How is CogStat different?

krajcsi edited this page Dec 16, 2017 · 7 revisions

CogStat is different from other statistical packages in many ways. The main aim was to build a very efficient statistical package for cognitive psychologist, but the methods used here could be efficient for many other scientists as well.

Compiling the components of an analysis

While analyzing the data, for a specific task, usually a quite well defined set of analyses are used. For example when comparing two groups, one might want to look at the raw data, look for outliers, check normality, compute some descriptive statistics, calculate the effect size, estimate the population parameters, run hypothesis tests, etc. In most statistical packages these functions can be found behind different menus, and takes quite some time to run all these analyses.

In CogStat, most relevant analyses are compiled, and the menus reflect complete tasks rather than single analysis. For example, in group comparison function after setting the grouping and dependent variables, CogStat automatically presents the data graphically, computes the descriptives, runs the hypothesis tests, etc.

This solution most importantly saves time. You'll notice how easy it is to run an analysis in CogStat. Second, it decreases the chances, that something important aspect of your data is missed. “Blind” analysis is less likely, because you can see your data more completely, without additional efforts.

Automatic hypothesis test selection

There are many useful summaries of hypothesis tests in textbooks and on the Internet that guides students and researchers which test should be used. However, in most cases it is quite straightforward which test should be used in a sense that it could be easily algorithmized, which means that a computer could easily do it. Using CogStat there is no need to look for the appropriate tests, because CogStat knows how to do it, and depending on the question, the variables, and the measurement level of the data, it chooses the appropriate hypothesis test.

Again, one main advantage of this solution is that CogStat is easy to use, and running analyses is much faster than in other packages. Automatic test selection also decreases the possibility that assumptions are not checked. Additionally, some less known test can be also used. Finally, developers and contributors of CogStat can suggest a consensus about the appropriate tests and their use when multiple solutions are proposed.

More efficient output

Most of the statistical packages display many details of an analysis, and many of those details are hardly used by the researchers. These details make using the output more difficult. E.g., have you ever made mistakes by coping the ANOVA results to your manuscript converting it to APA format? Do you spend some time to find the appropriate table that includes the relevant data, while the data of some other tables are never used?

CogStat displays the information that is mostly required by researchers, and CogStat displays these results in a format that is used by researchers, currently in APA format.

Again, this feature saves time, partly because results can be directly copied to your manuscript. It is also easier to apply the APA format more precisely, and there is no need for checking the correct format for a specific hypothesis test. It is also easier to find the appropriate information you need. Additionally, it is harder to mix up the results, and use some incorrect result based on the statistical package output.

Less known or insufficiently used analysis are more accessible

There are many known problems with the current statistical analysis practice. And there are several remedies proposed in the recent decades. in line with those solutions, CogStat, for example,

  • computes effect size and confidence intervals (work in progress, these results are available only for some of the analyses)
  • uses specific hypothesis tests for single case studies
  • is prepared to run EZ diffusion analysis (work in progress, see this function in an upcoming release)

Informative graphs

Graphs should help us to understand the data and understand statistics. Several rarely used methods are used. e.g.:

  • Individual data are displayed on graphs to show if there are outliers, or if there are some unexpected patters, or to show hints about the reliability of the data.
  • Repeated measures data are connected between neighboring variables to stress that the data are related.

Other features

  • CogStat can be used in Jupyter Notebook
  • Localized to several languages
  • Free, so you don't have to pay for it to use
  • Open source, therefore you can look at the code (if you understand code) and also you can modify it
  • Cross platform, so you can use it on Windows, Linux and Mac

Cool? Download and install CogStat.

You have some ideas to make it even cooler? Join the methodological or technical developers to improve CogStat.

Clone this wiki locally
You can’t perform that action at this time.
You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.
Press h to open a hovercard with more details.