You have been asked to develop a set of questions based on a class you took in the last term. Prepare a test containing at least twelve questions with at least t

Responsible Test Administration.html

Responsible Test Administration

Administering a test brings with it many responsibilities. Whether it is being given in an individual or group setting, several duties need to be fulfilled in order for a test to be given in a completely ethical manner.

The test administrator needs to ensure that all procedures and protocol are being followed both by the individuals taking the test as well as the administrator. It is critical to follow the allotted testing time and to make sure that the environment in which the test is being given suits all test-takers and is quiet and comfortable. After collecting all tests, answer sheets and test materials must all be stored securely and individuals' confidentiality honored.

Creating Test Items.html

Creating Test Items

Before beginning to design any assessment tool, it is critical to have clearly defined the construct in measurable detail. After operationalizing this construct, the next step is to actually create the items to be included in the test.

There are three broad approaches to designing items:

Empirically generated items are chosen because they have been found to accurately demonstrate or measure the construct, even if the test maker does not understand why they work.

Theory-driven approaches require a thorough review of literature surrounding the construct to be measured. Items are specifically developed to match with the prevailing theory about the construct.

A combination of the two previous approaches is the rationally developed approach. Although this approach often begins with an idea for a construct which is deeply rooted in a theory, the research done to gain empirical support renders it neither purely empirically- or theoretically-driven.

One way to measure attitudes is to use paired comparisons. Individuals are presented pairs that they must choose from. It is often a forced-choice, meaning that they have to choose one or the other, even if they want to choose both or neither.

Ranking items along some dimension is another way to measure individuals' attitudes. The person is asked to place items in order based on their preference or judgment.

Using items in interval-level categories, a successive intervals approach allows us to see the true interval level of each item as well as the interval-level boundaries of the response categories.

A slightly different approach is the behavioral assessment technique. A common type of scale used in the behavioral approach is the behavioral observation scale, which is commonly seen in the workplace to evaluate employees.

Analysis & Scoring.html

Analysis & Scoring

Open-ended responses can be difficult and time-consuming to assess because the depth of the information attained can be very great, and oftentimes does not fit neatly into one category for analysis.

Closed-ended questions require participants' responses to fit into a category decided upon by the creator of the test, and allow for analysis to be a much more straightforward process. A benefit of using closed-ended items is that they can easily be scored and statistically analyzed in order to draw conclusions from the data.

Frequency distributions, chi-square statistics, cross-tabulations and other statistical analyses are used to display and interpret information collected from the item responses.

Continuous responses are a way to avoid the dichotomous testing items, which limit individuals to two primary options. Summated-rating scales, likert scales, and visual scales are often used in order to give individuals more options, and, ideally, to obtain the responses that most accurately express an individual's actual opinions and feelings.

When scoring tests, it is important to keep in mind the different types of scoring that will be used depending on whether the format is open-ended or closed-ended. Also, considering whether the assessment would be most effective if attitudinal vs. behavioral will allow for accurate interpretation of results. Deciding which scale will allow for the most accurate responses is another vital aspect of scoring.

Finally, reviewing the types of issues and errors that are commonly made when using the measures you have chosen is an important step in finalizing a test.