STATS PLUS A GENERAL STATISTICS PACKAGE FOR THE APPLE II/IIe/IIc by Stephen Madigan, Ph.D, and Virginia Lawrence, Ph.D. Human Systems Dynamics 9010 Reseda Boulevard, Suite 222 Northridge, CA 91324 (818) 993-8536 Copyright 1982 by Human Systems Dynamics, Northridge, CA. All rights reserved. No part of this program or documentation may be reproduced in any form or by any means without permission in writing from the publisher. Program Warranty The professional staff of Human Systems Dynamics is dedicated to producing accurate and reliable statistical analysis software. Our customer support policies include five major points. Returns. You may return the program package for a full refund within 10 days of purchase if you find it does not suit your needs. Please keep your receipt if you buy from a dealer. If the original diskettes are defective when you receive them, we will replace them free of charge. Program Errors. If you find an error in the program, please notify us in writing or by phone. We will supply corrected versions of the program free of charge. Program Updates. Registered owners (those who return the OWNER REGISTRATION CARD) are entitled to program updates andextensions at low cost. Contact us if you require special program modifications. Charges for such modifications will depend on the nature and extent of required program changes. Replacement Copies. Registered owners may obtain replacement copies of program diskettes which have become damaged for $9.00. System Requirements. The STATS PLUS package requires an Apple II/IIe/IIc with ROM Applesoft, 48K, one or two disk drives and a printer for hard copy output. You can also use STATS PLUS with an Apple III in Apple II Emulation Mode or with the Franklin Ace. Apple II is a registered trademark of Apple Computer, Inc. TABLE OF CONTENTS 1. INTRODUCTION 1.1 Using STATS PLUS: An Overview 1.2 Special Program Features 1.3 System Setup 2. DATA PREP 2.1 The DATA PREP Programs 2.2 Choosing File Types 2.3 Random Access Files 2.4 Sequential Files 2.5 Data Transformations 2.6 Search and Select 3. STATISTICS 3.1 Using STATISTICS 3.2 Data Access 3.3 Using STATISTICS: An Example 3.4 Descriptive Statistics 3.5 Frequency Distribution 3.6 Correlation Matrix 3.7 Simple Linear Regression 3.8 Multiple Regression 3.9 T-Tests 3.10 Analysis of Variance 3.11 Nonparametric Statistics 3.12 Contingency Tables 3.13 Cross-tabulation 4. STATS PLUS EXAMPLE 5. REFERENCES 6. ERROR MESSAGES 7. INDEX 1. INTRODUCTION 1.1. Using STATS PLUS: An Overview When you boot the STATS PLUS program disk, you are presented with the MAIN MENU. The MAIN MENU gives you a choice between SYSTEM SETUP, DATA PREP, AND STATISTICS. Use SYSTEM SETUP to record your printer type, slot number, and number of disk drives. SYSTEM SETUP can be modified later if your system changes. 1. SYSTEM SETUP 2. DATA PREP 3. STATISTICS ENTER AN OPTION (1-3): DATA PREP and STATISTICS are the two main parts of the STATS PLUS program package. DATA PREP does file creation and data management. STATISTICS performs the statistical analysis procedures. Your choice of DATA PREP or STATISTICS depends on what you want to do with your data. A. If you have a small amount of data for statistical analysis, go directly to STATISTICS and the statistics option of your choice. Within each analysis option, you can enter your data by keyboard, edit your data quickly, then do the statistical analysis. B. If you have a great deal of data to be analyzed in a number of different ways, go to DATA PREP. B.1 Enter your data with random access files if: you prefer case x case data entry; you want to do a cross-tabulation; you want to search your data and select out cases; you have missing data. After searching and selecting your random access file data or performing cross-tabulations, you can change your data file type. Use the random access to sequential file option in DATA PREP to change your data to sequential files for use with the statistical analysis programs. B.2 Enter your data with sequential files if you are interested only in performing statistical analyses on numerical data. If you later decide that the capabilities of random access files would be useful, use SQF-RAF to convert your sequential files to random access files. Random access files and sequential files are described in greater detail in section 2.1. Just make your choices from the menus within each program. If you are searching for a particular procedure, look in the index at the back of this documentation or use the card in the front pocket of the binder 1.2 SPECIAL PROGRAM FEATURES The STATS PLUS program package is menu-driven. The programs present a list of actions and procedures from which you make a selection. There are no commands for you to learn. At the same time, several program features are available which are not displayed as menu items, but which are very important for ease of program use. A. The Abort Response. If you enter a double question mark (??) in response to a program prompt, the program branches to the EXIT MENU. This special response is very useful if you decide not to continue with a procedure or if you have made errors in specifying a problem. B. The Catalog Response. You can view the catalog of the data diskette by entering C? whenever the program prompts you to enter a file name. C. Invalid Responses to Prompts. Program prompts usually indicate the kind and range of response required. For example, the PRINTER COPY prompt appears as PRINTER COPY (Y,N). A response other than Y or N is rejected by the program. A "bell" sounds and a repeat of the prompt appears. D. Terminating Data Entry. You can terminate keyboard data entry by entering a single asterisk (*) instead of a data value. E. The Continue Response. Whenever you see a right arrow (=>) at the bottom of your screen, press the right arrow key to continue. 1.3 SYSTEM SETUP In SYSTEM CONFIGURATION you create, on the program diskette, a file which contains information about your system: number of disk drives, location of program and data diskettes, volume number, printer controller slot and printer interface type. After you have established the system configuration with this file, the programs automatically read the file whenever you run the programs. Unless you change your computer setup, you never have to change the system configuration again. The program diskette always has a record of your printer type, number of drives and volume number of the data diskettes. PROCEDURE After you boot the diskette from DRIVE 1 SLOT 6, the program title screen appears, followed by the MAIN MENU. MAIN MENU 1. SYSTEM SETUP 2. RUN DATA PREP 3. RUN STATISTICS ENTER OPTION (1-3): Enter 1 and press RETURN to get to SYSTEM SETUP. SYSTEM CONFIGURATION 1. PROGRAM DISKETTE IN DRIVE 1 2. DATA DISKETTE IN DRIVE 1 3. VOLUME NUMBER 254 4. PRINTER SLOT 1 5. PRINTER INTERFACE PARALLEL A prompt, asking if you want to change the configuration, follows. If this configuration fits your system, reply N and the main menu will reappear. To change configuration values, respond YES to the prompt. Indicate which configuration item number you want to change and then enter the new specification. For example, to establish a configuration for two drives, with the program diskette to be read from DRIVE 1 and a data diskette in DRIVE 2, you should enter the following: ANY CHANGES (Y,N) Y WHICH ITEM (2-5) 2 NEW SPECIFICATION (1,2) 2 The NEW SPECIFICATION prompt displays the available responses for each different configuration. Respond N to ANY CHANGES when you have entered all required changes. After the system configuration file (named SCF) has been updated, the MAIN MENU appears again. NOTE The program diskette specification is always 1 and cannot be changed. ONE-DRIVE SYSTEMS If you have only one disk drive, the system must be configured so that both PROGRAM and DATA diskettes are read from DRIVE 1. The main programs will prompt you to insert program or data diskettes at appropriate points. Note that is highly recommended you use separate data diskettes for your data. Use of the program diskette for data storage results in excessive wear of the disk. TWO-DRIVE SYSTEMS With two drives, the program automatically reads programs from disk DRIVE 1 and read files from, and write files to, the disk in DRIVE 2. VOLUME The VOLUME NUMBER refers to volume number of the data diskette. The value supplied is the standard default value. The program diskette volume number is 254 and cannot be changed. PRINTER If you are using an Apple III, you can specify Slot #7 for your printer although #7 is not included in the list of available responses. 2. DATA PREP DATA PREP is a set of data base programs for data file management in statistical applications. DATA PREP handles both random-access and sequential files, and produces data files that can be accessed directly by all Human Systems Dynamics statistical analysis packages. In addition to the basic procedures of data file creation, review/edit and printing, DATA PREP also includes data file transformation procedures. The transformations include commonly used statistical transformations of data as well as procedures for combining data fields arithmetically. Finally, DATA PREP features a search and select procedure in which random-access files can be searched on up to four fields. The search criteria for each field can be up to five specific values or ranges. While DATA PREP does not qualify as a full scale data base management system for business or mailing lists, you will find it extremely effective in statistical applications. DATA PREP is also very easy to use. 2.1 THE DATA PREP PROGRAMS. When you choose DATA PREP from the MAIN MENU, you get the DATA PREP menu with 5 programs. 1. RANDOM ACCESS FILES 2. SEQUENTIAL FILES 3. TRANSFORMATIONS - RAF 4. TRANSFORMATIONS - SEQ 5. SEARCH AND SELECT 6. RETURN TO MAIN MENU ENTER OPTION (1-6): This section gives you a short description of each DATA PREP program. Read the details in later sections. RANDOM ACCESS FILES Using the RANDOM ACCESS FILE program, you can create random access files, enter data, review data, edit data, print files, merge files, and create sequential files from random access files. SEQUENTIAL FILES Within this program, you can create sequential files, enter data, review data, edit data, print files, delete files, or join files. TRANSFORMATIONS -- R.A. FILES You can use this program to perform arithmetic transformations of numeric data in random access files. You can also rank the data, and find ranks. You can create computed fields by combining fields through addition, subtraction, multiplication, or division. TRANSFORMATIONS -- SEQ. FILES You can use this program to perform all of the previously described transformations on your sequential files. SEARCH & SELECT The SEARCH & SELECT program allows you to search random access files on up to 5 fields. You can also select records for display, printing, or creation of a new file. 2.2 CHOOSING FILE TYPES DATA PREP includes procedures for handling two types of files supported by the Apple: SEQUENTIAL (TEXT) FILES, and RANDOM ACCESS FILES. You can see the general nature of these two types of files in the following example. Suppose you have a set of test scores on each of 50 individuals with 3 scores for each person. PERSON # VERBAL QUANT SPATIAL 1 300 420 400 2 450 620 640 3 600 520 510 * * * 50 420 520 560 A simple random access file can hold all of the data for these 50 people. The 3 scores for each person are stored in one record in the file. Thus, number of records is the same as the number of cases. In the example being discussed, note that there are three measurements or data points for each record. The random access file for these data would have just one file of 50 records with three fields per record. To store the same data in sequential files, you would need three files, one for each test score. As a sequential file, each file has its own name, and each contains the data for one test score for each person measured. You would have one VERBAL file with 50 scores, one QUANT file with 50 scores, and one SPATIAL file with 50 scores. More details on files, entering data into files, and file access for analysis are given in the sections on random access and sequential file operations. Before you read those sections, you should read the next section carefully. The next section outlines some considerations in the selection of one of the two possible file methods. WHICH FILE SHOULD BE USED? In deciding which file type to use, you should consider the inherent differences between files and how you want to use those files. Your initial choice of file type does not commit you to that file type forever. You can always use the RANDOM-ACCESS FILES program to convert random-access files to sequential files and vice versa. There are 6 major considerations in deciding on a file type. A. Ease of Use. It is easier to set up sequential files. Generally, all that is needed is a file name for a set of data that you have entered on the keyboard. The program creates the file and writes the file to the data diskette. In contrast, a random access file must be planned in advance. In addition to naming the file, you have to specify the NUMBER OF FIELDS and the WIDTH of each field. The width of the field is the maximum number of characters that you will want the data field to accommodate. For example, suppose you wanted to make a GRADE BOOK file to hold 3 test scores for some number of students. You also want to be able to sum the test scores for each student and have the sums entered into the file. You also want to find the ranks of the total scores. You would have to set up the file with this in mind by setting up the file with enough fields to hold future data. While you might initially have data only for the first test score when you set up the file, you should plan ahead for the file structure you will need later. The next example shows the file structure you might use, and how you should enter that structure. OPTION 1. SET UP NEW FILE FROM RAF MENU. FILE SPECIFICATIONS FILE NAME: TESTS FIELDS PER RECORD: 5 SPECIFICATIONS FOR FIELD TEST FIELD NAME WIDTH 1. VERB 3 2. QUANT 3 3. SPAT 3 4. SUM 4 5. RANK 2 This example illustrates the importance of field width planning and specification. Field widths of 3 characters defined for fields 1, 2, and 3 would be suitable if the test scores varied only from 0 to 999. The SUM field, however, could easily require 4 characters. Finally, note that a field width of 2 is declared for the RANK field. This would be sufficient for up to 99 students; but it would not do for 100 or more students since the ranks for some students would require 3 characters. An insufficient field width can have very bad results. For example, a rank score of 130 would appear in a width 2 as only 13 because of truncation. Leave enough space in your fields to contain all possible characters. This is particularly true if you will be performing data transformations. B. File Speed. You can perform some operations much faster and easier with a random access file. In particular, retrieving a particular record from a file is much faster with a random access file than it is with a sequential file. C. File Space Requirements. Random-access files may use more storage space on a data diskette than a sequential file. This occurs when you specify a field width larger than necessary for the number of characters in the field. Unless you are dealing with a very large set of data that threatens to use up more than one disk, the increase in space needed by such a random-access file is of no consequence. D. Choice of Operations. Some operations in DATA PREP are available for random-access files, but not sequential files. For example, you can delete a single record from a random-access file with DATA PREP, but you cannot delete a single score from a sequential file with DATA PREP. E. Choice of Statistics. Some of the STATS PLUS statistical analysis procedures require one file type or the other. For example, the CROSS-TAB program in DATA PREP works only with a random-access file. Other Human Systems Dynamics programs such as ANOVA II, require sequential files. F. Type of Data. It is possible to store both numeric and alphabetic data in both random-access files and sequential files. However, all Human Systems Dynamics statistical analysis routines using sequential files accept only numeric data. If you want to include non-numeric data in a file, use a random-access file. Then you can also search records and cross-tabulate records based on non-numeric or numeric data. Random-access files can handle missing data. If you have a number of observations per person, with a few missing observations, use random-access files. To perform a later statistical analysis, select out records with complete data in the fields to be analyzed. DATA PREP lets you have the best of both worlds, because it includes procedures for generating sequential files out of the fields of a random access file. DATA PREP also lets you take a set of sequential files and transform them into fields of a single random-access file. If you decide later that you need more fields, you can add fields to your random-access file. You can also add more records at any time. You can do just about anything you want to yoor data with DAl PREP 2.3 RANDOM-ACCESS FILES Choosing Option 1 of the DATA PREP MENU brings you to the RANDOM-ACCESS FILES program, RAF. RAF consists of a set of procedures for creating files, reviewing, editing, printing and combining files. There are 12 different procedures. To understand random-access files more clearly, think of these files as collections of file cards. Each record or file card has all the data for one person. Think of a field from a record as containing the information on a single line of a file card. Field 1 might contain name; field 2, sex; field 3, income; etc. Record #1 would then contain all of the information you have on person #1. That data in record #1 would be neatly arranged in fields, just as you should have the data for person #1 neatly arranged on separate lines on a file card. From the DATA PREP MENU choose RANDOM-ACCESS FILES to get the RANDOM-ACCESS FILES MENU. RANDOM-ACCESS FILES PROCEDURES 1. SET UP NEW FILE 7. MAKE SUBFILE 2. FILE DESCRIPTION 8. RAF => SQF 3. DATA ENTRY 9. SQF => RAF 4. REVIEW/EDIT 10. PRINT FILES 5. MERGE FILES (V) 11. DELETE FILES \ 6. MERGE FILES (H) 12. CATALOG 13. EXIT ENTER AN OPTION (1-13): NOTES Applesoft files can contain up to 32,000 characters. To calculate the number of records which you can have per file, get the sum of the field widths in a single record of your file. Then divide 32,000 by the sum of your field widths to get the maximum number of records allowed in your file. Apple disks can hold about 100,000 characters, so one disk can hold 3 maximum size random-access files. If you have missing data in some fields, choose a key for missing data. If your real data runs from 1 to 10, you could choose 99 for missing data. Use the key later to select out cases with missing data. OPTION 1. SET UP A NEW FILE You must use Option 1 to set up a new file before entering data. The program presents a series of prompts. These are illustrated for you with the correct responses to set up a file for some test data. FILE SPECIFICATIONS FILE NAME: TESTS FIELDS PER RECORD (1-20): 5 SPECIFICATION FOR FILE TESTS FIELD NAME WIDTH 1 VERB 3 2 QUANT 3 3 SPAT 3 4 SUM 4 5 RANK 2 You first assign a name to the your new random-access file. The name must begin with a letter. The name can contain up to 15 characters. You indicate how many fields you want per record, then supply, for each field, a six character name. This name functions as a field title. The fourth prompt asks you to give the field width which is the maximum number of characters the field can hold. As discussed previously, you must define field widths carefully, based on the kind of data values you will be entering. The program truncates from the right any data that you try to enter into a field that is not wide enough to accommodate the data. After you have defined all of the data fields, the program creates a small description file on your data diskette. This file contains the file description information you have just supplied. The program stores the descriptive file on the data disk with the data file name and the suffix -D. When you create a descriptive file for data, the descriptive file appears in your catalog as FILENAME-D. You never have to refer to the descriptive file, but it must reside on the same data diskette as the random-access data file described. Whenever you tell the program to do anything with your random-access files, it reads the descriptive file first. OPTION 2. FILE DESCRIPTION This option displays or prints a description of the random-access file you name. For each file created, you will find it handy to have a printed copy of the description for reference when you perform various file operations. FILE NAME: TESTS FILE NAME TESTS NUMBER OF RECORDS 0 FIELDS PER RECORD 5 RECORD LENGTH 20 FIELD NAME WIDTH 1 VERB 3 2 QUANT 3 3 SPAT 3 4 SUM 4 5 RANK 2 OPTION 3. DATA ENTRY You use this option to enter new records into a random-access file. Once you have supplied the file name, the program indicates the current record number. Then you receive prompts to enter the data for each field in the new record. Simply enter the data for each field and press RETURN after each entry. If you do not want to enter data for a given field, just press RETURN instead of a data value. When all of the data for a record have been entered, the program prompts data entry for the next record. You can use this option to enter the first data into your file. Later, you can use this option to add more data to your file. As you enter data, the program interrupts occasionally to write your data to disk. When you finish entering data, enter * in response to the prompt for the first field of the next record. Entering * terminates data entry. The program then finishes writing your data to the data diskette. Follow the example to get a clearer picture of the procedure. RANDOM-ACCESS FILE NAME: TESTS DATA FOR RECORD 1 (* TO END) 1 VERB 300 2 QUANT 420 3 SPAT 640 4 SUM 5 RANK EDIT (Y,N): N DATA FOR RECORD 2 (* TO END) 1 VERB 450 2 QUANT 620 3 SPAT 640 4 SUM 5 RANK EDIT (Y,N): N * * DATA FOR RECORD 11 (* TO END) 1 VERB * WRITING DATA TO DISK OPTION 4. REVIEW/EDIT Option 4 is used to review and modify the contents of records. First you name the file to be reviewed, then the program loads the file from the data diskette. Simply keying RETURN causes the program to retrieve the first record. Subsequent RETURNs result in display of consecutive records. If you wish to review a specific record, type its number and then RETURN. For each record called up either way, the contents and field names are displayed. You can also see the list of options at all times. REVIEW/EDIT RANDOM-ACCESS FILE NAME: TESTS WHICH RECORD (1-10):3 (* TO END) RECORD 3 1 VERB 601 2 QUANT 520 3 SPAT 510 4 SUM 5 RANK E EDIT D DELETE C CONTINUE E WHICH FIELD (1-5): 1 NEW VALUE: 600 E EDIT D DELETE C CONTINUE C WHICH RECORD (1 10):* (* TO END) The (C) CONTINUE response leads to a repeat of the RECORD NUMBER prompt. The edit response (E) starts the editing procedure. When you select EDIT, the program prompts you to supply the number of the data field you wish, and the new data value you want placed in that field. When you have entered these responses, you get the EDIT, DELETE and CONTINUE prompts again. After reviewing a record, you can press D to DELETE that record from your file. You can use delete to drop your earliest data, remove records accidentally entered twice, or to remove records with too much missing data. Press * to terminate review/edit. The program incorporates your editing into the files and writes to the disk file when you terminate the review/edit procedure. NOTE You can enter up to 20 fields per record. OPTION 5. MERGE FILES (V) With this option you can combine two existing random-access files into one new file without affecting the two original files. This file combining is done vertically. The vertical merge files option is particularly useful when you have accumulated data on more people in a study. You must have two existing files which are identical in number of fields, field names, and field widths. The number of records may differ. To join files, just name the two files involved, then supply a name for the resulting file. The program automatically creates your new file with a total number of records equal to the sum of the number of records in file 1 and the number in file 2. MERGE FILES (V) ENTER FILE NAMES FIRST FILE: TESTS SECOND FILE: T2 NEW FILE NAME: TESTS2 CREATING FILE DESCRIPTION TESTS2-D This option combines our file TESTS which has 10 records with the second file, T2. T2 contains 10 records. The resulting file, TESTS2, has 20 records, as many records as the two original files combined. Note that this form of file combination can be applied only to two files with the same record length and identical field names, field sequences and field widths. Before the program joins the files, it compares the two descriptive files. If the descriptions contain different information about field names or field widths or numbers of fields, the program rejects the attempt to join them. OPTION 6. MERGE FILES (H) This option allows the horizontal merging of two already existing files. You might use horizontal file merge if you suddenly discovered that you have scores for all students on two more tests. You can start by entering those 2 new tests as the only two fields in a new file. Then you can merge the new file horizontally with your old file. With the horizontal merge, you produce a file containing all of the data for each person, old data and new data. This option requires that the 2 files to be joined have the same number of records. The files do not have to have the same number of fields. To perform a horizontal merge, choose Option 6, then namethe 2 files to be merged. When prompted, supply a file name for the resulting file. MERGE FILES (H) ENTER FILE NAMES FIRST FILE: TESTS2 SECOND FILE: EXTRA NEW FILE NAME: TESTS6FLD CREATING FILE DESCRIPTION TESTS6FLD-D OPTION 7. MAKE SUBFILE With Option 7 you can create a new file out of a subset of the fields of an existing random-access file. For example, you might want to create a subfile out of the TESTS2 file. The subfile might contain only the data on verbal and quantitative scores. The program would create a new file with all of the records of the TESTS2 file, but with only the two specified data fields. CREATE SUBFILE RANDOM-ACCESS FILE NAME: TESTS2 NEW FILE NAME: VQ FIELD NAME WIDTH 1 VERB 3 2 QUANT 3 3 SPAT 3 4 SUM 4 5 RANK 2 NUMBER OF FIELDS IN NEW FILE: 2 ENTER FIELD NUMBERS, 1 PER LINE 1. 1 2. 2 CREATING FILE DESCRIPTION VQ-D CREATING FILE VQ NOTE The program writes the subfile to the same data diskette as the original file. You must be sure that you have space for the subfile on your data disk. A data disk holds 3 maximum size random-access files. If your disk is already close to full, use FID on your System Master disk to copy your original file to an empty data disk. Then create your subfile. OPTION 8. RAF ==> SQF This option allows you to generate sequential files from the fields of a random-access file. This procedure is especially useful for statistical application programs that require sequential files. In general, the program creates a sequential file containing the data values in a given field for all of the records in the random-access file. In the following example, sequential files are created from the VERB and QUANT fields of the TESTS file. You can follow the example to see the method used. When prompted by the program, you specify the number of files to be generated, the random-access field numbers, and the sequential file names that correspond to each field. CREATE SEQUENTIAL FILES RANDOM ACCESS FILE NAME: TESTS FIELD NAME WIDTH 1 VERB 3 2 QUANT 3 3 SPAT 3 4 SUM 4 5 RANK 2 FILES FROM HOW MANY FIELDS (1-5): 2 ENTER FIELD NUMBERS (1-2) 1 2 ENTER NAMES FOR SEQUENTIAL FILES VERB: VERB QUANT: QUANT CREATING SEQUENTIAL FILES VERB => VERB QUANT => QUANT OPTION 9. SQF ==> RAF When you use Option 9, you can take a set of sequential files and convert the set into a single random-access file. Each existing sequential file forms one field of the new random-access file. The procedure requires you to describe the new random-access file, then supply the names of the sequential files to be converted. In the following example, the sequential-file version of the VERB data is formed into a new random-access file named NEW. FILE SPECIFICATIONS FILE NAME: NEW FIELDS PER RECORD (1-20): 1 FIELD NAME WIDTH 1 VERB 3 CREATING FILE DESCRIPTION NEW-D SEQUENTIAL FILE LENGTH (1-500): 10 ENTER SEQUENTIAL FILE NAMES 1. VERB VERB CREATING FIELD 1 NOTES All sequential files being converted to a single random-access file must be the same length. Be sure to describe your new random-access file with field widths wide enough to accept the full data values coming from the sequential files. 1 1 OPTION 10. PRINT FILES You can produce a printer copy of an entire file with this option. You need only name the file to be printed. With an 80-column printer, about 68 characters can be printed horizontally. If this does not accommodate all of the data fields per record, the remainder are printed after the printable fields for all records have been printed. OPTION 11. DELETE FILES This option deletes random-access files from a data diskette. After you name the file to be deleted, the program deletes the data file and its descriptive file. The program does not allow you to delete any of the program files on the program diskette. To end the deletion procedure, type * when prompted for the next file name. Remember that you can type C? to see the catalog on your data disk. DELETE FILES FILE NAME (* TO END): V FILE NAME (* TO END): * NOTES You can use option 11 to delete random-access files only. If you name a sequential file for deletion, you will get a FILE NOT FOUND error because the program looks for both the random-access data file and its descriptive file. If you want to delete sequential files, use the DELETE procedure in the SEQUENTIAL DATA FILES program. OPTION 12. CATALOG Option 12 produces a catalog of the data diskette. OPTION 13. EXIT Option 13 exits to the EXIT MENU. 2.4 SEQUENTIAL FILES PROCEDURES You can use SEQUENTIAL FILES for the creation, review and editing of sequential data files. From the DATA PREP menu, choose Option 2, SEQUENTIAL FILES. SEQUENTIAL FILES PROCEDURES 1. DATA ENTRY - CASE X CASE 2. DATA ENTRY - FILE X FILE 3. EDIT - CASE X CASE 4. EDIT - FILE X FILE 5. JOIN FILES 6. DELETE FILES 7. PRINT FILES 8. CATALOG 9. EXIT ENTER AN OPTION: OPTION 1. DATA ENTRY - CASE X CASE The CASE X CASE method involves entering data into several files simultaneously. Each case corresponds to a subject. B1 B2 B3 S1 5 7 8 S2 3 8 6 S3 2 8 5 S4 3 3 4 S5 7 8 9 Select the CASE X CASE data entry method, option 1, from the FILE CREATION MENU. Then specify how many files you want to create, the file length, and the file names. NUMBER OF FILES TO CREATE (1-12): 3 FILE LENGTH (2-600): 5 FILE NAME 1: B1 FILE NAME 2: B2 FILE NAME 3: B3 The screen clears and presents a new display. CASE #1 B1 B2 B3 The screen cursor then lines up under B1. You start to enter your data. Type 5, then RETURN. The cursor lines up under B2. Press 7 and RETURN, etc. CASE #1 B1 B2 B3 5 7 8 The screen then changes to accept data for case #2. CASE #2 B1 B2 B3 The program interrupts this process after every 50 cases, writes the entered data to the data disk and returns to the prompting until all cases have been entered. With CASE X CASE data entry, you can specify creation of up to 12 files at one time, and up to 600 cases per file. When you have entered as many cases as specified in FILE LENGTH, data entry is finished. Your files are created on the data disk and the FILE CREATION MENU appears. OPTION 2. DATA ENTRY - FILE X FILE This data entry/file creation procedure requires you to assign a name to the file to be created and then enter the data on the keyboard. FILE NAME: AAA ENTER DATA FOR AAA (* TO END) 1. 5 2. 6 3. 8 4. 5 5. * Like the CASE X CASE data entry, FILE X FILE entry accepts 600 data points per file. If you want to create longer files, use the JOIN FILES option to combine files into the long file that you require. See Section 2.4.5. Note that you end data entry by typing an asterisk (*) instead of a number. The file will be created and stored as soon as * is entered. The program then asks if you want to create another file. CREATE ANOTHER FILE (Y,N) N Data may be edited using options 3 and 4 from the FILE CREATION menu. Data cannot be edited during data entry and file creation. OPTIONS 3 AND 4. REVIEWING AND EDITING FILES After you have created your files, you can review and edit them. This can proceed on a file or case basis. To review/edit an entire file, select option 4 from the FILE CREATION menu, EDIT - FILE X FILE. Then name the file you want reviewed. The first 30 data points in the file will be displayed, followed by the EDIT prompt. To review on a case x case basis, select option 3 from the FILE CREATION menu, EDIT - CASE X CASE. State the number of files (up to 12) and the file length. The program asks you which case you want to review from the available cases in the files. REVIEW/EDIT (* TO END) WHICH CASE/SUBJECT: You can respond either by giving a case number or by pressing RETURN. To review one particular case, give a case number. Press RETURN if you want to review the first case in the files. After reviewing and editing the first case, press RETURN again in response to WHICH CASE. You can then review the second case. By simply pressing RETURN in this way, you can review and edit all cases in your files. When you have reviewed all cases, just answer WHICH CASE with an asterisk. This response terminates CASE X CASE review/edit and brings the FILE CREATION menu back to the screen. NOTES The changes or corrections you make do not modify the values on the disk until the review/edit process is finished. Thus, if you make corrections using case x case review and then re-review the data, your changes will not appear on the screen. To review corrections, you must exit from REVIEW/EDIT and then select REVIEW/EDIT again. Up to 12 files and 600 cases can be reviewed in any individual run of REVIEW/EDIT. Up to 100 corrections can be made in any one run of REVIEW/EDIT. Any more than 100 corrections will cause the program to halt and you to lose all your corrections. CASE X CASE review is slower than FILE X FILE review because CASE X CASE review requires a greater amount of disk file activity. OPTION 5. JOIN FILES This option lets you create a new file by joining a series of files. This procedure may be useful if you want to create a file for SINGLE FILE INPUT for analysis of variance by combining a number of files. To use this option, specify the number of files to be joined, the names of those files and the name to be assigned to the resulting file. NUMBER OF FILES: 4 FILE NAME 1: A1B1 FILE NAME 2: A1B2 FILE NAME 3. A2B1 FILE NAME 4: A2B2 FILE NAME FOR RESULT: AB2X2 If each file contained five entries, the resulting file AB2X2 would contain 20 entries. Files are entered into the new file in the order in which they are named. OPTION 6. DELETE FILES You can delete files by choosing option 6 and then supplying the names of the files you want to delete. Signal the end of file deletion with an asterisk (*) in response to the file name prompt. The program first unlocks the file and then deletes it. The program does not allow you to delete any of the program files. OPTION 7. PRINT FILES This option produces a formatted listing of a set of files of your choice. The table of data printed has as many columns as files and as many rows as subjects or cases. Columns are headed with the file names. The number of columns printed across the page depends on the field widths of the numbers being printed. The maximum number of variables which can be printed across a page is ten. If there are more than ten variables, the extras are printed after all cases for the first ten variables have been printed. You can specify common or specific printing format. A field width prompt appears. The program accepts any integer response from 2 to 15. The field width determines the number of places allocated to each number to be printed. The number of spaces needed by a number includes one space for the sign of the number and one space for the decimal point. You are also prompted to give the desired number of places to the right of the decimal point. The number of available places ranges from zero to four. You answer prompts for number of files and file names. You do not have to specify file length. All records in a file will be printed. As soon as the file names are entered, file printout begins. SEQUENTIAL FILES PROCEDURES 1. DATA ENTRY - CASE X CASE 2. DATA ENTRY - FILE X FILE 3. EDIT - CASE X CASE 4. EDIT - FILE X FILE 5. JOIN FILES 6. DELETE FILES 7. PRINT FILES 8. CATALOG 9. EXIT ENTER AN OPTION: 7 SPECIFY FILES NUMBER OF FILES (1-12): 6 FILE NAME 1 A1B1 CHECKING FILE A1B1 FILE NAME 2 A1B2 CHECKING FILE A1B2 FILE NAME 3 A1B3 CHECKING FILE A1B3 FILE NAME 4 A2B1 CHECKING FILE A2B1 FILE NAME 5 A2B2 CHECKING FILE A2B2 FILE NAME 6 A2B3 CHECKING FILE A2B3 SPECIFY FORMAT COMMON FORMAT (Y,N): Y FIELD WIDTH (2-15): 3 DECIMAL PLACES: 0 RECORD A1B1 A1B2 A1B3 A2B1 A2B2 A2B3 1. 3 5 7 5 8 12 2. 2 3 8 5 12 14 3. 2 4 10 3 10 9 4. 0 4 8 7 11 12 OPTION 8. CATALOG Option 8 produces a catalog of the data diskette. OPTION 9. EXIT Option 9 exits to the EXIT MENU. 2.5 DATA TRANSFORMATIONS DATA PREP includes two programs for transforming numerical data in files. One program works on sequential files, and one works on random-access files. When you choose data transformations, you are asked for a file name. Then the transformations menu appears. DATA TRANSFORMATIONS 1. ADDITION 8. ARCSIN 2. SUBTRACTION 9. RANK FIELDS 3. MULTIPLICATION 10. FIND RANKS 4. DIVISION 11. ADD FIELDS 5. POWER 12. SUBTRACT FIELDS 6. EXPONENTIAL 13. MULTIPLY FIELDS 7. NATURAL LOG 14. DIVIDE FIELDS 15. CATALOG 16. EXIT ENTER AN OPTION (1-16): Most of the transformations are self-explanatory in terms of the operations performed and the input required. The 14 available transformation procedures fall into four groups in terms of the required procedures and results. A. Options 1-5 require you to supply a numeric constant, which is then applied to all of the records in a set of files or fields of a file. B. Options 6-8 apply the indicated mathematical transformation to file elements. Exponential raises e to the X. C. Options 9 and 10 deal with ordering data: option 9 re-orders data in terms of descending numerical order; option 10 produces the ranks of data in a given field or file. D. Options 11-14 allow you to add files or fields; subtract one field or file from another; multiply and divide one field or file by another. While the same set of transformation procedures is available for sequential and random-access files, there is one basic difference in how the two transformation programs are used. For sequential files, the program writes the transformed data to a new file. With random-access files, the program writes the transformed data to a field of the same file. Follow a few examples of data transformation to make this difference clear. The examples show you how to specify transformations. These examples make reference to data represented by three sequential files: verb, quant, and spatial. The same data are also in one random access file with three data fields and five records. TESTS is the name of the random-access file. VERB QUANT SPATIAL 200 420 400 450 620 640 600 520 510 * * * * * * * * * 420 510 560 Example 1. Division by a Constant Random-Access Files. Suppose you want to divide by 5 the fields VERB and QUANT in TESTS. Select the division option, Option 4. Then follow the prompts to perform the transformation. DATA TRANSFORMATION RA FILES FILE NAME: TESTS ENTER AN OPTION (1-16): 4 FIELD # NAME 1 VERB 2 QUANT 3 SPAT 4 SUM TRANSFORM HOW MANY FIELDS: 2 ENTER DATA FIELDS AND RESULT FIELDS DATA FIELD RESULT FIELD 1. 1 5 2. 2 4 TRANSFORMING DATA... Specify the number of fields to transform, then give the constant divisor. give the number of each field to be transformed and name the result field for the transformed data. The program carries out the division on every record, then writes the data to the result fields. Sequential Files. When using sequential files, you name the files you want to transform, then enter a new file name for the result. DATA TRANSFORMATION SEQ FILES ENTER AN OPTION (1-16): 4 NUMBER OF FILES: 1 FILE NAME 1: VERB CHECKING FILE VERB FILE NAME FOR RESULT: V/5 Example 2. Subtracting Files Sequential Files. In this example, you can follow the process of arithmetic operations on files and fields. The example creates a sequential file that contains the difference between corresponding elements in files VERB and V/5. You just name the two files in the correct order. The second named file is subtracted from the first named file. Then give a file name for your new result file. DATA TRANSFORMATION SEQ FILES ENTER AN OPTION (1-16): 12 NUMBER OF FILES: 2 FILE NAME 1: VERB CHECKING FILE VERB FILE NAME 2: V/5 CHECKING FILE V/5 FILE NAME FOR RESULT: VERB-V/5 Example 3. Adding Fields Random-Access Files. You can add fields VERB, QUANT and SPAT in the random-access file TESTS2. DATA TRANSFORMATION RA FILES FILE NAME: TESTS2 ENTER AN OPTION (1-16): 11 FIELD # NAME 1 VERB 2 QUANT 3 SPAT 4 SUM 5 RANK ADD HOW MANY FIELDS (1-5): 3 ENTER FIELD NUMBERS 1. 1 2. 2 3. 3 RESULT FIELD: 4 TRANSFORMING DATA TRANSFORMING RANDOM-ACCESS FILES. The transformation process is relatively simple for sequential files. The transformation -procedure requires a little more thought for random-access files. You should keep three points in mind. A. The fields for results of a transformation must already exist in the random-access file. You must define a field for later results when you first set up the random-access file. The TRANSFORMATIONS program does not create new fields in the file. It just creates new values for fields. B. Your planning for results fields must include FIELD WIDTH considerations. While fields VERB, QUANT and SPAT in the example are two characters wide, the result field must be wide enough to accommodate the sum. Otherwise, the decimal point and digits to the right of the decimal are lost. As usual, the program truncates data values to fit the pre-defined field width. C. You do not have to write data to an unused or blank field. You can specify any field as the result field. If you choose a field containing data, the program overwrites the original data with your new results. NOTES MULTIPLICATION & DIVISION BY A CONSTANT. The constant must be non-zero. POWER. The constant you enter is the exponent. For example, your constant is 2 to obtain squares. Fractional exponents are also permitted. Use .5 to obtain a square root transformation. The program can also accept negative exponents. Entry of -1 produces reciprocals. ARCSIN. The program uses the derived function described in the Apple Basic Programming Manual, p. 103. RANKING. Option 9 returns the original data values ranked in descending numerical order. Option 10 returns the RANKS of the original data without averaging of tied ranks. For example, if a file contained the values 5,7,3,8,3, Option 9 returns 8,7,5,3,3, and Option 10 returns 3,2,4,1,5. ADDING FILES AND FIELDS. With sequential file transformation, you can only add two files at a time. With transformaton of a random-access file, you can add up to ten fields. All files involved must be the same length. SUBTRACTION OR DIVISION OF FILES OR FIELDS. With these procedures, the order in which you name files is important. For subtraction, the values in the second named file are subtracted from the values in the first named file. With division, the values in the first named file are divided by the data values of the second named file. TRANSFORMATION ERRORS. It is possible that a file or field may contain data values that cannot be operated on validly by the chosen transformation. The program handles these problems differently depending on file type. A. With sequential files, the program reports the problem and identifies the file producing the problem. The program stops trying to transform the file, and erases any values computed and written to that point in that file. The program then goes on to transform the next file named. B. With random access files the program writes a series of asterisks (***...) into the result or destnation field of the record involved, then continues with the transformation. The program presents a report of the occurrence of errors at the end of the transformation process. 2.6 SEARCH & SELECT With the SEARCH & SELECT program (S&S), you can select records from a random-access file. You base your selection on file contents and you specify the selection criteria. The program identifies records that meet the selection criteria. It can then take one of four actions with the records found. 2.6.1 THE RESULTS OF SEARCH & SELECT A. CREATE A NEW FILE The program creates a new file containing only the records found by the search procedure. ALL data fields in the selected records are written to the new file. This file carries the name that you supply. Remember that S&S selects whole records and not fields within records. Use the SUBFILE option of the RAF program to select out fields. B. DISPLAY RECORDS This option displays the contents of records on the TV monitor. This option displays each record as it is found rather than scanning the whole file before taking an action. As it finds each record, the program displays its contents. The search continues when you press any key. C. PRINT RECORDS With this option, you can obtain a printer copy of your selected records. The format is the same as the PRINT FILES option of RAF, except that each record is numbered with its record number in the original file. D. DELETE RECORDS Using this option, you can delete any records from your original file if the records meet your selection criteria. The resulting file has the same name and structure as the original, but fewer records. NOTE The SEARCH AND SELECT program does not produce any kind of reports or summaries of file data. Use the CROSS-TABULATION program if you want to select records and obtain statistical reports on record contents. 2.6.2 SEARCH METHOD AND SEARCH SPECIFICATIONS The most important aspect of using S&S is the definition of the search procedure. This involves specifying the fields to be examined (up to 4 at a time), and the data values in each data field. You can select out a data field in two ways. You can match to an exact value, or fit to a range of values. Either way, you can review and edit your specifications. The entire procedure becomes much clearer if you follow the examples. EXAMPLE 1. PRINT THE RECORDS OF ALL CASES WITH A SCORE OF 600 IN FIELD #1. Use the TESTS2 random-access file from previous examples in the DATA PREP section. This search requires the program to select records based on one field, #1. The program looks for records containing 600 in field #1. SEARCH & SELECT 1. CREATE A FILE 2. DISPLAY RECORDS 3. PRINT RECORDS 4. DELETE RECORDS 5. CATALOG 6. EXIT ENTER AN OPTION: 3 FILE NAME: TESTS2 FIELD # NAME 1 VERB 2 QUANT 3 SPAT 4 SUM 5 RANK SELECT ON HOW MANY FIELDS? (1-4): 1 FIELD #: 1 VERB SPECIFIC VALUES (S) OR RANGE (R): S HOW MANY VALUES (1-5): 1 ENTER VALUES 1. 600 SPECIFICATIONS OK (Y,N): Y SCANNING FILE TESTS2 RECORD 3 VERB 600 QUANT 520 SPAT 510 SUM 1630 RANK EXAMPLE 2. DELETE RECORDS WITH A SCORE OF 300 IN VERBAL AND A SCORE FROM 500 TO 800 IN QUANT. This requires a search on two fields, VERB and QUANT. You search the VERB field for one specific value and the QUANT field for a range of values. First select Option 4. FILE NAME: TESTS2 FIELD # NAME 1 VERB 2 QUANT 3 SPAT 4 SUM 5 RANK SELECT ON HOW MANY FIELDS? (1-4): 2 FIELD #: 1 VERB SPECIFIC VALUES (S) OR RANGE (R):S HOW MANY VALUES (1-5): 1 ENTER VALUES 1. 300 FIELD #: 2 QUANT SPECIFIC VALUES (S) OR RANGE (R): R FROM: 500 TO: 800 SPECIFICATIONS OK (Y,N):Y RECORD # 38 VERB 300 QUANT 520 SPAT 580 SUM 1400 RANK NOTE When you base your selection on a range of values, you must specify a minimum and a maximum value. A record passes the selection test on a field if its data value is numerically equal to or greater than the minimum, and less than or equal to the maximum. EXAMPLE 3. CREATE A NEW FILE WITH 0-400 IN VERB, 0-400 IN QUANT AND 0-400 IN SPAT. This requires a search based on three data fields per record. This example shows the use of an AND selection based on a range of values. SEARCH AND SELECT 1. CREATE A FILE 2. DISPLAY RECORDS 3. PRINT RECORDS 4. DELETE RECORDS 5. CATALOG 6. EXIT ENTER AN OPTION: 1 FILE NAME: TESTS2 FIELD # NAME 1 VERB 2 QUANT 3 SPAT 4 SUM 5 RANK SELECT ON HOW MANY FIELDS? (1-4): 3 FIELD #: 1 SPECIFIC VALUES (S) OR RANGE (R): R FROM: 0 TO 400 FIELD #: 2 SPECIFIC VALUES (S) OR RANGE (R): R FROM: 0 TO 400 FIELD #: 3 SPECIFIC VALUES (S) OR RANGE (R): R FROM: 0 TO 400 SPECIFICATIONS OK (Y,N): Y 7 RECORDS FOUND NEW FILE NAME: TESTS400 CREATING FILE DESCRIPTION TESTS400-D CREATING FILE TESTS400 The entire search and select procedure in this example can be summarized easily. The program selects a record if the first field to be examined contains values in the range 0 to 400, and the second field contains values in the range 0 to 400, and the third field contains values in the range 0 to 400. You can conduct a search on as many as four data fields simultaneously. With the specific values procedure, up to five different values can be searched per field. You can also mix the specific values and range of values procedures in one search. 3. STATISTICS The STATISTICS section of STATS PLUS performs a large number of statistical analyses. Read Sections 3.1 and 3.2 for a general overview of the operation of STATISTICS. Then follow the example in Section 3.3 to make everything clearer. When you are ready to perform a specific analysis, choose that analysis by number from the STATISTICS MENU. Then follow the documentation instructions for that analysis along with the screen prompts. STATISTICS 1. DESCRIPTIVE STATISTICS 2. FREQUENCY DISTRIBUTION 3. CORRELATION 4. SIMPLE LINEAR REGRESSION 5. MULTIPLE REGRESSION 6. T-TESTS 7. ANALYSIS OF VARIANCE 8. NONPARAMETRIC TESTS 9. CONTINGENCY TABLES 10.CROSS-TABULATION 11.RETURN TO MAIN MENU ENTER AN OPTION (1-11): 3.1 Using STATISTICS Doing a statistical analysis with STATISTICS involves two major steps, analysis selection and problem definition. GENERAL PROCEDURE A. Analysis Selection. Select the STATISTICS option from the MAIN MENU of STATS PLUS. B. Problem Definition. Most of the STATISTICS procedures request some preliminary information. The prompts you receive are very explicit, just type in the answer. Descriptive Title. This can be any 40 characters which will effectively title the output for you. Number of Data Samples. This determines the number of data samples which you want to process. Sample Names. Sample names are used only to label your results. These are not the names of the files to be read. Rather, you can label your samples with names different from their file names. The sample names can be up to six characters long. Data Access. You specify your data as either keyboard, sequential or random-access. Printer Copy. You are asked to respond YES if you want printer output of your results. NOTE If your random-access file contains cases with missing data, you must make a subfile of the cases with no missing data before analyzing with STATISTICS. Use SEARCH & SELECT within DATA PREP to select cases containing data within a specified range. If your missing cases are 0 or blank, select on a range from .5 to your data maximum. If your real data includes 0, then your missing data should be coded with something like 999. Select out the appropriate ranges on the variables you plan to analyze, create the subfile, and use that subfile in STATISTICS. 3.2 DATA ACCESS In most cases, you can choose data entry from keyboard, from random-access files, or from sequential files. Keyboard data entry allows you to use the programs in a calculator mode. You can review and edit your data, but you do not need to create any data files. Keyboard entry is particularly handy when you have a small data set and do not need data files for later analyses. In the following table, you can see a summary of the data access methods available for the statistical analysis routines of STATISTICS. OPTION KBD SEQ RAF Desc. Stats. Y Y Y Freq. Dist. Y Y Y Correlation Y Y Y Lin. Regress. Y Y Y Mult. Regr. Y Y Y t-Tests Y Y Y Anal. of Var. Y Y N Non-par. Y Y Y Conting. Tables Y N N Cross-tab. N N Y The type of data required by each analysis option is very obvious. Just answer each prompt as it appears. Go through the example in section 3.3 for a close look at the entire procedure. 3.3 STATISTICS: AN EXAMPLE 1. DESCRIPTIVE STATISTICS 2. FREQUENCY DISTRIBUTION 3. CORRELATION 4. SIMPLE LINEAR REGRESSION 5. MULTIPLE REGRESSION 6. T-TESTS 7. ANALYSIS OF VARIANCE 8. NONPARAMETRIC TESTS 9. CONTINGENCY TABLES 10.CROSS-TABULATION 11.RETURN TO MAIN MENU ENTER AN OPTION (1-11): 6 T-TESTS 1. SINGLE SAMPLE 2. INDEPENDENT SAMPLES 3. RELATED SAMPLES 4. EXIT ENTER AN OPTION (1-4): 2 ENTER SAMPLE/VARIABLE NAMES 1. WT1 2. WT2 HYPOTHESIZED DIFFERENCE: 0 PRINTER COPY (Y,N): N DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILES ENTER AN OPTION (1-3): 1 DATA FOR WT1 DATA FOR WT2 * TO END 1. 45 1. 55 2. 56 2. 52 3. 51 3. 47 4. 57 4. 49 5. 43 5. 53 6. 41 6. 57 7. 44 7. 56 8. 48 8. 58 9. 49 9. 59 10.50 10.52 11.* 11.* EDIT (Y,N):N EDIT (Y,N): N SAMPLE N MEAN S.D. WT1 10 48.400 5.337 WT2 10 53.799 3.910 HYPOTHESIZED DIFFERENCE: 0 OBTAINED DIFFERENCE: -5.400 T(18)=-2.580 P=.0180 STANDARD ERROR =2.092 POINT-BISERIAL CORRELATION=.520 == > After viewing the results of the independent samples t-test, push the right arrow key. The screen then shows the EXIT MENU. 1. RERUN PROGRAM 2. EXIT TO MAIN MENU 3. EXIT TO DATA PREP 4. EXIT TO STATISTICS 5. END ENTER AN OPTION (1-5):4 3.4 DESCRIPTIVE STATISTICS This option computes basic descriptive statistics on up to 8 data variables at a time. The program reports sample size, sum, mean, standard deviation, minimum and maximum values. Each sample may contain up to 600 data points. The 8 data samples analyzed at one time need not contain equal numbers of data points. PROCEDURE. A. Choose DESCRIPTIVE STATISTICS. Enter a title and the number of variables to be processed. B. Assign names to each variable. C. Select a data access method. If keyboard, enter data by keyboard. Otherwise, supply file names. D. Obtain your output automatically. If you have only monitor output, use the CONTINUE response for additional display. For printer output, you receive your results printed in a table. Each row shows all results for each variable. NOTE The standard deviations are unbiased population estimates. DESCRIPTIVE STATISTICS DESCRIPTIVE TITLE: DEMONSTRATION REPORT ON HOW MANY SAMPLES (1-10): 2 PRINTER COPY (Y,N): Y ENTER SAMPLE/VARIABLE NAMES 1. MALE 2. FEMALE DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILE ENTER AN OPTION (1-3): 1 DATA FOR MALE DATA FOR FEMALE * TO END 1. 2 1. 8 2. 4 2. 9 3. 5 3. 7 4. 3 4. 10 5. 6 5. 13 6. 7 6. 9 7. 8 7. 8 8. 4 8. 11 9. 3 9. 14 10.1 10.10 11 .* 11 .* EDIT (Y,N): N EDIT (Y,N): N DEMONSTRATION SAMPLE N SUM MEAN S.D. MIN MAX MALE 10 43.00 4.30 2.214 1.00 8.00 FEMALE 10 99.00 9.90 2.234 7.00 14.00 3.5 FREQUENCY DISTRIBUTION ~ This option produces a frequency distribution for a single data sample. Output can include basic descriptive statistics, frequencies or proportions, and a full percentiles report. You can choose your lower interval limit and your interval sizes. PROCEDURE. A. Enter a title, variable name, and your data access method. B. After you enter your data, the program reports sample size, and the minimum and maximum values in the sample. C. The program asks you to supply a starting value. This value is used by the program to define the lower limit of your distribution. You can use the reported minimum or a lower value. This choice of lower limit allows you to compare distributions on a common scale even though they do not have equal minimum values. D. The program calculates and gives you the minimum interval size. Since the plotting routine can plot up to 15 intervals, minimum interval size is chosen to limit the number of intervals to 15 or fewer. E. Enter your choice of interval size. F. Indicate your choice of frequencies or proportions. G. The program reports the distribution, including cumulative frequencies or proportions. Descriptivestatistics and percentile points appear next. H. You can choose graphics next to produce a bargraph or a frequency polygon with your own labels. Any keyboard character can be used for the labels. I. You can press any key to remove the graph from the screen. Then specify whether you want to create a binary file of your graph. If you create a binary file of your graph, you can dump that graph to a graphics printer later. J. To print your graph, you must first get out of STATS PLUS. Put any system disk in your drive #1, then press RESET twice. The system disk boots, and you are out of STATS PLUS. Use your own dump routine or use GRAFPAK from SmartWare. If you want to use THE GRAPPLER, run the following program to load your graph back into memory. Then follow the instructions that come with THE GRAPPLER. 5 HOME: D$="" 12 HGR2: TEXT: N$="" 15 INPUT "ENTER FILENAME ";N$ 20 PRINT D$;"BLOAD"N$",A$2000" 30 POKE - 16300,0: POKE - 16304,0 Please note in line #5 only, when you enter D$="", you actually enter D$=control key D". Just press the control key at the same time as the D. You will see only D$="" on the screen. Run the program, view the graph and follow instructions for the grappler. NOTE This program works correctly only with whole number data. If you have decimal data, transform the data first. GRAFPAK is a copyright of SmartWare. THE GRAPPLER is a trademark of Orange Micro. FREQUENCY DISTRIBUTION DESCRIPTIVE TITLE: DEMONSTRATION ENTER SAMPLE/VARIABLE NAME 1. TEST PRINTER COPY (Y,N): Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILE ENTER AN OPTION (1-3): 1 DATA FOR TEST * TO END 1. 12 11. 10 21. 7 2. 13 12. 22 22. 7 3. 4 13. 25 23. 17 4. 15 14. 23 24. 12 5. 16 15. 21 25. 29 6. 21 16. 20 26. 15 7. 23 17. 17 27. 14 8. 17 18. 13 28. 18 9. 19 19. 12 29. 17 10.11 20. 9 30. 8 EDIT (Y,N): Y DATA ENTRY: 2 NEW VALUE: 14 EDIT (Y,N): N 31. * SAMPLE: TEST N=30 MINIMUM: 4 MAXIMUM: 29 STARTING VALUE: 1 MINIMUM INTERVAL SIZE: 2 DESIRED INTERVAL SIZE: 5 - FULL PERCENTILES REPORT (Y,N): N SAMPLE: TEST SCORE F CF P CP 1-5 1 1 .03 .03 6-10 4 5 .17 .20 11-15 9 15 .30 .50 16-20 8 23 .27 .77 21-25 6 29 .20 .97 26-30 1 30 .03 1.00 N=30 MEAN=15.60 S.D.=5.90 PERCENTILE RANK SCORE 10TH 7.50 25TH 11.33 50TH 15.50 75TH 20.19 90TH 23.83 GRAPHICS DISPLAY (Y,N): Y BARGRAPH OR POLYGON (B,P): B FREQUENCIES OR PROPORTIONS (F,P): F TITLE FOR GRAPH: DEMO LABEL HORIZONTAL AXIS: TEST [Graphic, pg. 47] CREATE PICTURE FILE (Y,N): Y FILE NAME: FREQ.PIC 3.6 CORRELATION MATRIX This option computes up to 8 Pearson correlations at one time. If you use keyboard data entry, you can only have up to 100 cases per sample. However, if you use data files, you can use as many cases as your file holds. PROCEDURE A. Enter a descriptive title, number of data samples or variables, and the sample names. B. Select a data access method. If you select keyboard, enter your sample size when prompted. C. Enter data by file name or by keyboard. Keyboard data entry is case x case. You enter the case 1 data for all samples, then continue through the rest of the cases. D. Obtain results. The results include sample names, sizes, means, standard deviations, and an upper triangular correlation matrix. The columns of the array correspond to samples 2-k. The rows correspond to samples 1-(k-1). The program truncates sample names to 3 characters for the display. NOTES The program shows the correlation coefficients to 3 digits, without the decimal point. Correlations with an absolute value less than .001 appear as 000. Correlations of +1 or -1 appear as 999 or -999, respectively. For correlations, all samples must be the same length. If your data files are of varying lengths, the program stops reading data when it first strikes an end of file indication. CORRELATION DESCRIPTIVE TITLE: CORRELATION DEMO NUMBER OF DATA SAMPLES (2-8): 5 ENTER SAMPLE/VARIABLE NAMES 1. AAA 2. BBB 3. CCC 4. DDD 5. EEE PRINTER COPY (Y,N): Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILES ENTER AN OPTION: 1 SAMPLE SIZE (3-100): 8 CASE #1 1. AAA 12 2. BBB 8 3. CCC 14 4. DDD 22 5. EEE 13 EDIT (Y,N):N ETC. CORRELATION DEMO SAMPLE N MEAN S.D. AAA 8 9.87 4.15 BBB 8 8.37 2.87 CCC 8 17.50 8.70 DDD 8 14.87 3.79 EEE 8 9.37 3.50 BBB CCC DDD EEE AAA -047 -371 -617 -135 BBB 210 771 489 CCC -476 -065 DDD -454 3.7 SIMPLE LINEAR REGRESSION This program performs a least-squares regression analysis for a single predictor (independent) variable and a single criterion (dependent) variable. The analysis reports: descriptive statistics on each variable; regression coefficients, coefficients of correlation and determination;standard error of estimate; and an optional test on the significance of the slope coefficient. A high resolution scatterplot is also available. PROCEDURE A. Enter title and variable names. B. Select a data access method. Note that the program requires you to name the predictor variable first and then the criterion variable. C. After the basic analysis, you can have a significance test performed on the estimated slope value. Supply a hypothesized slope value, then the program reports a t-statistic and the two-tailed probability value. The usual hypothesized slope value is 0. D. You have the option to have observed, predicted and residual scores reported, either on the monitor or through the printer. E. You can choose graphics to produce a scattergram of your data. If you choose the scatterplot option, the program first writes a temporary file of your data to the data diskette. It then loads the scatterplot program, reads the temporary data file and displays the scattergram. You can supply any labels for the horizontal and vertical axes. In addition, you can cnoose to nave the regression line plotted. F. You can press any key to remove the graph from the screen. Then specify whether you want to create a binary file of your graph. If you create a binary file, you can dump that graph to a graphics printer later. G. To print your graph, you must first: exit from STATS PLUS. Put any system disk in your drive #1 then press RESET twice. The system disk boots and you are out of STATS PLUS. Use your own dump routine or use GRAFPAK from SmartWare. If you want to use THE GRAPPLER, you can follow the directions on page 45. NOTE If you are using a one-drive system, be sure to follow the screen prompts carefully in the scatterplot routine. GRAFPAK is a copyright of SmartWare. THE GRAPPLER is a trademark of Orange Micro. SIMPLE LINEAR REGRESSION DESCRIPTIVE TITLE: KERLINGER ENTER SAMPLE/VARIABLE NAMES 1. X 2. Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILES ENTER AN OPTION (1-3): 1 DATA FOR X DATA FOR Y * TO END 1.2 1. 1 2.4 2. 2 3.3 3. 3 4.5 4. 4 5.6 5. 5 6.* 6. * EDIT (Y,N):N PRINTER COPY (Y,N):Y KERLINGER SAMPLE N MEAN S.D. X 5 4.00 1.581 Y 5 3.00 1.581 Y=.900X-.600 R=.900 R SQUARED = .810 STANDARD ERROR =.796 TEST ON SLOPE (Y,N):Y HYPOTHESIZED SLOPE VALUE: 0 T(1)=3.576 P=.0357 PREDICTED SCORES AND RESIDUALS (Y,N):Y PRINTER COPY (Y,N):Y OBSERVED PREDICTED RESIDUAL 1. 1.000 1.200 -.200 2. 2.000 3.000 -1.000 3. 3.000 2.100 .900 4. 4.000 3.900 .100 5. 5.000 4.800 .200 SCATTERPLOT (Y,N):Y SCATTERPLOT LABEL, HORIZONTAL AXIS: X VARIABLE LABEL, VERTICAL AXIS: Y VARIABLE [Graph 2, pg. 53] CREATE PICTURE FILE (Y,N): Y FILE NAME: LINREG 3.8 MULTIPLE REGRESSION This program performs a full multiple regression analysis for two predictor variables and one criterion or dependent variable. Program output includes descriptive statistics on each variable; correlations and 1st order partial correlations; analysis of variance table; regression coefficients and significance tests. PROCEDURE A. Assign names to predictors and criterion variable. B. Select data access method and enter data. Be sure to assign data to predictors and criterion as prompted. The order in which you enter predictor variables dictates their order in the regression. C. The program displays descriptive statistics and correlation, the analysis of variance table, regression coefficients and their t-tests. D. You can request predicted and residual scores. MULTIPLE REGRESSION DESCRIPTIVE TITLE: TWO FACTOR ENTER SAMPLE/VARIABLE NAMES 1ST PREDICTOR: VERB 2ND PREDICTOR: QUANT CRITERION: SPAT PRINTER COPY (Y,N): Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILES ENTER AN OPTION (1-3): 3 DATA ENTRY: RANDOM-ACCESS FILE FILE NAME: TESTS FIELD # 1. VERB 2. QUANT 3. SPAT 4. SUM 5. RANK ENTER FIELD NUMBERS 1ST PREDICTOR: 1 2ND PREDICTOR: 2 CRITERION: 3 SAMPLE MEAN S.D. VERB 519.00 154.59 QUANT 581.00 128.87 SPAT 563.00 135.07 CORRELATIONS QUANT SPAT VERB .876 .732 QUANT .779 PARTIAL CORRELATIONS QUANT SPAT VERB .567 .129 QUANT .345 MULTIPLE CORRELATION=.786 SOURCE SS DF MS REGRESSION 101562.64 2 50781.32 RESIDUAL 62647.35 7 8949.62 TOTAL 164209.99 9 R SQUARED =.618 F(2,7)=5.67 P=.03 COEFF S.E. T P INTERCEPT 105.28 VERBAL .185 .424 .43 .654 QUANT .621 .508 1.22 .261 PREDICTED AND RESIDUAL SCORES (Y,N): N 3.9 T-TESTS With this option, you can perform any of three types of t-tests. A. Test a single sample mean against a population mean B. Test the difference between two independent sample means. C. Test the difference between two correlated or dependent sample means. The t-test option results give you means and standard deviations, t-ratio, exact p-value, and the estimate of standard error used in the test. PROCEDURE. A. Select one of the three t-test options. B. Specify the hypothesized difference between means. The usual difference specified is 0. C. Enter sample names and whether or not you want printer output. D. Choose the data access method, and enter data or file names. NOTES Samples may be equal or unequal in size for the independent groups option. The program uses a pooled or weighted estimate of error for the independent groups design. Each data sample may contain no more than 500 cases. The t-test for 2 independent samples produces the point-biserial correlation in addition to the usual t-test results. T-TESTS 1. SINGLE SAMPLE 2. INDEPENDENT SAMPLES 3. RELATED SAMPLES 4. EXIT ENTER AN OPTION (1-4):1 ENTER SAMPLE/VARIABLE NAME 1. DATA HYPOTHESIZED MEAN:27 PRINTER COPY (Y,N):Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILE ENTER AN OPTION (1-3):2 DATA ENTRY: SEQUENTIAL FILES DATA FILE NAME: AAA SAMPLE N MEAN S.D. DATA 5 17.600 4.449 HYPOTHESIZED MEAN: 27 OBTAINED DIFFERENCE: -9.399 T(4)=-4.723 P=.0104 STANDARD ERROR = 1.989 3.10 ANALYSIS OF VARIANCE The analysis of variance program performs a complete analysis of variance for 1 or 2 factor designs, with equal or unequal sample sizes. The program assumes that the designs are completely randomized designs with an independent sample of observations in each treatment condition. Program output consists of a full ANOVA summary table with F ratios, p-values and a table of descriptive statistics for each cell of the design. PROCEDURE A. Define the experimental design in terms of number of factors and equal or unequal sample sizes. B. For unequal n designs, enter the sample size per condition. C. Name each factor in the design. The name may contain as many as six characters. Specify the number of levels for each factor and then assign a three character name for each level. D. For unequal n designs, the program prompts you to specify sample size for each treatment condition. E. Select a data access method. The ANOVA program permits only keyboard or sequential file methods. With the keyboard method, you simply enter all data for each condition when prompted. You can edit if necessary. Sequential file data entry requires that your data already reside in a single sequential file. The data in the file must be ordered appropriately for your design. You can see the correct data order in the following 2x3 design with unequal sample sizes and N = 21. b1 b2 b3 a1 3 5 7 1 5 6 2 6 4 a2 5 4 4 5 3 5 6 3 6 6 2 7 You would store these data in a sequential file with 21 records in the following order: 3 1 2 5 5 6 7 6 4 5 5 6 6 4 3 3 2 4 5 6 7 You can easily create this kind of file by using the SEQUENTIAL FILES program in DATA PREP. If your experimental design has a large number of treatment conditions and a small number of cases per condition (eg. n = 1), sequential file input is much more efficient than keyboard input. F. With TV monitor output, the screen displays each effect in the design, one at a time, with the terms and total SS (in equal n designs) displayed last. With printer output, you get a full ANOVA table in the standad format. G. The final step is the display of descriptive statistics which gives n, mean, s.d. and cell sum of squares for each condition. With printer output, these appear in a formatted table which includes marginal means. NOTES The program can also analyze two factor designs with n = 1 per cell. In such designs the interaction effect is used as the estimate of error. The sequential file must contain at least as many records as implied by the design specifications. Otherwise an end-of-data message occurs and the program branches to the exit menu. Each factor can have from 2 to 12 levels. Maximum sample size per experimental condition is 100. A fixed-effects model is assumed in computation of the F-ratios. ANALYSIS OF VARIANCE DESCRIPTIVE TITLE: ANOVA NUMBER OF FACTORS (1-2): 2 EQUAL OR UNEQUAL SAMPLE SIZES (E,U): U FACTOR 1 FACTOR NAME: A NUMBER OF LEVELS (2-12): 2 ENTER LEVEL NAMES 1. A1 2. A2 FACTOR 2 FACTOR NAME: B NUMBER OF LEVELS (2-12): 3 ENTER LEVEL NAMES 1. B1 2. B2 3. B3 ENTER SAMPLE SIZES A1B1: 3 A1B2: 3 A1B3: 3 A2B1: 4 A2B2: 4 A2B3: 4 DATA ACCESS 1. KEYBOARD 2. SEQUENTIAL FILE ENTER AN OPTION: (12):1 DATA FOR A1B1 DATA FOR A1B2 1. 3 1. 5 2. 1 2. 5 3. 2 3. 6 EDIT (Y,N): N DATA FOR A1B3 DATA FOR A2B1 1. 7 1. 5 2. 6 2. 5 3. 4 3. 6 4. 6 EDIT (Y,N): N DATA FOR A2B2 DATA FOR A2B3 1. 4 1. 4 2. 3 2. 5 3. 3 3. 6 4. 2 4. 7 EDIT (Y,N): N PRINTER COPY (Y,N):N SOURCE: A SS .570 DF 1 MS .570 F .56 P .4721 SOURCE: B SS 12.67 DF 2 MS 6.33 F 6.20 P .0109 SOURCE: AB SS 29.810 DF 2 MS 14.90 F 14.58 P <.001 SOURCE: ERROR SS 15.33 DF 15 MS 1.02 SOURCE: TOTAL SS 57.238 DF 20 MS 2.861 N MEAN S.D. A1B1 3 2.00 1.00 A1B2 3 5.33 .58 A1B3 3 5.66 1.53 A2B1 4 5.50 .58 A2B2 4 3.00 .82 A2B3 4 5.50 1.29 3.11 NONPARAMETRIC STATISTICS - STATS PLUS includes six of the most frequently used nonparametric test procedures. (1) Mann-Whitney U Test (2) Signed Ranks Test (3) Kruskal-Wallis Test (4) Friedman ANOVA by Ranks (5) Spearman Rho (6) Kendall Tau (7) Exit Each of these tests involves some kind of ranking of raw data. Since the programs treat all input data as raw data, the ranking is handled entirely by the programs. You do not need to rank the data before entering it. The ranking operation can be a fairly time-consuming process for large samples, although the programs use an efficient ranking method, a modified shell sort. During the sort, the programs are finding data ranks, keeping track of data group origin, and handling tied ranks. NOTES General Procedure. All six test procedures have the _ same general input requirements and output. Tied Ranks. When you have data resulting in tied ranks, the programs apply the corrections for tied ranks as described by Siegel (1956). P-Values. The programs compute test statistic probabilities as if the data sample were large enough to warrant inferences based on t and z distributions. When the data samples are relatively small, the p values appear with an asterisk (*) beside them. In such cases, you should consult published tables of the test statistic in question to determine more exact probability values. 3.11.1 MANN-WHITNEY U TEST You can apply this test to two independent variables of equal or unequal size. It is the nonparametric analog of the independent groups t-test. The Mann Whitney U Test does not calculate ties. PROCEDURE. A. Supply the name of two varaibles. B. Select the data access method. The variables you enter may be of equal or unequal length. C. For each variable, the program reports the sum of ranks, based on a ranking of both variables combined. The program also reports the U1 and U2 statistics. D. The significance test is based on the smaller of the U1 and U2 values. For sample sizes greater then 20, the significance of an observed value of U is determined by calculating the Z test statistic as explained by Siegel, 1956. MANN-WHITNEY U TEST DESCRIPTIVE TITLE: EXPERIMENT 3G ENTER SAMPLE/VARIABLE NAMES 1. XX 2. YY PRINTER COPY (Y,N):Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILE ENTER AN OPTION (1-3): 3 FILE NAME: TABLE FIELD # 1. X 2. Y 3. Z ENTER FIELD NUMBERS 1. 1 2. 2 MANN-WHITNEY TEST EXPERIMENT 3G SAMPLE N RANK SUM XX 25 734.00 YY 28 697.00 U1=291.00 U2=409.00 Z=-1.05 P=.2935 PRINT RAW DATA AND RANKS (Y,N):N 3.11.2 WILCOXON SIGNED RANKS TEST You can apply this test to data from two related variables,as a nonparametric analog of the related groups t-test. The program finds the differences between pairs of observations and then ranks the differences. PROCEDURE A. Enter the name of two varaiables. B. Select a data access method and enter data. C. The program reports the rank sum of differences and the z test statistic. WILCOXON SIGNED RANKS TEST DESCRIPTIVE TITLE: JULY EXPERIMENT ENTER SAMPLE/VARIABLE NAMES 1. T1 2. T2 PRINTER COPY (Y,N):Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILES ENTER AN OPTION (1-3): 3 FILE NAME: DATA FIELD # 1. TEST1 2. TEST2 3. TEST3 ENTER FIELD NUMBERS 1. 1 2. 2 JULY EXPERIMENT DATA SAMPLES T1, T2 N=12 SUM OF RANKED DIFFERENCES = 60 NUMBER OF TIES: 4 NUMBER OF ZERO DIFFERENCES: 0 P=.0958* PRINT RAW DATA AND RANKS (Y,N): N 3.11.3 KRUSKAL-WALLIS TEST You can apply this test to data from two or more independent variables. This test is a parallel to a one factor analysis of variance for a randomized design. The program accepts data for up to 5 groups with equal or unequal sample sizes. The Kruskal-Wallis Test does not compute ties. PROCEDURE A. Supply names for all variables. B. Select the data access method. The samples you enter may be of equal or unequal length. C. The program reports the rank sums for each group, information on the number of tied ranks, and the chi-square test statistic with its probability value. KRUSKAL-WALLIS TEST NUMBER OF DATA SAMPLES (2-5): 3 DESCRIPTIVE TITLE: SEMESTER 1 ENTER SAMPLE/VARIABLE NAMES 1. TEACH 2. ADMIN 3. ADMIN2 PRINTER COPY (Y,N):Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILES ENTER AN OPTION (1-3):3 FILE NAME: TRIALS FIELD # 1. TEACHER 2. ADMIN 3. ADMIN2 ENTER FIELD NUMBERS 1.1 2.2 3.3 KRUSKAL-WALLIS TEST SEMESTER 1 SAMPLE N RANK SUM TEACH 5 22 ADMIN 5 37 ADMIN2 4 46 H=6.40 P=.0399 PRINT RAW DATA AND RANKS (Y,N): N 3.11.4 FRIEDMAN ANALYSIS OF VARIANCE BY RANKS You can apply this test to data consisting of two to five related data samples. The test is comparable to one factor repeated measures analysis of variance. PROCEDURE A. Assign names to each data sample. B. Select the data access method. Enter data, making sure that the same number of data points are entered for each group. The program rejects an attempt to apply this test to samples of unequal size. C. The program performs the ranking operation. The ranking is done for all observations for each case or subject. D. The results show rank sums and the value of the chi-square test statistic. FREIDMAN ANOVA BY RANKS NUMBER OF DATA SAMPLES (2-5): 4 DESCRIPTIVE TITLE: FRIEDMAN ENTER SAMPLE/VARIABLE NAMES 1. I 2. II 3. III 4. IV PRINTER COPY (Y,N): Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILE ENTER AN OPTION (1-3): 1 * TO END DATA FOR I DATA FOR II 1. 9 1. 4 2. 6 2. 5 3. 9 3. 1 4. * 4. * DATA FOR III DATA FOR IV 1. 1 1. 7 2. 2 2. 8 3. 2 3. 6 4. * 4. * EDIT (Y,N):N RANKING DATA FRIEDMAN SAMPLE N RANK SUM I 3 11.00 II 3 5.00 III 3 4..00 IV 3 10.00 CHI-SQUARE = 7.400 P = .0591 PRINT RAW DATA AND RANKS (Y,N): N 3.11.5 SPEARMAN RHO This program computes the Spearman rank-order correlation coefficient for two variables. Since the program determines ranks before computing the correlation, the input data need not consist of ranks. PROCEDURE A. Assign names to your two variables. B. Select the data access method and enter data. C. The program produces rho and its significance test value. The program tests rho against an hypothesized population value of zero. SPEARMAN RHO DESCRIPTIVE TITLE: SPEARMAN ENTER SAMPLE/VARIABLE NAMES 1. AUTH 2. SSS PRINTER COPY (Y,N):Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILE ENTER AN OPTION (1-3): 1 DATA FOR AUTH DATA FOR SSS * TO END 1. 82 1. 42 2. 98 2. 46 3. 87 3. 39 4. 40 4. 37 5.116 5. 65 6.113 6. 88 7.111 7. 86 8. 83 8. 56 9. 85 9. 62 10.126 10. 92 11.106 11. 54 12.117 12. 81 13. * 13. * EDIT (Y,N):N SPEARMAN DATA SAMPLES AUTH, SSS N = 12 TIES: AUTH 0 SSS 0 RHO =.8252 P = .001 NUMBER OF TIES = 2 PRINT RAW DATA AND RANKS (Y,N): N 3.11.6 KENDALL TAU Like the Spearman rho, this test is a measure of correlation based on ranks and the ordering of data. The program counts the number of concordant and discordant sets of paired observations. Each pair is compared to all other pairs. PROCEDURE A. Assign sample names to your two samples. B. Select the data access method and enter data. Your data must be two samples of equal size. C. The program reports the numbers of concordant and discordant comparisons, tau value and the z test statistic. KENDALL TAU DESCRIPTIVE TITLE: KENDALL ENTER SAMPLE/VARIABLE NAMES 1. J1 2. J2 PRINTER COPY (Y,N): Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILE ENTER AN OPTION (1-3):1 DATA FOR J1 DATA FOR J2 * TO END 1. 3 1. 3 2. 4 2. 1 3. 2 3. 4 4. 1 4. 2 5. * 5. * EDIT (Y,N): N KENDALL DATA SAMPLES J1, J2 N = 4 CONCORDANT COMPARISONS: 2 DISCORDANT COMPARISONS: 4 TIED COMPARISONS, J1 0 TIED COMPARISONS, J2 0 TAU = -.333 P = .5041 3.12 CONTINGENCY TABLES The contingency tables program offers three common tests of frequency data. (1) Fisher Exact Test (2) One Variable Chi-Square Test (3) Two Variable Chi-Square Test These programs accept data in the form of frequency counts from keyboard only. There is no provision for direct file input. 3.12.1 FISHER EXACT TEST Option #1 performs a Fisher exact test on a 2x2 table of frequency data, reporting the exact 1-tailed probability of the given distribution of data and all more extreme distributions (Siegel, p.96). The total frequency may not exceed 32 for a given table. PROCEDURE. A. Supply names (up to 6 characters) for each row and column. B. Enter the observed frequencies into each of the four cells displayed on the screen. C. The program displays the row sums and column sums, and then computes a series of hypergeometric probabilities, displaying these finally as a 1-tailed probability. NOTES If the total frequency is more than 32, the program does not attempt to compute the exact probability. Use the two-variable chi-square procedure for total frequencies greater than 32. FISHER EXACT TEST ENTER ROW AND COLUMN NAMES ROW 1: A ROW 2: B COLUMN 1: X COLUMN 2: Y ENTER OBSERVED FREQUENCIES X Y TOTAL A 4 8 12 B 7 11 18 TOTAL 11 19 30 ONE TAILED PROBABILITY p = .288 PRINTER COPY (Y,N): Y 3.12.2 ONE VARIABLE CHI-SQUARE The one variable chi-square option performs a chi-square test for a single variable with up to ten categories or levels. The program reports observed and expected frequencies. It also reports chi-square for each category, the total chi-square, and its exact p-value. PROCEDURE. A. You supply a variable name up to 12 characters long, then the number of categories. You can specify from two to ten categories. B. The program asks if the expected values are uniform. If you respond YES, the program computes the expected value for each category by dividing TOTAL FREQUENCY by NUMBER OF CATEGORIES. C. Enter observed frequencies for each category. If you respond NO in step B to the UNIFORM FREQUENCIES prompt, you also enter the expected frequency for each observed frequency. D. The program computes chi-square for each category and reports total chi-square and the p-value. E. The printer copy option prints the full chi-square table in horizontal format. NOTE If expected frequencies are not uniform, the sum of the expected frequencies you enter must equal the total (observed) frequency. You can choose to apply the Yates' Correction test. ONE VARIABLE CHI SQUARE APPLY YATES' CORRECTION (Y,N): Y DESCRIPTIVE TITLE: CHI VARIABLE NAME: TRIALS NUMBER OF CATEGORIES (2-10): 4 UNIFORM EXPECTED FREQUENCIES (Y,N): Y CATEGORY OBSERVED EXPECTED CHI NAME FREQUENCY FREQUENCY SQUARE X1 13 12.75 .00 X2 10 12.75 .40 X3 20 12.75 3.57 X4 8 12.75 1.42 51 51 CHI-SQUARE(3)=5.39 p=.1445 PRINTER COPY (Y,N): N 3.12.3 TWO VARIABLE CHI-SQUARE The two variable chi-square option performs an analysis of a two-variable contingency table, testing an independence hypothesis. Each variable may contain from two to ten categories. The program reports: observed and expected frequencies and chi-square for each cell; row sums, column sums and total frequency; total chi-square, p-value and the contingency coefficient. PROCEDURE A. For each variable, supply a variable name of up to 12 characters. Then give the number of categories which can range from two to ten. Supply a name for each category. The name can be up to four characters long. B. Enter the observed frequencies in response to each prompt. The prompts, consisting of the appropriate row and column names, proceed row by row. C. If you select TV monitor display of results, the screen shows results for each cell, arranged in a column. As many as 17 cells are displayed at once. The CONTINUE response displays successive sets of results, followed by total chi-square, p-value and contingency coefficient. The printer copy option produces a rows by columns table with marginal sums. Marginal sums are not shown with TV monitor display of results. NOTES You can choose to apply Yates' Correction. Frequencies greater than 999 may cause formatting problems in printed output. TWO VARIABLE CHI SQUARE APPLY YATES' CORRECTION (Y,N): N DESCRIPTIVE TITLE: CHI VARIABLE NAME, ROWS: POLITICS NUMBER OF CATEGORIES (2-10): 3 ENTER CATEGORY NAMES 1. REP 2. DEM 3. IND VARIABLE NAME, COLUMNS: SEX NUMBER OF CATEGORIES (2-10): 2 ENTER CATEGORY NAMES 1. FEM 2. MALE ENTER OBSERVED FREQUENCIES REP FEM 15 REP MALE 22 DEM FEM 43 DEM MALE 13 IND FEM 35 IND MALE 30 PRINTER COPY (Y,N): Y CELL OBSERVED EXPECTED CHI NAME FREQUENCY FREQUENCY SQUARE REP F 15 21.78 2.11 REP M 22 15.22 3.02 DEM F 43 32.96 3.06 DEM M 13 23.04 4.37 IND F 35 38.26 .28 IND M 30 26.74 .40 CHI SQUARE (2) = 13.234 P = .001 CONTINGENCY COEFFICIENT = .278 3.13 CROSS TABULATION XTAB is the cross tabulation program which generates reports on the contents of random-access data files. XTAB can identify each record in your file as belonging to one of a set of categories or cross categories. After records have been classified or categorized, the program gives you the reports. XTAB can report category frequencies or percents, plus basic descriptive statistics. So that you can learn XTAB functions and terminology easily, a number of examples show you how to generate reports on a file called SURVEY. The file is a random-access file withsix fields per record. FILE NAME SURVEY NUMBER OF RECORDS 40 FIELDS PER RECORD 6 RECORD LENGTH 20 FIELD NAME WIDTH 1 SEX 1 2 AGE 2 3 EDUC 2 4 INCOM 6 5 #TRIPS 2 6 SATIS 1 You could use XTAB to answer many questions concerning the SURVEY file. How are the satisfaction ratings distributed? What are the average satisfaction ratings for male and female respondents? How are the respondents distributed in terms of a sex x age frequency classification? What are the average frequency and satisfaction data for males with university education? These examples illustrate the different types of file search and report capabilities of XTAB. XTAB can perform from a 1-way to a 5-way cross classification of a file, with a maximum of 200 categories. Each dimension (or variable or data field) in the classification can have from 1 to 9 categories. XTAB can report descriptive statistics on 1 to 3 data fields for each classification. There are three basic tabulation and cross-tabulation methods available. A. SINGLE CATEGORY. You can report the frequency of occurrence of records that fit a single category. For example: how many male respondents are there in the file? This directs XTAB to count the number of records with in Fielci #1. The single category can be defined in terms of 1 to 5 fields. For example, report the number of records of females with incomes greater than 15,000 and satisfaction ratings of 5. In this case, XTAB examines three fields to report on a single category. B. MULTIPLE CATEGORIES, ONE VARIABLE. In this mode, XTAB reports a frequency distribution. For example, you can classify all records in terms of the satisfaction rating given. Assuming that the ratings vary from 1 to 5, XTAB would report the frequency of occurrence of each possible rating. C. MULTIPLE CATEGORIES, MULTIPLE VARIABLES. This is a true cross-tabulation, in which records are classified according to a joint classification scheme. For example, you might want a 2-way frequency distibution of satisfaction ratings. SATISFACTION RATING 1 2 3 4 5 MALE 0 3 9 6 2 FEMALE 5 4 8 4 3 For any of the three basic tabulation modes, XTAB can report descriptive statistics as well as simple frequency data. You can perform a cross-classification in terms of respondent sex and education level AND find the average income, frequency and satisfaction for each classificaton. 3.13.1 EXAMPLES This section includes examples of each basic procedure. You should review the first example carefully. It provides instructions for defining classification schemes for XTAB and for setting up DATA CODE VALUES for classifying records. EXAMPLE 1. FREQUENCY DISTRIBUTION Suppose you want to know how satisfaction ratings are distributed. The rating data in the SURVEY FILE range from 1 (poor) to 5 (excellent). Select FREQUENCY COUNT from the report options. CROSS TABULATION REPORT OPTIONS 1. FREQUENCY COUNT 2. FREQUENCY AND PERCENT 3. STATISTICS 4. CORRELATION 5. CATALOG 6. EXIT ENTER AN OPTION (1-6): 1 FILE NAME: SURVEY FIELD# NAME 1 SEX 2 AGE 3 EDUC 4 INCOM 5 #TRIPS 6 SATIS CROSS-TAB ON HOW MANY FIELDS (1-6): 1 ENTER FIELD NUMBER 1. 6 NUMBER OF CATEGORIES (1-9): 5 CATEGORY # NAME CODE 1 POOR 1-* 2 FAIR 2-* 3 GOOD 3-* 4 V.GD 4-* 5 EXCEL 5-* SPECIFICATIONS OK (Y,N): Y PRINTER COPY (Y,N): Y SCANNING FILE SURVEY CLASSIFICATION N POOR 1 FAIR 7 GOOD 17 V.GD 10 EXCEL 5 The table shows the number of records found for each category and the total records categorized. Records that do not fit any of the defined categories are given as UNCLASSIFIABLE RECORDS. If you had selected report option #2, frequency and per cent, the table for the previous frequency distribution would include the per cent of all classifiable records that each category frequency represents. NOTES You can perform this frequency distribution on discrete or continuous data. You can ask for a frequency distribution for class test scores ranging from 0 to 100, or you can do a distribution on five rating scores. CATEGORY NAME is just a descriptive label which you choose. The label can be up to six characters long. CATEGORY CODE is your specified search code. To perform a cross-tabulation, you must give a search code based on the data in the field being searched. To search for all cases with a score of 1 in the field of interest, your search code is 1-*. To search for all cases coded with M, use the code M-*. To search for all cases with a score falling in the range 50 to 100, use the search code 50-100. To search for scores falling in a negative range, specify the smaller number first. For example, to search for numbers in the range -8 to -2, use the search code -8--2. When you specify the search codes for a variable you may use both specific values and ranges. However the code used for specific values must be entered like a range. For example, the specific value 7 should be entered as 7-7. EXAMPLE 2. 2-WAY CROSS TABULATION _ This example illustrates a 2-way cross-tabulation in which you classify each record according to the data values it has in two data fields. It also illustrates the use of range search codes. Suppose you want to classify the records in the SURVEY file in terms of sex of respondent (male, female) and income. You want to have four income categories: 5000-14999, 15000-24999, 25000-34999, and >= 35000. You want a report of frequencies and per cents. CROSS TABULATION REPORT OPTIONS 1. FREQUENCY COUNT 2. FREQUENCY AND PER CENT 3. STATISTICS 4. CORRELATION 5. CATALOG 6. EXIT ENTER AN OPTION (1-6): 2 FILE NAME: SURVEY FIELD# NAME 1 SEX 2 AGE 3 EDUC 4 INCOM 5 #TRIPS 6 SATIS CROSS-TAB ON HOW MANY FIELDS (1-5): 2 ENTER FIELD NUMBERS 1. 1 2. 4 VARIABLE SEX NUMBER OF CATEGORIES (1-9): 2 CATEGORY # NAME CODE 1 MALE M-* 2 FEMALE F-* SPECIFICATIONS OK (Y,N): Y VARIABLE INCOME NUMBER OF CATEGORIES (1-9):4 CATEGORY 3 NAME CODE 1 ONE 5000-14999 2 TWO 15000-24999 3 THREE 25000-34999 4 FOUR 35000-150000 SPECIFICATIONS OK (Y,N):Y PRINTER COPY (Y,N): Y SCANNING FILE SURVEY CROSS-TAB ON FILE SURVEY INCOME ONE TWO THREE FOUR SEX ---------------------------- MALE 1 3 4 2 10 5% 15% 20% 10% 50% FEMALE 3 5 1 1 10 15% 25% 5% 5% 50% ---------------------------- UNCLASSIFIABLE RECORDS: 0 Note that you specify a tabulation on TWO data fields, and supply the field numbers involved. The search codes for SEX are M-* and F-*. The search codes for INCOME are range search codes. For example, any record with a value from 5000 to 14999 in the income field would be assigned to category ONE of the income classification. As this example indicates, both types of search codes may be used in a given search and classification problem. The printer copy of the table shows the frequency and percent for each cell of the joint classification, and the row and column sums and percents. You can perform a cross-tabulation on as many as 5 variables/fields at a time. For example, the sex x income problem could be extended to a 3-way sex x education x income classification. When the classification is based on more than two variables or fields, the program output is organized as a series of two-way tables. EXAMPLE 3. SINGLE CATEGORY You may want to obtain a report on a single categorization of file data, such as a simple count of the number of males, or the number of females with post-graduate education, or the number of males with frequency scores of 10 or greater and satisfaction ratings of 4 or 5. For this kind of report, you just indicate that there is one category tor each field/variable, and supply the search code(s) required. In the following example, the file is searched to report the number of females with university education . CROSS TABULATION 1. FREQUENCY COUNT 2. FREQUENCY AND PERCENT 3. STATISTICS 4. CORRELATION 5. CATALOG 6. EXIT ENTER AN OPTION (1 -6): 1 FILENAME: SURVEY FIELD # NAME 1. SEX 2. AGE 3. EDUC 4. INCOM 5. #TRIPS 6. SATIS CROSS TAB ON H0W MANY FIELDS (1 -5): 2 ENTER FIELD NUMBERS 1. 1 2. 3 VARIABLE SEX NUMBER OF CATEGORIES ( 1 -9 ): 1 CATEGORY # NAME CODE 1 FEMALE F-* SPECIFICATIONS OK (Y,N ): Y VARIABLE EDUC NUMBER OF CATEGORIES (1-9): 1 CATEGORY # NAME CODE 1 UNIV UN-* SPECIFICATIONS OK (Y,N):Y CROSS-TAB ON FILE SURVEY EDUC UNIV SEX ------ FEMALE 11 11 ------ 11 11 UNCLASSIFIABLE RECORDS: 18 EXAMPLE 4. STATISTICS REPORTING In addition to classifying and counting records, XTAB can also report basic descriptive statistics for the records in each classification. For example: cross-classify records in terms of sex and education level (as in example #2), and report the statistics on the frequency and satisfaction data for each cell of the sex x education classification. To obtain this kind of report, you select report option #3, STATISTICS, indicate the number of fields for which statistics should be computed, and then supply the field numbers. All of the rest of the search & classification definitions are the same as they are for frequency reports. When you select the statistics option, the form of the report is different from the tabular format used for frequency reports. The statistics option report is a series of statistical summaries for each cell of the classification. Each statistics report is labelled with the cross-classification involved. The program reports the number of records for each classification, the sum, mean, standard deviation, and minimum and maximum values. It can report these statistics for up to 3 data fields simultaneously. CROSS TABULATION 1. FREQUENCY COUNT 2. FREQUENCY AND PERCENT 3. STATISTICS 4. CORRELATION 5. CATALOG 6. EXIT ENTER AN OPTION (1-6): 3 FILENAME: SURVEY FIELD # NAME 1. SEX 2. AGE 3. EDUC 4. INCOM 5. #TRIPS 6. SATIS STATISTICS ON HOW MANY FIELDS (1-3): 2 ENTER FIELD NUMBERS 1. 5 2. 6 CROSS-TAB ON HOW MANY FIELDS(1-5): 2 ENTER FIELD NUMBERS 1. 1 2. 3 VARIABLE SEX NUMBER OF CATEGORIES (1-9): 2 CATEGORY # NAME CODE 1 FEMALE F-* 2 MALE M-* SPECIFICATIONS OK (Y,N): Y VARIABLE EDUC NUMBER OF CATEGORIES (1-9): 3 CATEGORY # NAME CODE 1 GS GS-* 2 HS HS-* 3 UN UN-* SPECIFICATIONS OK (Y,N): Y PRINTER COPY (Y,N): Y SCANNING FILE SURVEY FEMALE GS #TRIPS SATIS N 3 3 SUM 14.00 12.00 MEAN 4.5 3.00 S.D. 4.94 1.41 MIN. 1.00 1.00 MAX 8.00 4.00 FEMALE HS #TRIPS SATIS N 6 6 SUM 12.00 10.00 MEAN 8.50 4.00 S.D. 3.60 2.80 MIN 1.00 1.00 MAX 8.00 5.00 FEMALE UN #TRIPS SATIS N 11 11 SUM 25.00 18.00 MEAN 3.50 3.00 S.D. 3.10 4.20 MIN. 2.00 2.00 MAX. 9.00 5.00 MALE GS #TRIPS SATIS N 9 9 SUM 30.00 16.00 MEAN 4.6 3.5 S.D. 4.25 3.8 MIN 2.00 1.00 MAX 10.00 5.00 MALE HS #TRIPS SATIS N 6 6 SUM 18.00 12.00 MEAN 6.33 3.00 S.D. 2.57 1.60 MIN. 1.00 2.00 MAX. 7.00 4.00 MALE UN #TRIPS SATIS N 5 5 SUM 21.00 16.00 MEAN 5.00 3.00 S.P. 2.63 2.42 MIN. 1.00 2.00 MAX. 6.00 5.00 4. STATS PLUS EXAMPLE This is a step by step example of the use of STATS PLUS. This example is given to quickly familiarize you with the program. For more detailed discussions of these procedures, please refer to the Index at the end of this documentation for the appropriate documentation sections. Follow the example step by step. Always follow every response with the RETURN key. STEP 1. Boot the STATS PLUS disk to get the MAIN MENU. MAIN MENU 1. SYSTEM SETUP 2. RUN DATA PREP 3. RUN STATISTICS 4. END Choose Option 1, SYSTEM SETUP. The screen shows the SYSTEM CONFIGURATION MENU. SYSTEM CONFIGURATION 1. PROGRAM DISK IN DRIVE 1 2. DATA DISK IN DRIVE 1 3. VOLUME NUMBER 254 4. PRINTER SLOT 1 5. PRINTER INTERFACE PARALLEL ANY CHANGES (Y,N)?: Make any necessary changes in the configuration. When you are through, answer N to the ANY CHANGES prompt. The program writes the specified configuration to the program disk, then returns you to the main menu. STEP 2. You are ready to work through an entire analysis. Make sure that you have a DOS 3.3 initialized disk ready for data storage. Insert that data disk when prompted by the program.When the prompt DATA DISK IN PLACE? appears, place the data disk in your designated data disk drive and press the RETURN key. In the example problem, you are creating a data file and performing a t-test on the data. The data consist of measures of strength of a material before and after a ~ stressing procedure. The data are in the following table. CASE BEFORE AFTER 1 12 9 2 15 11 3 13 10 4 12 10 5 16 11 6 12 11 7 14 9 8 13 10 9 15 9 10 12 8 STEP 3. With this problem in mind, look at the MAIN MENU on your screen. MAIN MENU 1. SYSTEM SETUP 2. RUN DATA PREP 3. RUN STATISTICS 4. END To set up a data file, choose Option 2, RUN DATA PREP. The screen next shows the DATA PREP MENU. DATA PREP 1. RANDOM-ACCESS FILES 2. SEQUENTIAL FIILES 3. TRANSFORMATIONS-R.A. FILES 4. TRANSFORMATIONS-SEQ. FILES 5. SEARCH & SELECT 6. RETURN TO MAIN MENU STEP 4. From the DATA PREP MENU, choose Option 1, RANDOM-ACCESS FILES to get the RANDOM-ACCESS FILES PROCEDURES MENU. RANDOM-ACCESS FILES PROCEDURES 1. SET UP NEW FILE 7. MAKE SUBFILE 2. FILE DESCRIPTION 8. RAF=>SQF 3. DATA ENTRY 9. SQF=>RAF 4. REVIEW/EDIT 10. PRINT FILES 5. MERGE FILES (V) 11. DELETE FILES 6. MERGE FILES (H) 12. CATALOG 13. EXIT STEP 5 From the RANDOM ACCESS FILES PROCEDURES MENU, choose Option 1, SET UP NEW FILE. When the program prompts you, give the new file the name STRENGTH. Since the data consist of 2 variables, specify 2 fields for the file. Give a name for each field and a field width equal to the number of characters necessary in that field. The field width specifies the number of digits which can be accepted in that field. FILE NAME: STRENGTH NUMBER OF FIELDS: 2 FIELD # NAME WIDTH 1 BEFORE 2 2 AFTER 2 The program writes this information to your data disk and - then returns to the RANDOM ACCESS FILES PROCEDURES MENU. RANDOM-ACCESS FILES PROCEDURES 1. SET UP NEW FILE 7. MAKE SUBFILE 2. FILE DESCRIPTION 8. RAF=>SQF 3. DATA ENTRY 9. SQF=>RAF 4. REVIEW/EDIT 10. PRINT FILES 5. MERGE FILES (V) 11. DELETE FILES 6. MERGE FILES (H) 12. CATALOG 13. EXIT STEP 6. To enter the raw data into the STRENGTH file, select Option 3 from the RANDOM-ACCESS FILES PROCEDURES MENU. Then give the file name when prompted. The program asks you to enter data, starting with the first case, RECORD #1. Enter the data for each case and edit any typos. Repeat this for all 10 cases, and enter an asterisk (*) when the program prompts you for the eleventh record. After receiving the asterisk, the program creates the STRENGTH data file. RECORD 1 1 BEFORE 12 2 AFTER 9 RECORD 2 1 BEFORE 15 2 AFTER 11 RECORD 3 1 BEFORE 13 2 AFTER 10 * * * * RECORD 11 1 BEFORE * STEP 7. To get back to the MAIN MENU, choose Option 13 from the RANDOM ACCESS FILES PROCEDURES MENU. The EXIT MENU appears. 1. RERUN PROGRAM 2. EXIT TO MAIN MENU 3. EXIT TO DATA PREP 4. EXIT TO STATISTICS 5. END Choose Option 4, STATISTICS. STATISTICS MENU 1. DESCRIPTIVE STATISTICS 2. FREQUENCY DISTRIBUTION 3. CORRELATION 4. LINEAR REGRESSION 5. MULTIPLE REGRESSION 6. T-TESTS 7. ANOVA 8. NONPARAMETRIC TESTS 9. CONTINGENCY TABLES 10.CROSS-TABULATION 11.RETURN TO MAIN MENU STEP 8. To calculate t-tests, choose Option 6, T-TESTS. Answer program prompts for type of data access, file name and type of output required. Choose related samples t-test because the data come from repeated tests. Make sure that your printer is turned on if you want printed output. T-TESTS 1. SINGLE SAMPLE 2. INDEPENDENT SAMPLES 3. RELATED SAMPLES ENTER AN OPTION (1-3): 3 ENTER SAMPLE/ VARIABLE NAMES 1. BEFORE 2. AFTER HYPOTHESIZED DIFFERENCE: 0 PRINTER COPY (Y,N):Y DATA ENTRY OPTIONS 1. KEYBOARD 2. SEQUENTIAL FILES 3. RANDOM-ACCESS FILES ENTER AN OPTION (1-3): 3 FILE NAME: STRENGTH FIELD # 1 BEFORE 2 AFTER ENTER FIELD NUMBERS 1. 1 2. 2 SAMPLE N MEAN S.D. BEFORE 10 13.400 1.506 AFTER 10 9.800 1.033 HYPOTHESIZED DIFFERENCE: 0 OBTAINED DIFFERENCE: 3.600 T(9)=7.562 P<.001 STANDARD ERROR = .476 POINT-BISERIAL CORRELATION = .930 That's all there is to it. You are probably ready now to do your own analyses, since all parts of the program prompt you. If you have any questions, just read the appropriate section in the documentation. The rest of the documentation presents all of the details of file creation and data transformation, along with alternative methods of entering data, and the available analyses. If you have any comments concerning this or any of our programs, please contact us. We are extremely interested in hearing your opinion. 5. REFERENCES These are the texts containing discussions and examples of the statistical analysis procedures in STATS PLUS. Conover, W. J. Practical Nonparametric Statistics. New York: Wiley, 1971. Bruning, J. L., & Kintz, B. L. Computational Handbook of Statistics. Dallas: Scott Foresman, 1977. Hoel, P. G., & Jessen, R. J. Basic Statistics for Business and Economics. New York: Wiley, 1982. Levin, R. I. Statistics for Management. Englewood Cliffs, N.J.: Prentice-Hall, 1978. Siegel, S. Nonparametric Statistics. New York: McGraw-Hill, 1956. 6. ERROR MESSAGES STATS PLUS error messages are ordinary Applesoft messages. Error # Meaning 1 language not available 2,3 range error 4 write protected 5 end of data 6 file not found 7 volume mismatch 8 I/O error 9 disk full 10 file locked 11 syntax error 12 no buffers available 13 file type mismatch 14 program too large 15 not direct command 55 illegal quantity