This specifies what part of the data record is to be used to classify persons for subtotal in Table 28.

 

 

With tab-separated data and the subtotal indicator in a separate field from the Person Iabel, specify the subtotal field as the person label field using NAME1=, then PSUBTOTAL=$S1W1

 

Format 1: PSUBTOTAL = $S..W.. or $S..E.. using the column selection rules.

$S..W.. e.g., $S2W13 means that the label to be shown on the map starts in column 2 of the person label and is 13 columns wide.

 

$S..E.. e.g., $S3E6 means that the label to be shown on the map starts in column 3 of the person label and ends in column 6.

 

These can be combined, and constants introduced, e.g,

PSUBTOTAL=$S3W2+"/"+$S7W2

 

If the person label is "KH323MXTR", the subgrouping will be shown as "32/XT"

 

Format 2: PSUBTOTAL=*

This is followed by a list of subgroupings, each on a new line:

 

PSUBTOTAL=*

$S1W1+$S7W2  ; Subtotals reported for person classifications according to these columns

$S3E5   ; Subtotals reported for person classifications according to these columns

*

 

Example: Subtotal by first letter of person name:

 

PSUBTOTAL=$S1W1

TFILE=*

27 ; produce the subtotal report

*

 

Here is a subtotal report (Table 28) for person beginning with "R"

 

"R" SUBTOTAL FOR 8 NON-EXTREME PUPILS

+-----------------------------------------------------------------------------+

|           RAW                          MODEL         INFIT        OUTFIT    |

|          SCORE     COUNT     MEASURE   ERROR      MNSQ   ZSTD   MNSQ   ZSTD |

|-----------------------------------------------------------------------------|

| MEAN      28.1      25.0        4.04    3.48       .91    -.5   1.04     .0 |

| S.D.       5.5        .0        6.63     .14       .31    1.1    .54    1.4 |

| MAX.      38.0      25.0       16.30    3.82      1.61    2.0   2.37    3.4 |

| MIN.      19.0      25.0       -6.69    3.38       .64   -1.6    .60   -1.2 |

|-----------------------------------------------------------------------------|

| REAL RMSE   3.63 TRUE SD    5.54  SEPARATION  1.52  PUPIL  RELIABILITY  .70 |

|MODEL RMSE   3.48 TRUE SD    5.64  SEPARATION  1.62  PUPIL  RELIABILITY  .72 |

| S.E. OF PUPIL MEAN = 2.50                                                   |

| WITH 2 EXTREME = TOTAL 10 PUPILS  MEAN = 3.05,  S.D. = 28.19                |

| REAL RMSE   8.88 TRUE SD   26.75  SEPARATION  3.01  PUPIL  RELIABILITY  .90 |

|MODEL RMSE   8.83 TRUE SD   26.77  SEPARATION  3.03  PUPIL  RELIABILITY  .90 |

| S.E. OF PUPIL MEAN = 9.40                                                   |

+-----------------------------------------------------------------------------+

  MAXIMUM EXTREME SCORE: 1 PUPILS

  MINIMUM EXTREME SCORE: 1 PUPILS

      LACKING RESPONSES: 1 PUPILS

                DELETED: 1 PUPILS

 

Example: You want an average score for each option on a multiple-choice test.

 

1. Include the item response string in the person label

 

Data record

Fred ABCDE

 

NI = 5

Item1=6

Name1=1

Namelength=10

 

2. Produce subtotals for every item

 

PSUBTOT=*

6  ; corresponds to Item 1

7

8

9

10

...

*

 

3. Specify Table 28

 

TFILE=*

28

*

 

4. Table 28 output is voluminous, but the numbers are there. Do a search for "| MEAN".