Item Construction

1. be short

2. be unambiguous

3. be definite, avoiding generalities that may create one type of ambiguity

4. be simple, containing one proposition, not several combined

5. avoid negatives, especially double negatives

6. use simple vocabulary, and

7. use conventional vocabulary

1. Multiple negatives: “Obtaining a statistically non-signisficant result does not indirectly imply that the effect detected is unimportant”

2. Double-barreled items (i.e., multiple points contained within the same item): “I support civil rights because discrimination is a crime against God.” “I enjoy educational psychology because it is relevant to teaching.”

3. Ambiguous pronoun references: “Murders and rapists should not seek pardons from politicians because they are the scum of the earth.”

4. Misplaced modifiers: “Our representatives should work diligently to legal- ize prostitution in the House of Representatives.”

5. Adjective forms instead of noun forms: “All teachers should attend a boredom class.” “All teachers should attend a boring class.”

6. Negatively-worded items: Introduce potential confusion. Often results in two-factorial solutions (positively-worded items on one factor, negatively- worded items on the other factor). See Weijters and Baumgartner (2012) for an overview. Weijters, B., & Baumgartner, H. (2012). Misresponse to reversed and negated items in surveys: A review. Journal of Marketing Research, 49(5), 737-747.

7. Absolute statements (e.g., “always”, “everything”)

8. Vague quantifiers (e.g., “sometimes”, “often”, “rarely”) because they are

   prone to idiosyncratic interpretations.

• Please read each statement carefully and respond how similar to you the statement is on the following scale from 1 to 6. 1=Very Different, 2=Different, 3=Somewhat Different, 4=Somewhat Like, 5=Like, 6=A Lot Like. A sample item was: When I have free time at home, I read for fun.

• On a typical day, how much time do you spend reading the following materials? Use this as your guide: A=Never, B=1-15 min, C=16-30 min, D=31-45 min, E=46-60 min, F=1-2 hours, G=3 hours+. Items included Lecture notes, Newspaper/Magazine, etc.

• How often do you use the following religious media? a) Every Week, b) Every Month, c) Occasionally, d) Never. A sample item was: Watching religious television.

=> you should not just limit yourself to Strongly Agree, Strongly Disagree

• Notice that in the last directions, I did not label any numbers except the first (Not at all) and the last (Extremely). However, participants had seven different numbers they could circle. Not all numbers in the response scale must have a label assigned to them“.

• Bandalos & Enders (1996): Reliability was highest for scales having a 5 to 7 points.

• Preston & Colman (2000): Examined response categories ranging from 2 to 11. They found test-retest validity was lowest for 2 to 4 point scales, was highest for 7 to 10 point scales, and decreased for scales with more than 10 response categories.

• Research by Lozano, Garcia-Cueto, and Muniz (2008) investigated reliability and validity of scales ranging from 2 to 9 response options with four different sample sizes, and found that the optimum number of statistically favored response scales were between 4 and 7 options.

I would therefore suggest using 4 or 6 response options.

=> even number

Why should we choose an even number of response categories?

Because there is evidence that, when using Likert scales with an odd number of response options, participants’ interpretations of the midpoint can vary widely with the most common responses being: “no opinion,” “don’t care,” “unsure,” “neutral,” “equal/both,” and “neither” (e.g., Kulas & Stachowski, 2009). Thus, the middle category is a point of indifference (“does not apply”) than the midpoint of the response scale

Items with an odd number of response options furthermore exhibit a high response latency, an “it depends” connotation, and a strong, negative relationship with item clarity (Nadler, Weston & Voyles, 2015).