Research Methods in Criminal Justice - Essay Example

The researcher relies on their intuition that what they are measuring is exactly what is required to be measured through a questionnaire.

Content validity emphasizes that each item can determine what it is expected to measure. The researcher must support with an explanation of each item on the questionnaire and what it is expected to measure.

Face validity can take items by face value and exactly assess the face value. Some items can have lower face validity than what it appears to have. It involves the assessment of normal people in the field. Content validity involves the assessment of experts in the field, but it is non-empirical and judgmental (Adler & Clark, 2014).

Reliability is measured by test-retest which is obtained by running the same test twice on the same test subject in a given period. The results from both tests are evaluated for the test of stability over that period. For example, a test done for a particular crime can be retested in the occurrence of another similar crime. The correlation will give a stability score.

Construct validity is an assessment based on variables that are related. Construct validity carries weight when proof or evidence is provided. Interrater reliability is used where results for one observer are strategically used in comparison with another using the same method. Interrater reliability is used when two or more measures are not reliable. They can compare correlatively those results using an interrater (Maxifield & Babbie, 2012).

A scale is defined as a measure of responding to a question, a compilation of statements or questions used to respond to a concept studied. There are three types of attitudinal scaling used in criminal justice.

The Thurstone scale uses paired comparisons and equal appearing intervals methods to scale. Judges are given several pairs of questions and choose which would most fit the concept of study. The one picked most becomes part of the questionnaire. The judges are required to give a number scale on the strength of each question. The researcher then takes the questions that showed the strongest agreement from the judges. The weak questions are eliminated when selecting the statements to use (Wainer & Braun, 2013).

Likert scale is a five to seven - point number scale from strongly agree to strongly disagreed.  A total score is taken of all items in the scale. Those items that fall in between are eliminated. The middle number is an undecided choice. Practically, numbers are determined according to personal taste or conviction. In the Guttman scale, the cumulativeness of the scale is not perfect because an agreement with one item will respond to an agreement with another item (Bayens & Roberson, 2010).

The Likert scale would be the most appropriate scale to use. Seven points scale will give interval data with minimal error and higher precision. The distances produced on the seven-point scale will not be equal but symmetrical.

Likert scale is the summation of a five or seven-point scale, with seven giving the best results. Choices range from the least agreeable to strongly agreeable. All answers to questions are required to be given by the respondents. The researcher sums up the total items.

The analysis is done on the final results. Choices need to be numbered as they will give the researcher options on how to sort the results out. For example, in numbers 1 to 7. 1 strongly disagreed, and seven strongly agree. Number 4 is undecided (Dantzker & Hunter, 2012).

Validity and reliability may seem much the same, but they are different. The validity of data measurements through questions brings out its effectiveness. Reliability brings out the accuracy of measured data, and the same results of the data re-measured. In the questionnaire, the inclusion of scaling types like Thurstone, Likert, and Guttman is important. Likert is the most popular method used for the measurement and assessment of data.