Scales and Measurement

Please download to get full document.

View again

of 5
17 views
PDF
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Types of Scales and Measures Thurstone scales. The Thurstone scale is made up of statements about a particular issue and each statement has a numerical value indicating the repsondent’s attitude about the issue, either favorable or unfavorable. People indicate which of the statements with which they agree and the average response is computed. First, you must be very clear about exactly what it is you’re trying to measure. Then, collect statements on the topic ranging from attitudes that are favo
Document Share
Documents Related
Document Tags
Document Transcript
  Types of Scales and MeasuresThurstone scales. The Thurstone scale is made up of statements about aparticular issue and each statement has a numerical value indicating the repsondent’s attitude about the issue, either favorable or unfavorable. People indicate which of the statements with which they agree and the averageresponse is computed. First, you must be very clear about exactly what it is you’re trying to measure. Then, collect statements on the topic ranging from attitudes thatare favorable to unfavorable. For this example, we will use same sexmarriage . Example statements areIt ’s should be against the law.   There’s absolutely nothing wrong with it.  Marriage is between a man and a woman.It should be a sin. It’s perfectly appropriate for two consenting adults.  It should be legalized.It can harm children.Same sex couples should have the same legal rights as a male/femalecouple. It’s just horrible.   It can’t do any harm.  Next, you have judges evaluate, on an 11 point response format (1 verynegative to 11 indicating very positive), what kind of attitude each of these statements reflects. For example it’s likely that the statement “It should be asin.” would be judged to r  epresent a very negative attitude while the statement “It should be legalized.” would be judged to represent a verypositive attitude. The idea is that you’d like to develop a set of items that not only reflect the entire continuum between 1 and 11, but that your judges whohelped you develop the set of items would have considerable consensusabout what level of attitude each of the statements reflected. In thisexercise, statements for which there is little consensus would be discarded. So, let’s assu me that the average ratings among our judges are as below:AverageIt ’s should be against the law. 1.7 There’s absolutely nothing wrong with it. 6.8Marriage is between a man and a woman. 2.4It should be a sin. 1.3 It’s perfectly ap propriate for two consenting adults. 8.4It should be legalized. 9.1  It can harm children. 2.0Same sex couples should have the same legal rights as 7.6male/female couples. It’s just horrible. 1.5 It can’t do any harm. 6.9Suppose that there was reasonable consensus among our judges for the above items. When administering the scale, we’d ask individuals to indicate which of the above they agreed with (the average level that our judges hadagreed upon would not be indicated on the scale when administered toindividuals). Finally, the average of those checked would be calculated to determine the individual’s attitude.  Problems with developing Thurstone scales include 1) it can be quite timeconsuming and expensive, and 2) examples for the mid-points of the scalefor which there is consensus among the judges can be difficult to obtain.Guttman scales.With a Guttman scale, you have a set of statements so thata respondent who agrees with any specific statement in the list will alsoagree with all previous statements. In other words, each statementsubsumes the lower order statements. For example, a scale designed tomeasure Post-Traumatic Stress Syndrome, presented below, was basedupon the Guttman technique and was published in the Journal of ClinicalPsychology (1323-1328, 1984).Scale score Scale item0 In service 1965-19751 Stationed in Vietnam2 Saw injury or death of US Serviceman3 Fired weapon/fired upon in combat4 Responsible for death of enemy military5 Wounded in combat6 Responsible for death of enemy civilian7 Served third tour of duty in VietnamSemantic Differential scales. These scales measure an individual’s reactions to stimulus words and concepts in terms of ratings on bipolar adjectives ateach end. For example, an eight point response system that is anchoredfrom hot on one end to cold at the other, or friendly at one end with hostile atthe other, or good at one end with evil at the other, or polite at one end withrude at the other, etc. Typically a scale would consist of a minimum of fivebipolar pairs and the score would be obtained by adding the numeric values  of the responses. An example of a semantic differential scale is presentedbelow. This scale, the Least Preferred Co-Worker scale, was used by FredFiedler (1964) as a measure of leadership style.Pleasant __ __ __ __ __ __ __ __ __ UnpleasantFriendly __ __ __ __ __ __ __ __ __ UnfriendlyRejecting __ __ __ __ __ __ __ __ __ AcceptingTense __ __ __ __ __ __ __ __ __ RelaxedDistant __ __ __ __ __ __ __ __ __ CloseCold __ __ __ __ __ __ __ __ __ WarmSupportive __ __ __ __ __ __ __ __ __ HostileBoring __ __ __ __ __ __ __ __ __ InterestingQuarrelsome __ __ __ __ __ __ __ __ __ HarmoniousGloomy __ __ __ __ __ __ __ __ __ CheerfulOpen __ __ __ __ __ __ __ __ __ GuardedBackbiting __ __ __ __ __ __ __ __ __ LoyalUntrustworthy __ __ __ __ __ __ __ __ __ TrustworthyConsiderate __ __ __ __ __ __ __ __ __ InconsiderateNasty __ __ __ __ __ __ __ __ __ NiceAgreeable __ __ __ __ __ __ __ __ __ DisagreeableInsincere __ __ __ __ __ __ __ __ __ SincereKind __ __ __ __ __ __ __ __ __ UnkindLikert, Likert-type, or Summated Rating scales. Individuals withouttremendous expertise are able to develop sophisticated measures using theLikert method for developing scales. Although srcinally developed with anumber of statements in which individuals indicated their extent ofagreement with response choices such as 1 = disagree, 2 = disagreesomewhat, 3= neither agree nor disagree, 4 = agree somewhat, 5 = agree,statements with other response choices such as unimportant to important can be used. Additionally, it’s been found that if the anchors at each end aregood anchors, the choice points in between the end anchors do not need tobe labeled. Thus, a scale that measures job satisfaction might look like Very VeryDissatisfied SatisfiedThe actual job itself. 1 2 3 4 5The degree to which you feel 'motivated' by  your job. 1 2 3 4 5Current career opportunities. 1 2 3 4 5The level of security in your present job. 1 2 3 4 5The extent to which you may identify with thepublic image or goals of your organization. 1 2 3 4 5The style of supervision that your superiors use. 1 2 3 4 5Your level of salary relative to your experience. 1 2 3 4 5The way changes and innovations areimplemented. 1 2 3 4 5 For more information and the technical details about how Likert scales aredeveloped, please see the Likert video clip . Ranking. When it’s important to differentiate between various items or  characteristics, ranking may be the appropriate method to use. For example,when considering job characteristics such as pay, being part of the team,good supervision, etc., we would likely say that they are all important. If,however, we needed to differentiate among these characteristics we couldask respondents to rank the items in order of their importance, with 1 beingthe most important, 2 being the next most important, etc. Rank scales areipsative, which means that the responses are not independent of each other.We can only have one 1, one 2, etc. The problem with ipsative scales is thatit limits the types of statistical analysis which can be used.Questionnaire design. Questionnaire design is part art and part science.Essentially the goals are to obtain valid responses and to maximize the numberof completed questionnaires. A question that always exists is how do thoseindividuals who completed the questionnaire diffe r from those who didn’t. Thefollowing guidelines apply to questionnaires, whether paper/pencil, online, or liveinterviews.White space. Have plenty of white space. Don't clutter pages with codes,etc. or anything that doesn't help respondents. Demographics. Ask for only the information you truly need. Don’t ask for  extraneous information that may serve to alienate the respondent.Physical form. Use print large enough to easily read; use black ink on whiteor pale yellow paper; use a distinct, authoritative letterhead for cover letter.Language. If mailed to non-English speakers prepare separate versions.Length. Keep it short! In general no more than the front and back of onepage. If it must be longer use a brochure style with a 4 page maximum. Becareful not to include so many questions that potential respondents aredissuaded from responding ; it’s always tempting to add just one more
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks