and Youden’s index.
Improved confidence interval estimators for likelihood ratio and predictive values.

Post comment

Announcing the Analyse-it Method Evaluation Edition 3.5

Fri, 19 Apr 2013 15:39:36 GMT

We are now starting to release test previews of a major update to the Analyse-it Method Evaluation edition. The new release will include many new features (we'll reveal more in the coming weeks) and will support 32- and 64-bit versions of Excel 2007, 2010, and 2013.

During the initial test phases we release development versions of the application to a small group of customers to ensure it installs and runs as expected on a wide range of PCs and configurations. The official beta test phase stage then follows where more customers are invited to download and use the software, while we iron-out the final few bugs before the official release. The official release is planned for summer 2013.

If you want to take part in the test phase, reply to this post or contact us and let us know what aspects of Analyse-it you use:

Analytical Linearity, Precision, Accuracy,
Diagnostic performance (ROC, binary test performance),
Reference ranges,
Agreement (Bland-Altman).

We will then invite you into the testing phase at the appropriate time over the next 4-6 weeks. You will be able to use the test version alongside your current Analyse-it so it will not disrupt your day-to-day work.

As a reward for your time, those who partipicate in the test phases will get a discount on the upgrade (or free upgrade for those who contribute the most) when the product is released later this year.

Post comment

Quantiles, Percentiles: Why so many ways to calculate them?

Mon, 18 Feb 2013 14:39:54 GMT

What is a sample quantile or percentile? Take the 0.25 quantile (also known as the 25^th percentile, or 1^st quartile) -- it defines the value (let’s call it x) for a random variable, such that the probability that a random observation of the variable is less than x is 0.25 (25% chance).

A simple question, with a simple definition? The problem is calculating quantiles. The formulas are simple enough, but a take a quick look on Wikipedia and you’ll see there are at least 9 alternative methods http://en.wikipedia.org/wiki/Quantile. Consequently, statistical packages use different formulas to calculate quantiles. And we're sometimes asked why the quantiles calculated by Analyse-it sometimes don’t agree with Excel, SAS, or R.

How are quantiles and percentiles calculated in Excel, SAS and R?

Excel uses formula R-7 (in the Wikipedia article) to calculate the QUARTILE and PERCENTILE functions. Excel 2010 introduced two new functions that use slightly different formulas, with different denominators: PERCENTILE.INC and PERCENTILE.EXC.

SAS, R and some other packages let you choose which formula is used to calculate the quantiles. While this provides some flexibility, as it lets you reproduce statistics calculated using another package, the options can be confusing. Most non-statisticians don’t know when to use one method over another. When would you use the "Linear interpolation of the empirical distribution function" versus the "Linear interpolation of the modes for the order statistics for the uniform distribution on [0,1]" method?

Why so many ways to calculate quantiles?

Many of the formulas to calculate quantiles were developed when today's computing power wasn’t available. Believe it or not some are nearly 100 years old! Now they’re merely historical curiosities, but some remain in packages like SPSS that have their roots in the 1970s.

During the development of Analyse-it we always ask: what’s the latest or best (in some sense) method we can use to calculate this statistic?

Filling Analyse-it with all the formulas invented would be easy. But that doesn’t help the average user – users that don’t have the knowledge, or can’t invest the time needed to research the most suitable method. Instead we look at published research to find the best method. If there is no single best method we implement a small set of alternatives that provide the best solution in specific situations – situations we can clearly define and explain.

Every statistical test and estimator included in Analyse-it has to pass this test to make the cut. We applied these principles when it came to quantiles.

Hyndman and Fan published a paper on calculating quantiles in 1996. It evaluated the methods used by popular statistics packages to calculate quantiles, with the intention to find a consensus on which all statistics packages could standardise. Of the 9 formulas used, 4 formulas satisfied five of the six properties desirable for a sample quantile and their derivations were deemed justified. Of those 4 formulas, Hyndman and Fan felt the "Linear interpolation of the approximate medians for order statistics" (method R-8 on the Wikipedia page) formula was best due to the approximately median-unbiased estimates of the quantiles, regardless of the distribution. Of the remaining formulas, 2 were also distribution-free but were not unbiased, and the other was approximately unbiased only for the normal distribution. They concluded that formula R-8 should be adopted as the standard across software packages.

That was 1996. Unfortunately little progress has since been made toward standardisation. Many statistical packages have a long history (even Analyse-it is over 15-years old now!) and most tend to stick to the same method to maintain backwards computability with older versions. Even R, a relative newcomer, doesn't use the recommended formula. It uses formula R-7 by default, for compatibility with S (http://stat.ethz.ch/R-manual/R-patched/library/stats/html/quantile.html). Minitab, SPSS and SAS use R-6.

Consistency between the statistics and plots.

To complicate the situation further quantiles and percentiles are also used in statistical plots. The Tukey box-plot, for example, uses the 1^st and 3^rd quartiles (0.25 and 0.75 quantiles) for the extent of the box element of the plot.

Frigge, Hoaglin and Iglewicz published a paper in 1989 that looked at how quantiles were calculated in 3 of the major statistical packages. They found that because each package used a different formula, each identified different observations as outliers. Confusing! They recommended statistical packages use the "Ideal or Machine Forths" formula for consistency, which is equivalent to using method R-8 to calculate quartiles.

What technique does Analyse-it use?

By now you can probably guess that we chose to use R-8 formula in Analyse-it.

Formula R-8 is recommended as the standard in both papers cited above, for descriptive statistics and plots. And using a single formula avoids the confusing situation you’ll sometimes see with other statistics packages, where the 1^st and 3^rd quartiles used for the box-plot differ from sample .25 and .75 quantiles.

Of course using formula R-8 can lead to some differences between the quantiles calculated by Analyse-it and other packages, though often you'll only see it with small sample sizes. If possible, as you can in R, we recommend you change the quantile calculation to use the R-8 formula. If not, you can be sure the statistics calculated by Analyse-it are absolutely correct. And you can cite this article as to why!

Further reading:

Hyndman, R.J.; Fan, Y. (November 1996). "Sample Quantiles in Statistical Packages". The American Statistician 50 (4): 361–365.

Frigge, Michael; Hoaglin, David C.; Iglewicz, Boris (February 1989). "Some Implementations of the Boxplot". The American Statistician 43 (1): 50–54.

2 comments

Analyse-it 3.10 now includes a statistical reference guide

Wed, 16 Jan 2013 15:28:32 GMT

Yesterday we improved the help in the Analyse-it Standard Edition and added a statistical reference guide. The guide tells you about the statistical procedures in Analyse-it, with help on using and understanding the plots and statistics. It’s a work in progress, and we intend to improve it further with your comments and feedback, but it’s important to understand the role of the guide.

Firstly, the guide is not intended to be a statistics textbook. While it covers key concepts in statistical analysis, it is no substitute for learning statistics from a good teacher or textbook.

Secondly, the guide does not include the mathematical formulas behind the statistics. While an understanding of the mathematics is useful, it is better to understand the practical application of statistics: when and where they can be used, and how to interpret the results. Software makes it unnecessary to know the exact formulas, and often the exact mathematics used in software differ from those in textbooks since optimised routines are used to ensure good performance and numerical precision.

Finally, the guide does not explain experimental study design - an area where many mistakes are often made. We recommend you invest in a good textbook on study design to save wasting time on research that is ultimately flawed and cannot be published.

The Statistical Reference guide is included in version 3.10 of the Analyse-it Standard Edition. If you have active maintenance you will be notified the update is available when you next start Analyse-it. When you have installed the update you can access the help from the Help icon on the Analyse-it tab. You can provide feedback on any topic in the help using the Send Feedback button on the toolbar within help.

If you prefer to read the guide on your e-reader we have published it in EPUB format (suitable for Apple iBooks and other e-readers) and MOBI format (suitable for Kindle readers).

Analyse-it Standard Edition 3.0 Improvements make it a Cost-Effective Alternative to Traditional Statistical Software

Wed, 03 Oct 2012 08:15:53 GMT

Leeds, England (PRWEB) October 03, 2012 -- Analyse-it Software, Ltd. today announced a major new release of their popular statistical analysis software, Analyse-it®. With support for Excel 2007, 2010 and the forthcoming Excel 2013, Analyse-it transforms Microsoft Excel into a cost-effective powerful statistical analysis and data visualization package. Statistics and plots are included for exploring and describing data, estimating parameters, testing hypotheses, uncovering relationships and fitting models.

"We’ve made major improvements to model fitting and regression analysis in Analyse-it 3.0", said Simon Huntington, Director of Statistical Products at Analyse-it. "The improvements make Analyse-it a serious rival to larger statistical packages costing up to 5 times the price. Unlike the alternatives though, Analyse-it lets you perform all your statistical analysis without having to leave, or export your data, from Microsoft Excel."

Amongst the hundreds of improvements, model fitting and regression analysis have been improved to support simple models such as linear, logarithmic, exponential and power regression and advanced models such as multiple linear regression and logistic regression. Relationships between variables can be visualized using the scatterplot matrix and partial residual leverage plots show the effect of each term when building the model. Residual plots, distribution plots, lag plots, and sequence plots are also included for checking model assumptions. And an influence plot helps to quickly identify outliers and influential points, based on Studentized residuals and Cook’s D.

Pricing for Analyse-it 3.0 starts at just US$ 249 for a 1-user perpetual licence, or US$ 99 for a 1-user annual subscription. Customers with maintenance can download version 3.0 at http://analyse-it.com/support/download/, or can get the upgrade by extending maintenance from just US$ 50.

For more information and to download a 30-day trial, visit:

http://analyse-it.com/products/standard/

To check if your maintenance includes a free upgrade to version 3.0, and if not to extend maintenance, visit:

http://analyse-it.com/support/download/

1 comment

Analyse-it Standard Edition 3.0 release candidate now available

Thu, 12 Jul 2012 09:35:06 GMT

Update 3rd October 2012

The Analyse-it Standard Edition version 3.00 is now available, see Analyse-it Standard Edition to purchase a licence or download a trial.

Original article

Today we pushed the release candidate of the Analyse-it Standard Edition v3.0 for Microsoft Excel 2007 & 2010, our statistical analysis software for Microsoft Excel.

The release candidate is feature complete and is intended to be the final, almost public release of the software.

The software is now validated against our library of thousands of tests to ensure the statistics and plots are accurate and correct. You can now use the software in your day to day work and, like previous versions, it can be used alongside your current version of Analyse-it until you become more comfortable with it.

The release candidate also includes a user-guide and five tutorials to guide you through using the software by example. We recommend you complete the tutorials first as they help you quickly understand how to use Analyse-it 3.0 and demonstrate many of the types of statistical analysis included. The send feedback feature is available from the help toolbar so you can easily send your suggestions on how we can improve the help.

Finally the performance of the software is much improved and installation is simpler as we’ve provided a combined installer for all supported versions of Excel: Excel 2007 32-bit, Excel 2010 32-bit and Excel 2010 64-bit.

The release candidate is available to everyone. If you already have a pre-release of Analyse-it v3.0 installed you should receive an update notification in the next few days. Alternatively you can download it now at:

http://analyse-it.com/support/test/rc/

In the next blog post we will announce pricing for Analyse-it 3.0, and upgrade pricing and eligibility for current users of Analyse-it. Those of you who helped during the pre-release testing of Analyse-it will also be rewarded with a discount on the upgrade.

Analyse-it Standard Edition 3.0 final beta and technical specification

Thu, 24 May 2012 13:50:53 GMT

Update 3rd October 2012

The Analyse-it Standard Edition version 3.00 is now available, see Analyse-it Standard Edition to purchase a licence or download a trial.

Original article

We’re pleased to release the final beta of the Analyse-it Standard Edition, v3.0. The beta is now publically open to anyone as we iron out any final issues and conduct final testing before release.

To download the beta, please visit:

http://analyse-it.com/support/test/beta/join.aspx

Screenshot: Analyse-it fit model analysis includes an influence plot to identify points with a substantial effect on the fitted model.

Many of you have asked what statistical tests and plots are included in the new release. The full specification is shown below. In coming weeks we'll announce pricing, upgrade pricing, and award free licences to the beta testers who contributed most time and effort.

Descriptive statistics
	Mean, Median, Variance, SD, Skewness, Kurtosis Quantiles / Percentiles Frequency distribution table Contingency table Correlation coefficients – r, rs, tau
Plots
	Histogram Frequency plot Dot plot Box plot Mean error bar plot CDF plot Normal Q-Q plot Scatter plot Scatter plot matrix Leverage plot Residual plots Outlier and influence plot Difference plot Mosaic plot Grouped frequency plot Stacked frequency plot Pie frequency plot
Hypothesis tests
	Location
		Z Student t Welch t Wilcoxon Wilcoxon-Mann-Whitney Sign ANOVA Welch ANOVA Kruskal-Wallis
	Dispersion
		X² Fisher F Bartlett Levene Brown-Forsythe
	Normality
		Shapiro-Wilk Anderson-Darling Kolmogorov-Smirnov
	Association
		Kendall test Spearman test Pearson test
	Proportions
		Binomial exact McNemar exact Fisher exact Score Z Pearson X² Likelihood ratio G²
Estimators
	Location
		Mean Median Hodges-Lehmann pseudo-median t-based confidence interval for mean Z-based confidence interval for mean Thompson-Savur confidence interval for median Tukey confidence interval for Hodges-Lehmann pseudo-median Mean difference t-based confidence interval for mean difference Z-based confidence interval for mean difference Welch-Satterthwaite t-based confidence interval for mean difference Hodges-Lehmann location shift Tukey confidence interval for Hodges-Lehmann location shift Tukey-Kramer confidence interval for mean difference Dunnett confidence interval for mean difference
	Dispersion
		Variance X²-based confidence interval for variance Variance ratio F-based confidence interval for variance ratio
	Proportions
		Proportion Odds Clopper-Pearson exact confidence interval for proportion / odds Wilson score confidence interval for proportion / odds Proportion difference (risk difference) Proportion ratio (risk ratio) Odds ratio Miettinen-Nurminen score confidence interval for proportion difference / proportion ratio / odds-ratio Newcombe score confidence interval for proportion difference Conditional exact confidence interval for odds ratio
	Correlation
		Pearson r Fisher Z confidence interval for Pearson r Spearman rs Kendall tau Samara-Randles confidence interval for Kendall tau
Regression and model fitting
	Linear regression Polynomial 2nd to 6th order Logarithmic regression Exponential regression Power regression Multiple linear regression Binary logistic regression Advanced multiple linear and logistic regression models with simple, crossed, polynomial and factorial terms, with categorical explanatory variables coded as dummy variables

Join the Analyse-it Standard Edition 3.0 beta!

Fri, 02 Mar 2012 14:57:22 GMT

Update 3rd October 2012

The Analyse-it Standard Edition version 3.00 is now available, see Analyse-it Standard Edition to purchase a licence or download a trial.

Original article

Today we released the first public beta of Analyse-it Standard Edition, v3.0.

Maybe we're biased, but it is an amazing product! It's been a heck of a lot of work, but the range of statistical tests and plots in Analyse-it v3 rival what the expensive, established statistical packages provide. In fact, in many cases we’ve surpassed what they offer. And the statistical plots go beyond what's available in any other Excel add-in.

The beta-test programme is open to everyone. Even if you don't currently use Analyse-it you are welcome to join. If you do use Analyse-it the beta will run alongside any existing version, so your day-to-day work won't be interrupted or affected.

To join, drop an e-mail to beta@analyse-it.com and follow us on Google+ to join the beta-testing circle. We will then send you a link with instructions on how to download, install and start using the software. As a reward for your time, all Google+ beta-test followers will get a discount on the upgrade when Analyse-it v3.0 is launched.

In the coming weeks we will reveal more about Analyse-it v3.0, including upgrade and new licence pricing. If you are currently considering whether to buy a licence for Analyse-it v2.0, don’t hesitate – you’ll get the upgrade to version 3.0 free of charge and avoid the price increase.

2 comments

Announcing Analyse-it Standard Edition 3.0

Wed, 16 Nov 2011 14:47:30 GMT

Update 3rd October 2012

The Analyse-it Standard Edition version 3.00 is now available, see Analyse-it Standard Edition to purchase a licence or download a trial.

Original article

If you follow us on Twitter (@Analyse_it) you will have seen that we released a new version of the Analyse-it Standard Edition, v3.0, to testing this week.

Screenshot: Analyse-it v3 scatter matrix with histograms and density ellipses

In coming months we’ll reveal more about the many new features, statistics & plots in Analyse-it Standard Edition 3 (see image above). Pricing and upgrade costs will also be announced, though many, including anyone buying a licence today, will receive the upgrade free.

To join, get involved, and start testing early releases of Analyse-it 3 e-mail support@analyse-it.com. The only commitment is you need to use the application for a few hours over the coming weeks and will need either Excel 2007 or Excel 2010 (on any version of Windows). You can use the test version alongside your current Analyse-it so it will not disrupt your day-to-day work.

If you don’t already, you can subscribe to the blog, or follow us on Twitter (@Analyse_it), to get the latest information. We plan to use Twitter more in future to post small updates and notifications, and use the blog for longer articles.

5 comments

Analyse-it 2.22 adds support for Excel 2010

Thu, 12 Aug 2010 19:41:27 GMT

Update 4th April 2012

The Analyse-it Standard Edition version 3.00 now supports 32-bit and 64-bit versions of Excel 2010. For more information see Analyse-it Standard Edition .

Original article

It's been a few months since we released Analyse-it 2.22, which added compatibility with, what was then, the Excel 2010 release candidate. Now Excel 2010 is available it seems many are upgrading from Excel 2003 and are contacting us to ask whether Analyse-it is compatible. It is!

Interestingly, starting with Office 2010, Microsoft is providing 32- and 64-bit versions of Microsoft Excel. Until now Excel has been a 32-bit application only (going back to Excel 5 which was a 16-bit application). Excel 64-bit natively supports the 64-bit microprocessors becoming more common in desktop PCs and allows you to work with truly enormous quantities of data.

Unfortunately most Excel add-ins, such as Analyse-it, are 32-bit applications and are not compatible with the 64-bit versions. It will take sometime until add-ins like Analyse-it natively support 64-bit – it has taken Microsoft many years to take the first step. For this reason, Microsoft recommends you only choose to install Excel 64-bit if you really need it. And 99.99% of users don’t. Excel 32-bit supports over 1million rows by over 65,000 columns and up to 2GB workbooks – that’s a lot of information!

If you have to choose whether to install Excel 32- or 64-bit, and want to use Analyse-it, ensure you choose to install the 32-bit release. And before you think you'll have your cake and eat it – you can't install both 32-bit and 64-bit versions on the same PC. It's one or the other. See Microsoft’s advice: 32-bit vs 64-bit editions of Office 2010.

How to download and install 2.22

You can download Analyse-it version 2.22 now at:

http://analyse-it.com/support/download.aspx

When you’ve downloaded the update, simply install over your existing version of Analyse-it. There’s no need to uninstall the old version and you won’t need your product key to reactivate Analyse-it -- unless you’re upgrading from Analyse-it version 1.xx in which case you’ll need to request a product key to activate Analyse-it 2.20.

Network administrators can install Analyse-it version 2.22 by repeating the server installation steps as described at http://analyse-it.com/userguide/install_net.aspx

What's changed in this release?

Since many of you work in regulated environments, and need to re-validate changes made in each Analyse-it update, we have described the changes in the change log at:

http://analyse-it.com/support/changes.aspx

Subscribe for news of the latest releases

As we mentioned before, all major and minor updates to Analyse-it on this blog. If you’re not already a subscriber, subscribe to the blog by e-mail or via the RSS feed to be notified of new versions as soon as they’re released.

Microsoft launches Windows 7

Fri, 23 Oct 2009 16:01:19 GMT

Yesterday Microsoft launched Windows 7. It’s the next version of Windows, following on from Windows Vista, Windows XP, and Windows 2000.

As a software vendor we had early access to Windows 7 and have been using it daily for approximately 6-8 months. Our impressions are Windows 7 is very reliable, stable, much faster than Vista, and is an upgrade we wouldn’t hesitate to recommend.

Under the hood, dramatic improvements have been made. Start-up time is reduced so Windows 7 is ready to use from cold-start or hibernation far more quickly than Vista and even XP. In normal use it also feels much faster and more responsive.

Usability is improved significantly, with many slick user interface improvements especially for managing applications on the task bar. Many users will be relieved to know that the learning curve is shallow though – the user-interface isn’t such a departure from Windows Vista that you’ll be less productive for the days after upgrading.

Probably the biggest concern when upgrading is backwards compatibility: will it run the software and work with the hardware you already have. Unlike Vista, the answer seems to be yes.

As you might expect, Analyse-it is fully compatible. Office 2007 runs perfectly, as does Office 2003 so if you’re not ready for the Office 2007 Ribbon user-interface yet you won’t be forced to upgrade Office at the same time.

Generally it seems most applications released in the last few years will run without any problems. If you do have a particularly old Windows XP application that isn’t compatible, Windows 2007 Professional/Ultimate Editions offer Windows XP mode, a very slick feature whereby you can run an application in a virtual, yet fully fledged, Windows XP environment. The application runs in a true Windows XP environment, yet appears like any regular application as part of the Windows 7 desktop, on the start menu, and on the taskbar.

Finally, Windows 7 offers better out-of-the-box support for the myriad of PC hardware available, so in most cases you won’t be forced to search the internet for suitable hardware drivers.

To check the compatibility of any of your existing hardware or software, before upgrading, take a look at the Windows 7 Compatibility Center .

Analyse-it 2.20 released

Fri, 08 May 2009 14:51:59 GMT

Today we released the latest update to Analyse-it, version 2.20, which includes some minor fixes and major improvements. The major improvements affect users of Excel 2007, though we recommend all users download this update for the minor fixes included.

Microsoft Office/Excel 2007 Service Pack 2

Late last week Microsoft started to release Microsoft Office 2007 Service Pack 2, and made it publicly available on Tuesday this week. Service Pack 2 includes many improvements, including some worthy performance improvements for users of Microsoft Outlook.

Unfortunately, the changes made to Excel 2007 caused Analyse-it 2.14 (and earlier) to sometimes crash when you clicked Refresh or Collate on the Report toolbar. Analyse-it 2.20 fixes the problem, working around the bug introduced in Excel 2007 Service Pack 2.

If you want to download Microsoft Office 2007 Service Pack 2, you can download it from Microsoft or install it automatically using Windows Update.

Support for large Excel 2007 worksheets

We've also taken the opportunity in this release to add support for large Excel 2007 worksheets. Before Excel 2007 a worksheet was limited to 65,536 rows and 256 columns. Excel 2007 expands this so you can now have a worksheet as large as 1,048,576 rows and 16,384 columns.

Analyse-it now supports the same, so in Excel 2007 you can analyse very large datasets. Unfortunately, Excel charts are still limited to 32,000 data points though, so some Analyse-it charts such as the dot-plot, and scatter-plot, won’t display for very large datasets.

How to download and install 2.20

You can download Analyse-it version 2.20 now at:

http://analyse-it.com/support/download.aspx

Network administrators can install Analyse-it version 2.20 by repeating the server installation steps as described at http://analyse-it.com/userguide/install_net.aspx

What's changed in this release?

Since many of you work in regulated environments, and need to re-validate changes made in each Analyse-it update, we have described the changes in the change log at:

http://analyse-it.com/support/changes.aspx

Subscribe for news of the latest releases

3 comments

Price change: VAT reduced to 15%

Wed, 03 Dec 2008 13:35:04 GMT

The British government recently announced a 2.5% reduction in VAT (sales tax) on goods purchased from the United Kingdom (see where we are based). UK VAT was previously 17.5%, but from the 1st December 2008 until the end of 2009 it has been reduced to 15%.

Like many businesses, last Monday, we implemented the change.

Customers in the United Kingdom who aren’t VAT exempt, and those of you in Europe who are not VAT registered and must pay VAT, will now pay only 15% VAT.

For customers outside Europe, including those of you in the USA, Canada, and Australia, UK VAT still does not apply and you can continue to purchase licences without paying UK sales tax.

Spot the difference: Numbers stored as text

Thu, 27 Nov 2008 07:53:31 GMT

In clearly titling this blog post, we’ve probably already revealed the answer, but... Can you spot the difference between the two rows of values in the Excel spreadsheet shown below?

Sorry, it’s a trick question, because (visually) there is no difference. The difference is how the values are stored by Microsoft Excel. The value 57 in the cell on second row is actually stored as a text string, not a number.

When does Excel store numbers as text?

When you type a value into a cell, Excel looks at what you’ve typed and decides whether it’s a valid number. If it is, the value is stored as a number, and if not it’s stored as text (a string of characters).

Considering this, how is it possible for Excel to store a value that looks like a number, as text? There are a few ways. Most common is when you copy-paste data from another application, and the application providing the data fools Excel into believing the values should be stored as text. Similarly, if you import data from a database field that contained numbers stored as text, the numbers will be imported as text. Finally, you can force Excel to store a number as text by prefixing it with an apostrophe (‘).

The side-effects of numbers stored as text

So what difference does it make? Venturing into computer science briefly, computers represent and store numbers and text values very differently – numbers are stored in a compact binary representation, and text strings are stored as a string of individual characters. The problem is that mathematical operators and functions can only be applied to values stored as numbers. Text strings, even those that look like numbers, cannot be operated on mathematically.

If you use Excel’s SUM function on the two rows, you can see the problem:

The SUM of the second row doesn’t match the first, because SUM works only on numeric values. The cell containing the text string 57 on the second row is ignored by SUM.

If you weren’t aware of this issue before, you might be surprised. Regardless of how Excel stores a value, if it looks like a number you might expect SUM and other worksheet functions to treat it as such!? Because Excel doesn't, this can lead to very subtle and difficult to spot errors.

Microsoft’s attempt to mitigate the problem

Microsoft recognises this problem and have tried to mitigate it since Excel 2002. Excel 2002, 2003 and 2007 now show a green triangular indicator in the top-left of any cell that contains a number stored as text. If you click the cell to activate it, a small pop-up menu appears so you can convert the cell content to a number. See below:

For more information, see Convert numbers stored as text to numbers.

Analyse-it’s solution to the problem

We identified the problem shortly after releasing Analyse-it in 1997. In all versions of Analyse-it since, when Analyse-it reads your data from the Excel worksheet it treats any numbers stored as text as numbers, so they are included in the analysis.

We feel this is a better approach than Excel’s current solution, but the different approaches can lead to confusion. Recently a customer was validating Analyse-it and was surprised to find the Excel function CORREL gave a different answer to that shown by Analyse-it’s Pearson correlation. You can probably guess why. Some numbers on his worksheet were actually stored as text, and just like Excel’s SUM function, the CORREL function ignored the text values. That meant the correlation coefficient was wrong, and didn't match Analyse-it. Once we converted the numbers stored as text to actual numbers, Excel's CORREL function calculated the correlation coefficient properly. It then matched Analyse-it.

2 comments

Analyse-it speeds Method Validation at National Laboratory

Wed, 19 Nov 2008 11:27:47 GMT

Today we’re delighted to publish the second case study into the use of Analyse-it.

The case study features a national clinical laboratory in the USA that offers more than 2,000 tests and combinations to major commercial and government laboratories. They use Analyse-it to determine analytical performance of automated immunoassays for some of the industry’s leading in-vitro diagnostic device makers -- including Abbott Diagnostics, Bayer Diagnostics, Beckman Coulter and Roche Diagnostics.

Unfortunately we cannot name the end-user, or the organisation she works for, in the case study. Although she was delighted to feature in the case study, at final approval her organisation's committee preferred the names be withheld. Thankfully they have allowed us to use the case study, albeit anonymously.

You can read the case study online now or download the Adobe PDF version.

We would love to feature more customer stories in case studies. If you can get approval to participate – which we realise is very difficult in many industries – and have 20 minutes to spare for a telephone interview, please contact us at support@analyse-it.com.

Normal quantile & probability plots

Thu, 06 Nov 2008 15:26:28 GMT

In a previous post, Testing the assumption of Normality, we explained the tests provided in Analyse-it to determine if a sample has normal distribution. In that post, we mentioned that although hypothesis tests are useful you should not solely rely on them. You should always look at the histogram and, maybe more importantly, the normal plot.

The beauty of the normal plot is that it is designed specifically for judging normality. The plot is very easy to interpret and lets you see where the sample deviates from normality.

Interpreting the normal plot

As an example, let’s look at the distribution of systolic blood pressure, for a random group of healthy patients. Analyse-it creates the histogram (left) and normal plot (right) below:

Looking at the histogram, you can see the sample is approximately normally distributed. The bar heights for 120-122 and 122-124 make the distribution look slightly skewed, so it’s not perfectly clear.

The normal plot is clearer. It shows the observations on the X axis plotted against the expected normal score (Z-score) on the Y axis. It’s not necessary to understand what an expected normal score is, nor how it’s calculated, to interpret the plot. All you need to do is check is that the points roughly follow the red-line. The red-line shows the ideal normal distribution with mean and standard-deviation of the sample. If the points roughly follow the line – as they do in this case – the sample has normal distribution.

And that’s the real beauty of the normal plot compared to the histogram – it's very easy to interpret. Visually, the human eye can better judge the points against a straight-line. And, unlike the histogram, there’s less ambiguity. You don’t have to try judge histogram bar-heights against the normal overlay curve.

Variations on the normal plot

Analyse-it creates what is technically called a Normal Quantile plot. Quantile is just another word for a normal or Z-score and refers to what’s shown on the Y axis (in the case of Analyse-it).

There are actually four variations of the normal plot, or eight since depending on preference the X and Y axes are often swapped:

Normal quantile plot. Observations plotted against expected normal score (Z-score, known as quantiles)
Normal quantile-quantile plot (also known as normal QQ plot). Normal score (Z-score, known as quantiles) of the observations plotted against expected normal score (Z-score, known as quantiles)
Normal probability plot. Observations plotted against expected CDF (cumulative area under the normal curve, known as probability)
Normal probability-probability plot (also known as normal PP plot). CDF (cumulative area under the normal curve, known as probability) of the observations plotted against expected CDF (cumulative area under the normal curve, known as probability)

By their nature, normal plots based on probability fail to emphasise non-normality in extreme observations – in the tails of the distribution – as well as quantile based normal plots. Generally, probability/P-P plots are better to spot non-normality around the mean, and normal quantile/Q-Q plots to spot non-normality in the tails.

Thankfully, whichever of variation of the normal plot you’re faced with, interpretation is the same. If the sample is normal you should see the points roughly follow a straight-line.

In future posts we’ll show cases of skewed and peaked distributions, and explain how you can identify these problems from the histogram and normal plot.

1 comment

Using INDIRECT to refer to cells on Analyse-it analysis reports

Thu, 30 Oct 2008 15:19:34 GMT

A customer contacted us last week to ask how to refer to cells on an Analyse-it report worksheet, from a formula on another worksheet. The customer often used Analyse-it's refresh feature, to repeat the statistical analysis and update the statistics, and direct references to cells on the report were being lost on refresh.

As an example, suppose you have used Analyse-it linear regression to calculate the linear relationship between installation cost and the number of employees required, distance to the site, and the cost of machine being installed. Analyse-it would calculate the effect of each variable on the final cost, technically known as regression coefficients, which you can then use to predict installation costs for jobs in future.

You might setup a worksheet to predict and quote installation costs for future jobs. You could use an Excel formula to reference the coefficients directly from the Analyse-it report, for example:

= Employees * CostAnalysis!C17 + Distance * CostAnalysis!C18 + MachineCost * CostAnalysis!C19

By directly referencing the coefficients calculated by Analyse-it in your formula, you can be sure you’re using the exact values with no chance of error.

If you’ve used this technique before, you already know the problem. When you refresh the Analyse-it report, to repeat the analysis and recalculate the statistics, references to cells on the report worksheet are broken. In the above example, the references to cells on the CostAnalysis worksheet become #REF!, for example:

= Employees * #REF! + Distance * #REF! + MachineCost * #REF!

#REF! simply means the reference is broken and refers to a cell that no longer exists.

The reason the cells no longer exist is because when you click Refresh, Analyse-it repeats the analysis, creates a new report worksheet to present the statistics and charts, then finally deletes the old worksheet and replaces it with the new. The cell references are broken when Analyse-it deletes the old report worksheet.

Although not perfect there is a simple workaround for the problem using the INDIRECT function. Rather than refer directly to a cell, you wrap the cell-reference in INDIRECT(“..”), to indirectly refer to it. For example, the installation cost prediction formula would become:

= Employees * INDIRECT("CostAnalysis!C17") + Distance * INDIRECT("CostAnalysis!C18") + MachineCost * INDIRECT("CostAnalysis!C19")

If you now refresh the analysis, for example because you have collected more data, or found an error in the original data, the formula will still work. Cell references won’t be broken, and will refer to the latest calculated coefficients.

Analyse-it Cuts Project Time in Half at Swiss Lab

Tue, 21 Oct 2008 09:30:22 GMT

Today we’re delighted to publish the first case study into the use of Analyse-it.

Marco Balerna Ph.D., a Clinical Chemist at the EOC (Ente Ospedaliero Cantonale) in Switzerland, used Analyse-it when replacing the clinical chemistry and immunological analysers in EOC’s laboratories.

Since the EOC provides clinical chemistry services to five large hospitals and three small clinics in the region, it was essential the transition to the new analysers went smoothly. Marco used Analyse-it to ensure the analyser’s performance met the manufacturer’s claims, to ensure the reporting of patient results was not affected, and to comply with the regulations of the EOC’s accreditation.

Overall the project involved comparing performance for 110-115 parameters, comprising over 25,600 measurements with control materials and patient samples.

Marco was so impressed with Analyse-it and the time he saved, that he was very enthusiastic when we asked if we could feature his story in a case study. We would like to publically thank Marco for his co-operation in the case study. Grazie Marco! Salute!

You can read the case study online now or download the Adobe PDF version.

If you would like to feature in a future case study, on how you’re using Analyse-it, please contact us at support@analyse-it.com

Analyse-it 2.12 released

Tue, 14 Oct 2008 10:15:19 GMT

Last Friday we released the latest update to Analyse-it, version 2.12 -- a minor update, providing fixes to minor issues recently reported by customers. The update is available free.

If you're using version 2.11, and not experiencing any of the issues fixed (see the change history), then you can skip the the update if you wish. But if you're using an earlier version of Analyse-it, version 2.10 or earlier, we recommend you get the update.

If you’re unsure which version of Analyse-it you’re using, see our FAQ: How to find which version of Analyse-it you’re using.

How to download and install 2.12

Analyse-it automatically checks for updates every 15 days on start-up, and will tell you if an update is available to download. Firewalls can get in the way though, so if you haven’t got a notification yet, or want to download 2.12 right away, you can download at:

http://analyse-it.com/support/download.aspx

When you’ve downloaded the update simply install over your existing version of Analyse-it. There’s no need to uninstall the old version and you won’t need your product key to reactivate Analyse-it -- unless you’re upgrading from Analyse-it version 1.xx in which case you’ll need to request a product key to activate Analyse-it 2.12.

What's changed in this release?

We know many of you work in regulated environments and need to re-validate changes made in each Analyse-it update. To see exactly what’s changed since the version you validated, so you can re-validate just the affected statistical tests, please see the change log at:

http://analyse-it.com/support/changes.aspx

Subscribe for news of the latest releases

As we mentioned last time, we’ll announce all major and minor updates on this blog. If you’re not already a subscriber, subscribe to the blog by e-mail or via the RSS feed to be notified of new versions as soon as they’re released.

Print an Excel chart full-page

Thu, 25 Sep 2008 13:06:28 GMT

Although the charts in Analyse-it are large so they’re easy to read when printed, sometimes you need to print a chart to fill the full page. You can do so easily, without resizing the chart, in just a few steps:

Click anywhere on the chart you want to print. It doesn’t matter what part of the chart you select: the chart, a series, the legend – anywhere will do.
Analyse-it users: Click Print on the Analyse-it toolbar. Analyse-it will show a preview first, and from there you can print the chart.
Excel only users: Choose File > Print from the Excel menu bar to print, or choose File > Preview to preview it first. In Excel 2007, click the Office button then choose Print, or Print > Preview.
Excel will print the chart full page.

Chart size is only limited by the page size your printer supports.