Research Paper on Major Issues to Address in Item Construction

Introduction

Consideration for unidimensionality is important during item construction to ensure that systematic differences within its variance emerge of one latent variable. The fulfillment of this requirement mandates alignment to the local independence principle. It involves ensuring unidimensionality when correlated residuals are non-existent when one controls for latent construct. Secondly, item construction should reflect the intended measure derived using rational development method (Ziegler & Hagemann, 2015). Practically, the items from a similar scale should capture differences that were underlying the same construct.

Is your time best spent reading someone else’s essay? Get a 100% original essay FROM A CERTIFIED WRITER!

Save 25%

Attaining unidimensionality prioritizes a general position and gradually working into specific hypotheses. The psychological measures were chosen to derive data capable of providing information to the hypotheses. Again, one should guarantee that the test scores obtained from the psychological measures directly link to the constructs captured in the formulated hypotheses (Ziegler & Hagemann, 2015). Doing so would allow the items to the test to reflect differences that match the specific construct and not others. Addressing unidimensionality when constructing the items that constitute the test score yields soundness to the assessment process it is applied (Ziegler & Hagemann, 2015). The unidimensionality tests eliminate risk potential within the score.

The unidimensionality should extend beyond actual construct, positioning it within the nomological net to ensure such items belong to the proposed measure. Its fulfillment requires trying to prevent the items from loading with other traits. One should pay attention to constructing process to avoid loading the items with other traits. Failure to do so would leave constructs in woven nomological net since the existence of multiple overlaps erodes unidimensionality (Ziegler & Hagemann, 2015). It is beneficial to position the items formulated within a nomological net derived from the exact construct one seek to measure.

General Types of Response Scales Used In Measuring a Construct

Scales accommodates multiple options for responses. Their development acknowledges order categories in the responses realized within a population. Likert scaling accommodates continuum-based responses categorized into five categories including strongly approve, approve, undecided, disapprove and strongly disapprove (Trochim & Donnelly, 2006). Its variation accommodates scaling anchored on strongly disagree, disagree, undecided, agree and strongly agree. Variations allow the researcher expand the descriptors beyond the approval though retaining the equal intervals. Intensity Likert-type scale captures the degree of the information on agreement, satisfaction, and liking (Kline, 2005). Such may capture strongly agree, disagree, slightly agree, undecided and strongly disagree. Frequency Likert-type scales capture how often one may observe and experience behavior or symptom (Trochim & Donnelly, 2006). It sometimes reflects, frequently, seldom and hardly.

Visual analog scales accommodates variants derived through multi-categorical method to the responses. Their use requires the responded to mark on the line their score read from one endpoint. Pictorial scaling allows the participation of response lacking strong verbal skills including children, adults without language proficiency and lower literacy (Trochim & Donnelly, 2006). It requires the respondent to select facial expressions that closely resembles one's affective reaction and attitudes. Adjective rating scale anchors the responses using presumed opposites, including selfish versus selfless and happy against sad. The use of behaviorally-anchored rating scales helps assess work performance setting (McCrae & Allik, 2014). Normative scaling allows one to capture responses on attitude, interest, ability and personality measure. Its use assesses the normal distribution of scores within the population. Ipsative scale captures dependent responses where knowledge of the item influences the response. Such arises where choosing an option qualifies the reversed scale (Trochim & Donnelly, 2006).

Appropriate Response Scale

The application of summated-rating scale suits the instrument since it allows multiple options in the continuous responses. Its use acknowledges the applicability of ordered categories by inferring higher values to imply intensity of the information provided. As such, using a Likert scale would accommodate the attitude statements captured in five or seven response categories (Monette, Sullivan, & DeJong, 2011). Its use eliminates dependency on the labor-intensive scale scores in each construct item. Besides, it enables one to obtain assessment based upon simple summing derived upon the response categories.

The knowledge of the valence in item construct makes it possible to use reversely coded responses transposed to reflect the item. Its adoption in the instrument would accommodate variations in the categorized descriptors on a five-point scale on the agreement. The response scale makes it viable for a construct that captures negative valence since respondents would pay attention (Trochim & Donnelly, 2006). I consider that important since the inclusion of the negative valence prevents the respondents from selecting categories without attending to the respective items. It suits the inclusion of an evenly distributed negatively and positively worded categories to avoid skewed responses and interpretation.

Sample Items

1. The appreciation of members' efforts improves overall team performance.

2. Involvement of members in setting goals draws their efforts to accomplish the team purpose.

3. Higher levels of vigilance in member's roles deter laxity in team performance.

The five-point Likert scale suits the assessment of each item, to sum up the strength of one's response across the categories. Its use would indicate the agreement level anchored to the descriptor closely matches the respondent attitude and judgment. For that reason, the five-point Likert scale will accommodate reverse coding, hence capture negative and positive valence (Domino & Domino, 2006). It would involve anchored scale from 1 to 5 representing strongly disagree, disagree, undecided, agree and strongly agree respectively. The inclusion of the midpoint - undecided, will accommodate views of individuals indifferent to each item. However, the inclusion of negative and positive response categories attempts to avoid skewed scaling.

Rauthmann's Item Format Taxonomy to Analyze Partially-Constructed Three-Item Scale

Rauthmann argues a four-way interaction comprising reference point, general item format, construct indicator and conditionality to as important to ensure optimal capture of the construct. The taxonomy format argues that item formats often overlook wording, grammar, and syntax (Rauthmann, 2011). The format reveals the use of static sentences and descriptions that extend on behaviors frequency and valence on one's feelings. The items captured above utilizes possessive blend with valence descriptions within the five-point Likert scale that range from strongly disagree to strongly agree. The scale reveals descriptors that concern the valence from the respondents' feelings and attitudes to the static sentences. For example, "I disagree that involvement of members in goals settings..."

Again, the item content reveals construct-relevant indicators attributing participation, appreciation, and vigilance to avoid laxity in the team performance. The item content certifies conditional and unconditional element to ensure contextual specifications that team performance behaviors would occur. In particular, the Likert five-point scale shows the attitude towards involvement and high vigilance to trigger improved team performance. The inclusion of construct indicator affirms behavioral and contextual attributes captured under-involvement, appreciation, and vigilance to deter laxity. The evaluation of sample item constructs satisfied Rauthmann taxonomy format though captured within the Likert scale response categories.

References

Domino, G., & Domino, M. L. (2006). Psychological Testing: An Introduction. Cambridge: Cambridge University Press.

Kline, T. (2005). Psychological testing: A practical approach to design and evaluation. Thousand Oaks, Calif: Sage Publications.McCrae, R. R., & Allik, J. (2014). The Five-Factor Model of Personality Across Cultures. New York: Plenum Publishers.

Monette, D. R., Sullivan, T. J., & DeJong, C. R. (2011). Applied Social Research: A Tool for the Human Services. Belmont, CA: Brookscole.

Rauthmann, J. F. (2011). Not Only Item Content but also Item Format is Important: Taxonomizing Item Format Approaches. Social Behavior and Personality, 39(1), 119-128.

Trochim, W. M., & Donnelly, J. P. (2006). The Research Methods Knowledge Base. Mason: Cengage Learning.

Ziegler, M., & Hagemann, D. (2015). Testing the Unidimensionality of Items Pitfalls and Loopholes. European Journal of Psychological Assessment, 31(4), 231-237.