Each time, a test evaluates only one or a few more characteristics of an individual. Thus, interpreting final results should always be done in conjunction with the person’s history, nature, accomplishments, etc.
What makes a test or assessment good?
The test is objective
A good test s always objective and replicable. It is free from subjective elements and leaves no room for ambiguity of interpretation. A test’s items and scoring system, both need to be objective. If items are not objective, different people will answer or respond to the items differently. If the scoring system is not objective, different people can interpret or score the results differently.
Even projective tests need to have an objective scoring method. Even they need to have items that are consistent. When uniformity of items and scoring methods are maintained, we can expect results that are reliable and valid.
The test is reliable
When we say a test is reliable, all we are saying is that it can self-correlate. The test results should be consistent when it is administered more than once on the same sample after a reasonable period. Internal consistency is the consistency in performance in a single administration. Temporal consistency refers to retesting and achieving similar results.
In other words, if a man is tested for personality traits, he should score the same result or similar results even after a long duration of the same test. This makes the test reliable, and free from errors of measurement.
There are norms
A perfect test comes with certain criteria, related to age, grades, percentile and standard scores. It must be understood what standardization is, and what samples and populations are.
Population: A population segment is the target audience for a test. For instance, the target audience for a coding test is IT professionals. All IT professionals in the world make up the population for this particular test.
Sample: As the whole IT population cannot be used to standardize the test, we choose a representative sample of these IT professionals and call them ‘The Sample.’ This sample is drawn randomly, to avoid any bias or prejudice. When a test is standardized on a randomly drawn test sample, norms can be calculated and used for future scoring of tests.
How is a test constructed?
Write the items of the test
The second step in constructing a test is preparing a list of items that need to be included in the test. If the assessment involves a coding test, all questions that will be posed need to be listed. Each question is an item. An objective test will have objective items. Some of these are multiple-choice items, completion items, pictorial items, matching items, alternative response items, etc.
To write good items, many requirements are to be followed. Some of them are:
- The item writer must be a subject matter specialist and must have a thorough knowledge about the test items he is writing for.
- He must understand the sample and population very well. He must design the items according to the intelligence level of each test taker.
- He should know what makes a good item and what doesn’t.
- The item writer must have a vast vocabulary of simple yet professional words.
- Once the items have been listed, they need to be submitted to other experts for criticisms and suggestions.
Preliminary administration of the test
When the final list of items is prepared, the next step is to try them out, aka, ‘Experimental Try-Out’.
The first try-out is called the pre-try-out, and is usually done on 100 people representing the population.
The second try-out should have around 400 samples from the population, and these results are used in item analysis. Item analysis helps to ascertain the difficulty of the item, the discrimination index (or how different each item is), if the items are effective at testing what they intend to test, etc.
The third try-out is done to rule out any errors that were not picked up by the initial try-outs.
HR assessment tools need to follow test construction protocols
HR assessments and testing tools are increasingly being published online. As more companies realize the importance of testing online, it becomes necessary for businesses to get custom online testing kits to assess their applicants.
If all the methods that are listed above are followed, there is no reason why an online HR assessment tool is less useful than more expensive and lengthy printed testing methods. Moreover, online tests rate very high on practicability, which is a sign of a real test.
Online HR tests also score high reliability and validity coefficients, making them a lot more scientific than printed tests, which are prone to scoring and interpretation errors, lengthy administration time and other such issues.
Contact us for more information, and we can then have a chit chat on how our online assessment system can do wonders for you!