Psy 401 - Personality Research Methods

Dr. John A. Johnson

Instructions for Second Lab Project:

Due October 27, 1998

USING THE CALIFORNIA Q-SET AND CALCULATING RELIABILITY

 

Objectives

  1. Learn to translate personality impressions into structured Q-sort profile.
  2. Learn to enter Q-set data into a computer file and to calculate Cronbach alpha reliability.
  3. Learn how adding and subtracting raters affects composite reliability.

Sections marked with the group picture represent regular instructions for students who are taking this course with a full class of other students. Sections marked with the individual picture represent special instructions for students who are taking this course in an independent study format.

Procedure

We will first divide into groups of 4. Each group will receive a set of five editorials written by columnist Charley Reese. Pass the editorials around until each member of the group has read all four. If you like, you may jot down impressions of Mr. Reese's personality as you form them. Please do not discuss the editorials with the other members of your group. When you have finished reading the editorials, construct a 9-point Q- sort distribution of Mr. Reese's personality as instructed in class and as described on the CQ-set record sheet. Transfer the category values to the copy of the California Q-set Record sheet given to you in class. Throughout this procedure there should be no consultation between members of the group. In order to share each other's ratings, you will need to get the name and CAC access userid from each person in your group. At the very first opportunity, (HINT: like, today. right now.) create an Excel spreadsheet file from your Q-set record sheet. The file should be in the first column and the data portion should be 100 rows long. In the 101st row, type your computer userid (e.g., Rickie Lee Jones would type rlj123 if that were her userid) in this last cell. Save the file as a regular Excel file using your own userid when naming your file, e.g., Rickie Lee Jones would call her file rlj123.xls

When you have finished creating your file of q-sort information, you will need to send copies to the other members of your group. You can do this by saving all group members' files on diskette. But unless you are all working in the same computer lab at the same time, it would probably be much easier to email your Excel files (save in Excel format, rlj123.xls) to each other as attachments. In addition to the data files sent to you by persons in your group, you will also soon see in your mailbox two files from me. Their names are sub111.xls and bgs999.xls. After receiving everyone's files you can open them in Excel and copy the column of data into your own Excel file to create a single data file. Each column will have data from one person in your group; the 101st cell in each column will contain the userid of the person (e.g., rlj123). Save the file in regular Excel format, naming it totqdat.xls

Students who are taking the course in an independent format will receive the additional q-sort ratings as an Excel file, addqdat.xls. The instructor will send this file to you as an email attachment. You should add your own ratings to it and save it as totqdat.xls as described above.

From this point on, directions are the same for both students in a regular class and students completing this lab through independent study

Uploading the Q-Data File

The next step is to upload the data file to the mainframe computer. First, delete row 101, which contains the userids of each rater. You will still be able to refer back to your .xls formatted file if you lose track of which rater is in which column. Next, minimize the column widths with the Format->Columns->AutoFit command. Then save the file as either a .txt or .prn text file (totqdat.txt or totqdat.prn). The prn format is probably the better choice, because txt separates the columns with tabs and prn, with spaces. Files with tabs in them can be used on the mainframe, but can sometimes be confusing. Log on to PSUVM and upload your text file, specifying its name on the mainframe as TOTAL QDAT. Your uploaded file should be 100 lines long, and its name should appear in the form TOTAL QDAT A. I recommend numbering the prefix area of your file by typing NUM ON next to the command line arrow i.e., ====> NUM ON. This way you can confirm that your file is 100 values long.

Using the SPSS Reliability Procedure

Below is an SPSS program file, designed for the present example, for computing reliability. If you are reading an electronic version of this lab, you can copy and paste it into a file of your own. You can then edit it to fit your own data set and upload it to the mainframe. Instructions for doing this are given in the next section of these instructions. The remainder of this section describes what each line of the SPSS program file means.

UNNUMBERED
TITLE 'PSY 401 RELIABILITY EXAMPLE'
SET COMPRESSION=ON LENGTH=NONE WIDTH=80
FILE HANDLE DATAIN1 /NAME='TOTAL QDAT A'
DATA LIST FILE=DATAIN1 RECORDS=1
  /1 BVK122 1 HRM102 2 TLC189 3 TIR137 4 SUB111 5 BGS999 6
COMPUTE TOTRATE=BVK122+HRM102+TLC189+TIR137+SUB111+ BGS999
RELIABILITY VARIABLES=TOTRATE, BVK122 TO BGS999
  /SCALE(TOTRATE)=BVK122 TO BGS999/SUMMARY=TOTAL
FINISH

The first three lines merely define the formatting of the output should you be printing the results of your analysis (and eventually you will). The title placed between the single quotations '<title>' will appear at the top of each page of your output. The COMPRESSION= and LENGTH= commands save space, and the WIDTH= command can be used to limit the number of columns if you are printing to 8 1/2 by 11" paper. If you don't use the WIDTH= command, SPSS will print in 133 columns, the standard width for high-speed printer paper.

The FILE HANDLE command tells SPSS where to find your data file(s). In this case, we have only one data file, TOTAL DAT A, and we are calling it DATAIN1 (for data input file number 1) within the SPSS program. Actually, you can call it anything you wish. I use DATAIN1, and if SPSS needs to read files with additional variables in my study, I call them DATAIN2, DATAIN3, etc.

The DATA LIST command begins by pointing to data file you just named in the previous statement, FILE=DATAIN1. Next, the RECORDS= command tells SPSS how many lines (records) are required for each case. [Here, each Q-set item represents one case. More typically, each case represents one participant in the study.] In this example, we have only 6 ratings per item that fit easily on to one line, so RECORDS=1. For larger problems, you might have to use more than one record per case. Next follows the names of the variables on each record and where each variable is located. In our example, the "variables" are the raters, labeled with the userid of each rater. [Again, more typically, the variables in an analysis will represent a set of responses to test items or a set of test scores for a single research participant.] The line beginning with the indented /1 tells SPSS that on the first record we find the ratings of BVK122 in column 1, HRM102 in column 2, etc. [If your variables had required, say, three columns, the format would be /1 AAA111 1-3 BBB222 4-6, etc. If you had needed two records per case, the second format statement would begin with /2, directly under the line beginning with /1.]

IMPORTANT NOTE FOR FUTURE PROGRAM WRITING: The maximum length of any line in an SPSS program file is 80 columns. If one of the statements in a program looks like it will extend beyond 80 columns (say you have a lot of variable names on a single record), break before you get to column 80 and continue with the statement on the next line, indenting at least one space. SPSS assumes that any line not beginning in column 1 is a continuation of the previous line.

The COMPUTE command creates a new variable that is the total of all the individual ratings, which is necessary to do reliability statistics. Finally, the actual statistical procedure is found on the line with the RELIABILITY command. Here, RELIABILITY gives the names of the variables contributing to the reliability (the total of all ratings and each individual rating). The (indented) line that begins /SCALE(TOTSCALE)=BVK122 to BGS999 is actually part of the RELIABILITY command. The /SCALE subcommand tells SPSS which variable is the scale and which variables are the items. /SUMMARY=TOTAL line is tells SPSS which specific reliability statistics you wish it to compute, in this case, item-total statistics. This will be explained further in class.

Creating Your SPSS Program File

Just as with any other file, you can create an SPSS program file with the XEDIT editor. When you make up a name for your program files, it is a good idea to follow the convention that the file type or extension (second part of the file name) is SPSS. The first part of the file name can be anything that indicates to you what this program does. A good name for the present program might be something like RELQDAT SPSS A to indicate that your program does a reliability analysis on your QDAT file. Let me use that more specific name now.

You have two options for creating your RELQDAT SPSS file. If you are reading an electronic version of this lab with a web-browser or word processor, you can copy, edit, and upload the sample program above. To do this, highlight the sample program by clicking and dragging over all the line. Then choose Edit->Copy from the menu (or just type Ctrl-C). Open up either your word processor, or, even better, a simple text editor like Notepad or Wordpad. Use Edit->Paste (or Ctrl-V) to paste the copied program into your document. Next, change the variable names (raters' userids), in the last four lines before the FINISH command to the variable names for your data set.

The next step is to save the document, using any name you wish, as a plain text file. The Notepad editor automatically saves in plain text, so I recommend it for that reason. If you use a full-fledged word processor or Wordpad, you have to explicitly tell your editor to save in plain text format. Wordpad has a further annoyance of automatically adding the suffix ".txt" to the file name you choose when saving in plain text format if the name you choose doesn't end in ".txt" (unless you put the file name between double quotation marks). The bottom line is that Notepad is probably your best choice for a text editor.)]

After saving the file, you will need to upload it to the mainframe. Use the same procedure you used to upload your q-sort data. Set the name of the file you are uploading to RELQDAT SPSS A.

The alternative, to copying, editing, and uploading the sample program is to create the RELQDAT SPSS file from scratch. To do this, you can choose FILES from the PSUMENU Main Menu and then XEDIT from the FILES menu. When the window pops up asking for the file name, type in RELQDAT in the blank next to fname and SPSS in the blank next to ftype. The XEDIT screen should appear, and you can type INPUT (or just I for short) to begin writing your lines of instruction. (If you don't want to use PSUMENU, you can simply type in the XEDIT command from any bottom-line command line like this: =====> XEDIT RELQDAT SPSS A
and start typing your program.) When you have finished writing and editing all of the lines of your SPSS program file, you will want to save it with the FILE command on the command line:

=====> FILE

Running Your SPSS Program File

To run the program, either choose FILELIST from the Files Menu or type FILELIST from any command line, =====> FILELIST. You will get a list of files that includes all of your files, including the one you have just created, RELQDAT SPSS A. Since you just created it, it should appear at the top of your file listing. Typing the command SPSS to the left of the file and hitting RETURN or ENTER will send your instructions to the main SPSS processor, which will read and follow the instructions:

USR123 FILELIST A0 V 108 Trunc=108 Size=14 Line=1 Col=1 Alt=2

Cmd

Filename

Filetype

Fm

Format

Lrecl

Records

Blocks

Date

Time

SPSS

RELQDAT

SPSS

A1

F

80

12

2

10/17/98

22:32:32

 

TOTAL

QDAT

A1

F

80

100

2

10/14/98

9:44:23

If you have program errors, SPSS will tell you what they are. If you can't figure out how to correct the errors, send me an email note. When your program is error free, you will simply get a message like

END OF JOB: 11 COMMAND LINES 0 ERRORS 0 WARNINGS .3 CPU SECONDS

with MORE . . . in the lower right corner. Hit the CLEAR button, which will bring you back to your list of files, and press the F2 key to REFRESH the screen. You will now see a file with the file type LISTING and the same file name as the SPSS program file you just ran. LISTING is always the file type of a file representing output from SPSS.

USR123 FILELIST A0 V 108 Trunc=108 Size=14 Line=1 Col=1 Alt=2

Cmd

Filename

Filetype

Fm

Format

Lrecl

Records

Blocks

Date

Time

 

RELQDAT

LISTING

A1

V

133

205

3

10/17/98

22.34:03

 

RELQDAT

SPSS

A1

F

80

12

2

10/17/98

22:32:32

 

TOTAL

QDAT

A1

F

80

100

2

10/14/98

9:44:23

 

Note under the column Lrecl that the line record length of the listing file is 133 and, under the Format column, that the length format is V, which stands for variable. This means that lines are of variable length and could be as wide as 133 characters (the width of computer paper for the high-speed printers). LISTING files are always created with the potential for 133- character lines. But the WIDTH=80 specification on the third line of your SPSS program file ensures that the lines will actually be only 80 characters long, maximum. Files with this width are easier for word-processors to handle if you were to edit this file with the word processor of your personal computer. They are also easier to read when you XEDIT them, especially when you turn off the prefix area from the command line with =====> PREFIX OFF.

Reading and Interpreting Your Results

Let's take a look now at the file by placing the cursor to the left of its name in the file list and pressing the XEDIT key, F11. If you page through the file with the the F8 key (FORWARD) you will find, first, a short advertisement about the wonderful features and options of SPSS, followed by a listing of the lines from your instruction program. After the DATA LIST instruction line is printed, you will see a table listing your variables (raters, in this case) and the columns you've told SPSS to put them in:

1 0
2 0 UNNUMBERED
3 0 TITLE 'PSY 401 RELIABILITY EXAMPLE'
4 0 SET COMPRESSION=ON LENGTH=NONE WIDTH=80
5 FILE HANDLE DATAIN1 /NAME='TOTAL QDAT A'
6 DATA LIST FILE=DATAIN1 RECORDS=1
7 /1 BVK122 1 HRM102 2 TLC189 3 TIR137 4 SUB111 5
8 BGS999 6
0THIS COMMAND WILL READ 1 RECORDS FROM TOTAL QDAT A1

0VARIABLE

REC

START

END

FORMAT

BVK122

1

1

1

F1.0

HRM102

1

2

2

F1.0

TLC189

1

3

3

F1.0

TIR137

1

4

4

F1.0

SUB111

1

5

5

F1.0

BGS999

1

6

6

F1.0

The table above shows that SPSS thinks that BVK122's values start and end in column 1, HRM102's start and end in column 2, etc. It is a good idea to double-check this table to verify which rater is in which column of your data file.
Your listing file will next show the rest of your SPSS instructions and then report some details on how SPSS is handling the analysis; you needn't concern yourself with these details. The bottom line comes at the end of the file when the actual results are reported. It will look something like this:
0 R E L I A B I L I T Y A N A L Y S I S - S C A L E (T O T R A T E)
1. BVK122
2. HRM102
3. TLC189
4. TIR137
5. SUB111
6. BGS999
-ITEM-TOTAL STATISTICS

0

SCALE
MEAN
IF ITEM
DELETED

SCALE
VARIANCE
IF ITEM
DELETED

CORRECTED
ITEM_
TOTAL
CORRELATION


ALPHA
IF ITEM
DELETED

BVK122

24.3000

68.1919

.7522

.8321

HRM102

24.3900

67.6140

.7691

.8290

TLC189

24.7600

69.2145

.6872

.8435

TIR137

24.6400

73.3640

.5784

.8622

SUB111

24.2200

69.7693

.6871

.8436

BGS999

24.2900

74.3898

.5400

.8688

 

 

 

 

 

0
0RELIABILITY COEFFICIENTS
0N OF CASES = 100.0 N OF ITEMS = 6
0ALPHA = 0.8692

-
(By the way, if you have been wondering why the first column of the listing output is usually blank but occasionally has a 0 or 1 or - in it, these are "control characters" that affect the spacing and paging of output on the high- speed line printer. You can just ignore this first column, but if really bothers you, you can delete it with the COLEDIT (column edit) procedure, discussed later.)

What you want to look at first is the ALPHA = result at the end of the results to judge whether the overall alpha reliability is acceptable. You can next look at the ITEM-TOTAL CORRELATION column, which shows how well each individual's ratings matched the average of the rest of the group. The last column shows how the alpha reliability coefficient would be affected if you removed a rater's ratings from the analysis. A "good" rater is one whose ratings correlate highly with the remaining judges' ratings. An alpha reliability coefficient will normally decrease if you remove a good rater's ratings. The "weakest" rater is the one whose ratings contribute least to the overall reliability. Weak ratings do not correlate well with the remaining ratings, and the alpha will increase if a weak rating is removed (see the last column in the table, ALPHA IF ITEM DELETED).

Note that the summary item-total statistics don't tell you how the alpha reliability would change if you removed the ratings of two raters. The only way to assess this would be to XEDIT your program file and eliminate the names of the two raters in the RELIABILITY command. For example, if I wanted to remove the ratings of TIR137 and BGS999, I would change the subcommand /SCALE(TOTRATE)=BVK122 TO BGS999 to

/SCALE(TOTRATE)=BVK122,HRM102,TLC189,SUB111

and run the program again.

If you believe that your first alpha reliability coefficient from all of the raters is sufficiently high, you needn't worry about deleting raters and recalculating reliability. But if one or more set of ratings seriously detract from reliability, you should recalculate alpha without those ratings. Furthermore, any additional analyses you conduct (including identifying descriptive items, below) should be based only upon data from the acceptable raters.

Identifying the Most and Least Descriptive Items with COLEDIT

Now that you've established that your ratings have acceptable reliability or you have eliminated ratings until you have achieved acceptable reliability, it is time to summarize your findings of the person who was rated. Typically this is done by identifying the Q-set statements that are most descriptive (had the highest mean rating across reliable observers) and those that are least descriptive (had the lowest mean rating across reliable observers). The easiest way to find these items by computing means and sorting them with a command called COLEDIT.

First, to calculate the mean ratings for every item, XEDIT your RELQDAT SPSS program file. This file has already been set up to read the data in your TOTRATE QDAT file, so we will only change some of the lines, shown below:


UNNUMBERED
TITLE 'PSY 401 RELIABILITY EXAMPLE'  <--------------|should change title, even
SET COMPRESSION=ON LENGTH=NONE WIDTH=80             |though it doesn't affect
FILE HANDLE DATAIN1 /NAME='TOTAL QDAT A'            |the way the program runs.
DATA LIST FILE=DATAIN1 RECORDS=1 
  /1 BVK122 1 HRM102 2 TLC189 3 TIR137 4 SUB111 5 BGS999 6 
COMPUTE TOTRATE=BVK122+HRM102+TLC189+TIR137+SUB111+BGS999 
<------------------------------------------------will need to add a line here
RELIABILITY VARIABLES=TOTRATE, BVK122 TO BGS999        |need to change the command
  /SCALE(TOTRATE)=BVK122 TO BGS999/SUMMARY=TOTAL       |to compute descriptive statistics
FINISH

The results will look like the following:

 UNNUMBERED 
 TITLE 'DESCRIPTIVE STATS ON Q-SET ITEMS' 
 SET COMPRESSION=ON LENGTH=NONE WIDTH=80 
 FILE HANDLE DATAIN1 /NAME='TOTAL QDAT A' 
 DATA LIST FILE=DATAIN1 RECORDS=1 
 /1 BVK122 1 HRM102 2 TLC189 3 TIR137 4 SUB111 5 BGS999 6 
 COMPUTE TOTRATE=BVK122+HRM102+TLC189+TIR137+SUB111+BGS999 
 FLIP 
 CONDESCRIPTIVE VARIABLES=ALL 
 FINISH

IF you were to save this file by typing FILE, it would replace the older version of your reliability SPSS program--and we don't want to do that. So, instead of typing FILE on the command line, we will use the SAVE command as follows:
=====> SAVE CON SPSS A
In essence, you've created a new program file by modifying an old one, and giving it the name of your choice, CON SPSS A, by saving it with that name.

Now, the CON SPSS A file should appear if you type the FILELIST command or choose FILELIST from the Files Menu. You run it just like the reliability program, placing the cursor next to the name of the file in the FILELIST, and typing SPSS, followed by RETURN or ENTER. If the program reports errors, you need to find and fix them. If the run was error-free, hit the CLEAR key and the REFRESH function key, F2. You can look at the results by placing your cursor to the left of CON LISTING and typing the xedit key. As usual, the CON LISTING output will begin information you don't care about. what you want to see are the mean ratings for each item. So you may as well delete all the lines from the beginning up to where the means begin by typing DD anywhere in the prefix area of the first line:

=DD== 114-OCT-98 SPSS RELEASE 4.1 FOR IBM VM/CMS
                                                                                                       PAGE 1
===== 18:54:49 PENNSYLVANIA STATE UNIVERSITY IBM 3090-600S SPSS 4.1

===== 0FOR SPSS 4.1 PENNSYLVANIA STATE UNIVERSITY LICENSE NUMBER 20377
===== THIS SOFTWARE IS FUNCTIONAL THROUGH NOVEMBER 30, 1999.

===== TRY THE NEW SPSS RELEASE 4 FEATURES:
===== 0* LOGISTIC REGRESSION PROCEDURE          * CATEGORIES OPTION:
===== * EXAMINE PROCEDURE TO EXPLORE DATA       * CONJOINT ANALYSIS
===== * FLIP TO TRANSPOSE DATA FILES            * CORRESPONDENCE ANALYSIS
===== * MATRIX TRANSFORMATIONS LANGUAGE         * NEW LISREL AND PRELIS OPTION

the lower right corner of your screen will say: DD pending ... Now, use the FORWARD function key, F8, to page forward until you get to the item means. Type another DD in the prefix area of the line above these numbers, hit ENTER or RETURN, and the first part of the output will vanish.
(See below)

===== 7,272 BYTES REMAIN TO BE ACQUIRED.
===== -
===== 19-OCT-98 DESCRIPTIVE STATS ON Q-SET ITEMS                  PAGE 4
              |...+....1....+....2....+....3....+....4....+....5....+....6....+....7...
===== 14:25:18 PENNSYLVANIA STATE UNIVERSITY  IBM 3090-600S      SPSS 4.1
=====
=====
===== 0NUMBER OF VALID OBSERVATIONS (LISTWISE) =           9.00

=DD== 0

 

 

 

 

VALID

===== VARIABLE

MEAN

STD DEV

MINIMUM

MAXIMUM

   N LABEL

=====

 

 

 

 

 

===== CASE_LBL THIS IS A STRING (ALPHANUMERIC) VARIABLE.

===== VAR001

7.25

2.54

2.00

9.00

   9

===== VAR002

5.87

1.62

3.00

8.00

   9

===== VAR003

5.62

2.29

2.00

9.00

   9

===== VAR004

7.37

1.22

6.00

9.00

   9

===== VAR005

4.00

 .71

3.00

5.00

   9

===== VAR006

6.50

 .87

5.00

8.00

   9

 

 

 

 

 

 

Now, what we would like to be able to do is to sort the 100 Q-set items, labeled VAR001 to VAR100, from highest mean to lowest mean. This will allow us to quickly spot the most and least descriptive Q-set items. To do this, we need to use the COLEDIT command. You have just finished deleting all of the lines above the Q-set means. The next thing to do is to move the cursor in the prefix area next to VAR001, type a slash (/) in the area as follows

==/==  VAR001        7.25       2.54      2.00      9.00      9

hit ENTER or RETURN, and then issue the COLEDIT command from the
command line =====> COLEDIT

Your screen will change in that a list of column edit commands will appear at the top of the screen, and a row of dots will appear immediately above your Q-set item statistics:

CON      LISTING  C1  V 133  Trunc=133 Size=208 Line=96 Col=1 Alt=0
>> C=Copy, M=Move, F=Fol, P=Pre, I=Ins, D=Del, "=Dup, B=Blank, R=Rep, O=Ovl
>> U=Up, L=Low, AL=AlnL, AR=AlnR, V=Ver, H=Hex, ;=Tab, |=Trnc, <=SortA, >=SortD
***** ..........................................................................
...........................................................
      |...+....1....+....2....+....3....+....4....+....5....+....6....+....7...
=====  VAR001        7.25       2.54      2.00      9.00      9
=====  VAR002        5.87       1.62      3.00      8.00      9
=====  VAR003        5.62       2.29      2.00      9.00      9
=====  VAR004        7.37       1.22      6.00      9.00      9
=====  VAR005        4.00        .71      3.00      5.00      9
=====  VAR006        6.50        .87      5.00      8.00      9
=====  VAR007        7.50        .87      6.00      9.00      9
=====  VAR008        6.12        .60      5.00      7.00      9
=====  VAR009        4.37       1.32      3.00      6.00      9
=====  VAR010        4.75        .43      4.00      5.00      9
=====  VAR011        6.25        .97      5.00      7.00      9
=====  VAR012        5.62       1.11      4.00      7.00      9
=====  VAR013        3.25        .66      2.00      4.00      9
=====  VAR014        1.12        .33      1.00      2.00      9
  1=Help       2=AddLine    3=Quit       4=FileList   5=SChange    6=?
  7=Backward   8=Forward    9=PeekNext  10=Rgtleft   11=SplitJoin 12=Cursor
====>
                                                            COLED pending...

IT IS VERY IMPORTANT that all items and means are listed BELOW the scale line.
If this is not the case (e.g., if VAR001 is ABOVE the scale) go back to
regular XEDIT by hitting ENTER or RETURN twice. Then, move the cursor in
the prefix area next to VAR001, type a slash (/) in the area as follows

==/==  VAR001        7.25       2.54      2.00      9.00      9

hit ENTER or RETURN, and then issue the COLEDIT command from the command line =====> COLEDIT To sort the Q-set items from high to low (i.e., in descending order), place the cursor along the row of dots NEXT TO THE FIVE ASTERISKS until the cursor is directly above the column of means. Then type four greater-than signs (>) and hit ENTER or RETURN twice.

CON      LISTING  C1  V 133  Trunc=133 Size=208 Line=96 Col=1 Alt=0
>> C=Copy, M=Move, F=Fol, P=Pre, I=Ins, D=Del, "=Dup, B=Blank, R=Rep, O=Ovl
>> U=Up, L=Low, AL=AlnL, AR=AlnR, V=Ver, H=Hex, ;=Tab, |=Trnc, <=SortA, >=SortD
***** ...............>>>>.......................................................
...........................................................
      |...+....1....+....2....+....3....+....4....+....5....+....6....+....7...
=====  VAR001        7.25       2.54      2.00      9.00      9
=====  VAR002        5.87       1.62      3.00      8.00      9
=====  VAR003        5.62       2.29      2.00      9.00      9

The first time you hit ENTER, the rows of this file will be sorted from the Q-set item with the highest mean to the item with the lowest mean. The second time you hit ENTER, you return from COLEDIT to normal XEDIT. (Actually you can still execute normal XEDIT commands while doing COLEDIT, if you wish.) The result will look something like this:

|...+....1....+....2....+....3....+....4....+....5....+...
=====  VAR096        8.25        .83      7.00      9.00      9
=====  VAR077        8.00        .87      6.00      9.00      9
=====  VAR007        7.50        .87      6.00      9.00      9
=====  VAR004        7.37       1.22      6.00      9.00      9
=====  VAR098        7.37       1.49      4.00      9.00      9
=====  VAR001        7.25       2.54      2.00      9.00      9
=====  VAR095        7.12       1.69      4.00      9.00      9
=====  VAR093        7.00       1.22      5.00      9.00      9
 [. . . items in the middle deleted to save space . . .]
=====  VAR047        2.12        .93      1.00      3.00      9
=====  VAR021        1.87        .93      1.00      4.00      9
=====  VAR030        1.75       1.30      1.00      5.00      9
=====  VAR014        1.12        .33      1.00      2.00      9

At this point, you should save this file, CON LISTING A, by typing the FILE command at the command line: =====> FILE Then download this file, naming it con.listing, to either the hard disk or a floppy disk of your personal computer. Do the same for your RELQDAT LISTING A file, naming it relqdat.listing. You will be sending these two files as attachments to your email with the answers to the questions below.

Questions To Be Answered By Email

Send your answers to the following questions to me, j5j@psu.edu, by email. Attach the downloaded files, relqdat.listing and con.listing.

1. What is the alpha reliability for your set of ratings?

2. Is this level of reliability acceptable? Explain.

3. Who is the "best" rater in your group? The "worst" rater? (Use initials like WCB102. please, not names. Explain how you got these answers.

4. Prepare a composite personality description of Charley Reese by finding the 5 (or 6 or 7, if you like) MOST characteristic Q-set items and the 5 (or 6 or 7) LEAST characteristic Q-set items. Place the mean value for that item across all raters in parentheses. Your description should look something like this:

MOST CHARACTERISTIC
37. Is guileful and deceitful, manipulative, opportunistic. (8.50)
40. Is vulnerable to real or fancied threat, generally fearful. (8.50)
  7. Favors conservative values in a number of areas. (8.00) [etc.]

5. Make a few comments about the kinds of things you think Reese said in his editorials that led the raters to the personality impressions they formed.