home *** CD-ROM | disk | FTP | other *** search
Text File | 1996-03-28 | 10.4 KB | 282 lines | [ttro/ttxt] |
- ========================================================================
- G•Power for Macintosh, Version 2.1.1 April 1996
- ========================================================================
-
- Contents of this document:
-
- (1) Installation and registration
-
- (2) What G•Power does and how it works
-
- (3) How to get G•Power (i.e., where to check for future updates and versions for
- other platforms)
-
- (4) Disclaimer
-
-
- A more detailed description of the program and its background will be available
- soon in printed form:
-
- Erdfelder, E., Faul, F., & Buchner, A. (1996). GPower: A general power
- analysis program. Behavior Research Methods, Instruments, & Computers,
- 28, 1-11.
-
- This article should appear very soon.
-
- ========================================================================
- (1) Installation and registration
- ========================================================================
-
- There are three versions of G•Power:
-
-
- G•Power PPC
-
- is the version to use on Power Macintoshes. It is amazingly fast.
-
-
- G•Power
-
- should run on any 68K Macintosh using System 7 or higher. It has also been tested
- successfully on some Power Macintosh models on which it runs in emulation mode.
- Sorry, earlier system software versions are no longer supporeted.
-
-
- G•Power FPU
-
- is useful if your 68K Macintosh is equipped with a 68881 or 68882 arithmetic
- coprocessor (G•Power/FPU is several times faster than G•Power but it will not run
- on machines without FPU, both 68K or PowerMacs). Again, you must have System 7
- or higher.
-
- No special installation procedure is necessary to use G•Power. Simply copy the
- program to your harddisk. However, we recommend that you install only the
- version appropriate for your Macintosh. Remove the other versions.
-
- G•Power is free. You may to give the program to friends and colleagues who might
- find it useful.
-
- --------------------------------------------------------------------------------
- VERY IMPORTANT: If you want to include G•Power on a Shareware or Freeware CD-ROM
- you MUST ask the authors for permission. Contact us under the address given
- below.
- --------------------------------------------------------------------------------
-
- Note that there is an MS-DOS version of G•Power, so if your friends and
- colleagues use that platform, they can also work with G•Power if they like (see
- Section (3) of this file for the FTP site to check for updates and the MS-DOS
- version of G•Power).
-
- We encourage you to let us know your complete address (including your e-mail
- address if possible) and the version of G•Power you use. As a registered user,
- you will be informed about future program revisions. Your personal data will be
- used for this purpose only. Please send this information to:
-
- Axel Buchner
- FB I - Psychologie
- Universität Trier
- D-54286 Trier
- Germany
-
- E-mail: buchner@cogpsy.uni-trier.de
-
-
- We also ask you to report any errors or problems encountered with G•Power to the
- address mentioned in Section (4) of this file.
-
- What follows is a brief description of the program (it takes no more than five
- minutes to read it).
-
-
-
- ========================================================================
- (2) What it does and how it works
- ========================================================================
-
- G•Power performs high-precision statistical power analyses for the most common
- statistical tests in behavioral research, that is, t-, F-, and Chi2-tests. The
- power values obtained by “accuracy mode” calculations of G•Power are correct up
- to five significant digits at least. Although available for taking a quick look
- at a problem, we do not encourage “speed mode” calculations because they are less
- accurate.
-
- G•Power computes
-
- (a) power values for given sample sizes, effect sizes and alpha-levels (post hoc
- power analyses);
-
- (b) sample sizes for given effect sizes, alpha-levels, and power values (a priori
- power analyses);
-
- (c) alpha- and beta-values for given sample sizes, effect sizes and
- beta/alpha-ratios (compromise power analyses).
-
- It also displays graphically the relation between any two of the relevant
- variables.
-
- Note that in a priori power analyses, the resulting sample size is
- rounded to the next larger integer. Therefore, the actual power of
- these analyses is usually somewhat larger than the one specified.
- G•Power informs you how large the power of your test actually is given
- the calculated samplesize.
-
- In addition, G•Power calculates for you the critical value of the
- relevant test statistic and the noncentrality parameter of the
- noncentral distribution of your test statistic.
-
-
- ----------------------- The “Analysis window” -----------------------
-
- The “Analysis window” is used to specify the information necessary for a
- particular power analysis, and it displays the results of the last analysis.
-
- Click “Calculate” to perform a single analysis for the Type of Power Analysis,
- the Type of Test, and the parameters you specified.
-
- Click “Calc ES” (where ES is d, r, f, f2, or w, depending on the type of test
- currently selected) to open a dialog in which you can specify the values
- necessary to calculate the effect size for your power analysis from empirical
- data (e.g., from published studies).
-
- Click “Draw graph” to open a dialog in which you can specify the values necessary
- to display one parameter (e.g., the power) as a function of another parameter
- (e.g., the total sample size) over a range of values (e.g., from 20 to 220
- subjects) in the “Graph window”.
-
- If the “Graph window” is already open, you can simply double-click this window to
- open the dialog in which you can specify the graph parameters.
-
-
- ------------------------- The “Graph window” -------------------------
-
- The “Graph window” can be used to display one parameter (e.g., the power) as a
- function of another parameter (e.g., the total sample size) over a range of
- values (e.g., from 20 to 220 subjects).
-
- The header of the graph always specifies the type of test, and the fixed values.
- The note below the graph tells you what calculation mode you used. Default is
- “Accuracy mode” which is very accurate (up to five significant digits at least)
- but the price is that you will usually have to wait a little longer until you see
- a result. “Speed mode” calculations are fast but they may not be very accurate
- (particularly for ‘extreme’ values). If you selected “Speed mode”, do not trust
- the results too much!
-
- You can copy the graph using “Copy graph” from the “Edit” menu and paste it into
- your word processor or drawing program documents for saving or editing the graph
- information. Inside G•Power you cannot edit or save a graph.
-
- The scaling of the axes in the “Graph window” depends on the size of the window,
- so re-sizing the window may have the effect of a more fine-grained scaling.
-
- If the “Protocol (table)” option in the “Graph dialog” is on, a table of the
- values used to plot the graph is written to the Result window. This table can be
- imported into your favorite graphics package, either via the clipboard or by
- saving the table as a G•Power file (see below).
-
-
- ------------------------- The “Result window” -------------------------
-
- If the “Result window” is open, all parameters of all power analyses performed in
- the “Analysis window” are recorded here.
-
- The “Result window” is fully editable (using the commands from the Edit menu).
- You can either copy the values you need to the clipboard, or you can save the
- entire information of the “Result window” to a G•Power file using the Save or
- Save as commands from the File menu.
-
- Double-clicking such a file launches G•Power and opens the file so you can record
- more analyses in it. However, because G•Power files are of type TEXT, they may
- also be opened from inside your favorite word processor or graphics package.
-
-
-
- ========================================================================
- (3) How to get G•Power
- ========================================================================
-
- The most convenient way to get a copy of G•Power is to download the program from
- the public FTP server at the University of Trier, Germany. This is where you
- should check for updates and future releases, and for versions for other
- platforms of G•Power. However, we will also try to make available updates and
- future releases of the Macintosh version on an info-mac archive near you.
-
- There are two ways to access this server:
-
- (I) Via WWW:
-
- http://www.psychologie.uni-trier.de:8000/projects/gpower.html
-
-
- (II) Via FTP:
-
- Host: ftp.uni-trier.de
-
- User ID: anonymous
-
- Password: Your e-mail address.
-
- Directory:
-
- (a) if you want the Macintosh-version of G•Power:
-
- /pub/mac/local
-
- (b) if you want the MS-DOS-version of G•Power:
-
- /pub/pc/msdos
-
- Set the file type to binary (omitting to set the file type properly is a frequent
- error!).
-
- Start downloading (if you have to type in the file names, make sure you type them
- exactly as indicated below, this procedure is case-sensitive):
-
- (a) if you want the MACINTOSH-version of G•Power:
-
- GPower21.sit
-
- (b) if you want the MS-DOS-version of G•Power:
-
- GPower2i.exe (international version) or GPower2d.exe (German version)
-
- 7. Finish the FTP session: quit
-
- 8. Convert the compressed file to a standard Macintosh or DOS file:
-
- (a) Macintosh-version: Open the file GPower2.sit from a version of StuffIt or use
- the Stuffit Expander.
-
- (b) DOS-version (self-extracting archive): GPower2i.exe (international version)
- or GPower2d.exe (German version). This will create a subdirectory GPower in your
- current working directory which will contain all necessary files.
-
-
- ========================================================================
- (4) Address for correspondence
- ========================================================================
-
- Please report bugs and suggestions to improve the program to:
-
- Axel Buchner
- FB I - Psychologie
- Universität Trier
- D-54286 Trier
- Germany
-
- E-mail: buchner@cogpsy.uni-trier.de
-
-
-
- ========================================================================
- (5) Disclaimer
- ========================================================================
-
- We hereby disclaim all warranties relating to this software, whether expressed or
- implied, including without limitation any implied warranties of merchantability
- or fitness for a particular purpose. We will not be liable for any special,
- incidental, consequential, indirect or similar damages due to loss of data or any
- other reason, even if we or agents of ours have been advised of the possibility
- of such damages. In no event shall we be liable for any damages, regardless of
- the form of the claim. The person using the software bears all risk as to the
- quality and performance of the software.
-