The paper will specifically focus on composing items and response scales. The first part of the paper will determine main issues which the researcher must be able to address in the item construction. The second part will discuss types of response scales that are commonly used in measuring a construct and offer a brief example. The third part will indicate one of the most appropriate response scales that the instrument could use with justification. Next, three sample items will be constructed which represents the construct of choice and response scale. Lastly, the paper will use Rauthmann’s proposed item format taxonomy to evaluate the item scales created.
Issues to be focused on during item construction
In the process of item construction, there are major issues that must be taken into consideration. First, it is important to use a simple language where every participant will be able to understand the message quickly. Internal jargon should be avoided at all costs. It is true that various corporations have their abbreviations and acronyms for services and products that might not be familiar to consumers. Employees in most instances are more likely to respond to particular survey item if the questions are precise, straightforward and easy to answer besides being relevant to them (Fishman & Galguera, 2003).
Delegate your assignment to our experts and they will do the rest.
There is the need to ensure that there is a consistent response format. Each item in the survey section ought to use a similar rating scale. The participants might get confused if the answers change from a five point to a seven point scale. Therefore, the scales should be kept within the same direction (Trochim, 2006). For instance, if a rating of 'Strongly agree' is utilized as a response alternative in the first item, then it ought to be employed on all the subsequent items. It is imperative to ensure that the anchor for responses is consistent across the survey questionnaire.
Consistent wording should be employed. In this case, the researcher will have to use similar phrases for survey text (Zeigler & Hagemann, 2015). For instance, the questions should be set up with a particular lead sentence which is the phrase that will be used to lead off each question. The use of closed rather than the open-ended questions are highly encouraged whenever possible and ensure that the context is similar for all the participants. However, the pattern for the closed-ended questions is significantly dependent on the answer set presented. The inclusion of ‘other’ might not compensate for the omission of an essential answer, and in case the unimportant answer is included, it is vital to be overestimated.
The following are the major things that should be avoided when one is constructing items. First, leading questions should be prevented. One should avoid influencing the survey volunteers with a leading question. Biased questions might lead to the incorrect response, and the volunteers might be turned off by the leading question (Fishman & Galguera, 2003). One should avoid using more than one question at once also referred to as double barreled questions, for instance, "Sales reps are polite and responsive." This might force the respondent to be tempted to rate a particular attribute in a different way from his or her real feeling. Lengthy questions have to be avoided; however, it might be justified. Use of the word "If" should be avoided because it might alert one to potential confusing direction. Lastly, challenging and unfamiliar words should not be used because few participants are familiar with such words. Rather, the question should be specific and in a simple language.
Types of response scales to measuring a construct
Response scales is simply a particular composite measure that is entailed of various items which have a precise empirical and logical structure. Scales often uses the difference in intensity among the indicators of the variable identified. It is clear that there are several forms of response scales including:
Likert Scale- is commonly used the scale in the field of social science research developed by Rensis Likert. It offers a simpler rating system which is common to all forms of surveys. A common use of this response scale is the survey which asks the participants to express their personal opinions about something through stating the particular level of which they seem to either agree or disagree (Kline, 2005). For an individual to develop a suitable Likert scale the choices will be assigned an appropriate score for instance between 0-4 then answers for the items are included for each to achieve the overall Likert scale. For example, when measuring prejudice against women, the following statement will be used "Women should not be allowed to vote,” and then assign each of the listed response from strongly agree to strongly disagree on the score of between 0 and 4 strongly.
Bogardus Social Distance Scale- it was developed by Emory S. Bogardus as a tool to measure a person’s eagerness to participate in the social relationship with various forms of individuals. This scale, therefore, invites participants to show the degree at which they are accepting other groups. There is always a clear difference in intensity which suggests a structure among the items. If the participant seems to acknowledge a particular level of association, then it follows that he will significantly be more than willing also to accept the rest of the items on the list, but this is not always the case (Kline, 2005). When the rating for the responses is averaged, the lower score will indicate significant acceptance level compared to that of higher scores. For instance, when one is interested in the extent to which the US’s Christians are willing to be associated with Muslims, the questions below will be used:
Are you willing to live in with Muslims in the same country?
Are you ready to share the neighborhood with Muslims?
Are you willing to live with a Muslim next door?
Thurstone Scale- it was developed by Louis Thurstone as an alternative response scale. It is used in preparing a format meant to generate a group of indicators on a particular variable which is believed to have an empirical structure. For instance when one is researching about the concept, discrimination, and one would create a list of approximately ten items then request the respondents to assign a correct score to every listed item (Zeigler & Hagemann, 2015). The respondent will rank these items from the weakest indicator towards the strongest indicator. After the participants have completed scoring the items, the researcher will examine the score provided to every item to identify several items that the subjects seem to agree upon. In a situation where the items in the scale seem sufficiently developed and scored, it would follow that the efficacy of available data reduction would emerge.
Appropriate response scale that the instrument could use
According to Dulewicz, Higgs & Slaski, (2003), BarOn EQ-i self-report is this considered as a premier scientific measure used when self-reporting the emotional intelligence. The method specifically covers the core aspects of the emotional intelligence as related to both the verbal and non-verbal perception in addition to the emotion’s expression in self and others. In addition to this, BarOn EQ-i has the potential to measure an individual’s intrapersonal skills, stress management, adaptability and interpersonal skills.
BarOn EQ-i assessment is useful when the company wishes to focus on the employee’s initiative for development and even in facilitating selection and recruitment of new staff. Studies have shown that there exists strong link between employee job performance and emotional intelligence (Dulewicz, Higgs & Slaski, 2003). As a result of this, BarOn EQ-i assessment is considered as the most suitable tool for screening to help the human resource manager to select successful candidates. The manager can quickly develop a profile of the organization’s top performers with the aim of determining specific skills that are considered valuable to the firm and a specific job function.
Sample items that represent the construct of choice
According to BarOn, the concept of Emotional Intelligence can be understood as a range of an individual’s skills, noncognitive capabilities and even competencies which influences their ability to be successful in adjusting and coping with different environmental pressures (Dulewicz, Higgs & Slaski, 2003). Emotional intelligence, in this case, will be measured using BarOn EQ-i.
I have a good sense of why I have certain feelings at times
Strongly Disagree Strongly agree
| 1 | 2 | 3 | 4 | 5 | 6 | 7 |
I have a good a level of understanding of my emotions
Strongly Disagree Strongly agree
| 1 | 2 | 3 | 4 | 5 | 6 | 7 |
I often know my friend’s emotions based on their behaviour
Strongly Disagree Strongly agree
| 1 | 2 | 3 | 4 | 5 | 6 | 7 |
Use Rauthmann’s proposed item format taxonomy to analyze your partially-constructed three-item scale (i.e., point of reference, general item format, constructs indicator and conditionality).
Indeed none of the partially-constructed three-item scales achieve a widespread acceptance because they differ from one to the other (Rauthmann 2011). They are based on the underlying situational concepts, theoretical problems and meta-theories which are which cannot comparatively bet integrated or compared. This is because they have addressed different types of situational data. As evidently observed in the first partially-constructed three-item scale, from the three listed taxonomies, one can be said to pertain to feelings, other two pertain to the situational characteristics, and none pertains to situational cues. This is in contrary to most taxonomy since they capture situational characteristics secondarily with regard concretizing situation classes (Fishman & Galguera 2003). However, the underlying taxonomic features tie together various situations which are merely inferred from the fact that such situations are the same (Dulewicz, Higgs &, Slaski 2003). The partially-constructed three-item scales are problematic because they have not established on which dimensions the perceived person’s situations are similar. The participants seemed not to have rated any characteristics of the situations.
staticity
Awareness of others people's feelings
It is always good to raise the special topic of classification very critically. An item like “laughing with people” needs to be classified as an item of self-awareness to others (Fishman, J. A., & Galguera 2003). This is a valency approach with a behavioral indicator. This is in an unconditional format as seen above in the table. Nonetheless, based on the strict adherence of the rules, this item needs to have been classified as an item of staticity approach having a mental indicator as seen above in the texts. It is characterized by the behavioral specification of the mental indicator such as talking with regards to the unconditional format (Fishman, J. A., & Galguera 2003). Even though it is also correct and has an advantage of dissolving valency approaches, it helps in reducing the complexities that are associated with item format taxonomy.
References
Dulewicz, V., Higgs, M.&, Slaski, M. (2003). Measuring emotional intelligence: content, construct and criterion-related validity. Journal of Managerial Psychology. 18(5), p405. 16p. D.o.i: 10.1108/02683940310484
Fishman, J. A., & Galguera, T. (2003). Introduction to test construction in the social and behavioral sciences: A practical guide. Lanham, MD: Rowman & Littlefield. Read Chapter B2, and Review Chapters A5, A6, and B1
Kline, T. (2005). Psychological testing: A practical approach to design and evaluation. Thousand Oaks, Calif: Sage Publications. Review Chapters 2 and 3
Rauthmann, J. F. (2011). Not only item context but also item format is important: Taxonomizing item format approaches.
Trochim, W. (2006). The research methods knowledge base. Read Survey Research and Scaling sections.
Zeigler, M., & Hagemann, D. (2015). Testing the unidimensionality of items: Pitfalls and loopholes. Retrieved from: http://dx.doi.org.proxy1.ncu.edu/10.1027/10155759/a000309
