Instrument Scoring

HOME > SETUP > DEVELOPER'S INSTRUMENT GUIDE > HOW TO CODE AN INSTRUMENT > SCORING

---

1. Scoring Algorithm Script
2. Scoring Function
3. Reliability Scoring

---

1) Scoring Algorithm Script

If an instrument form should include a scoring algorithm, a scoring script can be coded manually to accompany the form. This scoring script will be executed by Loris automatically when a user saves data entered in the instrument form via their browser. The script should be stored as an executable file in the project/instruments/ directory, and the filename must be called Instrument_name.score.

It can be coded in any scripting language, but we suggest using our PHP example below. CommentID is passed as the first argument, after which it selects from Instrument_table, performs calculations and updates using the CommentID. To test your scoring script, run it from the tools directory and provide CommentID. Below is a sample scoring script for BMI calculation - this file can be copied from the docs/instruments/ directory:

<?php
/* Test_name Scoring
* Description of the scoring algorithm
* @category Instrument
* @package  Test_name
* @author   Author
* @license  Loris License */
require_once "../tools/generic_includes.php";
require_once "Database.class.inc";
$CommentID = $argv[1];
$db =& Database::singleton();
$query = "SELECT * from 'test_name' WHERE CommentID = :CommentID";
$WhereCriteria = array('CommentID'=>$CommentID);
$result        = $db->pselectRow($query, $WhereCriteria);
$scores = array();

//check unit of measurement
if ($result['unit_of_measurement'] == 'standard') {
    $query_std = "SELECT bmi_value FROM bmi_standard WHERE height_feet =:hgt_feet AND height_inches=:hgt_inches AND weight=:wgt_pounds";
    $Where     =  array('hgt_feet'=>$result['height_feet'], 'hgt_inches'=>$result['height_inches'],
'wgt_pounds'=>$result['weight_pound']);

$scores['bmi_value'] = $db->pselectOne($query_std, $Where);   

} else if ($result['unit_of_measurement'] == metric) {$query_metric = "SELECT bmi_value FROM bmi_metric WHERE height_cms=:hgt_cms'' AND weight_kgs=:wgt_kgs";
$Where = array('hgt_cms'=>$result['height_cms'], 'hgt_kgs'=>$result['weight_kgs']);
 $scores['bmi_value'] = $db->pselectOne($query_metric, $Where);}

if ($bmi_value <= 18.5) { $scores['bmi_category'] = 'Underweight';} 
else if ($bmi_value > 18.5 && $bmi_value <= 24.9 ) {$scores['bmi_category'] = 'Normal weight';} 
else if ($bmi_value >= 25 && $bmi_value <= 29.9) {$scores['bmi_category’] = 'Overweight';} 
else if ($bmi_value >= 30) {$scores['bmi_category'] = 'Obesity';}

//save scores
$result = $db->update('test_name', $scores, $WhereCriteria);

2) Scoring Function

If your instrument was coded manually in PHP, then your scoring algorithm can be implemented within your instrument’s PHP file as a function. The function must be named score. Below is an example of what a scoring function typically looks like:

/**
* Example scoring function
*
* @return void
*/
function score() {
    if ($this->_determineDataEntryCompletionStatus() == "Incomplete") {
        return;
    }
        
    $db =& Database::singleton();
    //Get raw questions point values
    $query = "SELECT * FROM " . $this->table .
        " WHERE CommentID='" . $this->getCommentID() . "'";
    $record = $db->pselectRow($query, null);
        
    $score = array(
        'score_1' => null,
    );
        
    $score['score_1'] = true;
    if ($record['abc_1'] <> 'yes') {
        $score['score_1'] = false;
    } 

    // save the scores
    $db->update(
        $this->table,
        $score,
        array('CommentID'=>$this->getCommentID())
    );
}

3) Reliability Scoring

Under construction.

To include instruments in the Reliability module, you must enable ReliabilityInstruments in the config.xml in your project directory.

<!-- Instruments for reliability module -->
<ReliabilityInstruments>
    <Instrument>
        <Testname>bmi</Testname>
        <Threshold>0.5</Threshold>
        <Displayname>BMI</Displayname>
    </Instrument>
</ReliabilityInstruments>

To enable reliability scoring, you must create a reliability instrument called NDB_Reliability_$TESTNAME$_reliability.class.inc in your project/reliability/ directory.