The problem of length is not limited to the questions, however. 

 Often, as in the CETA Follow-Up Instrument, there is a potential 

 need for identifying multiple gradations among response cate- 

 gories. When several response categories are used, two major 

 problems are often encountered: (1) remembering the categories 

 (reminding respondent of them can help here), and (2) the difficulty 

 of finding the proper words to connote intensity of feelings of 

 response. The latter problem is best addressed by sacrificing some of 

 the information by reducing the number of categories. This problem 

 and its potential solution is most prevalent among open ended, 

 attitudinal, opinion type, and historical recall questions (most 

 prominently in this survey the questions dealing with labor 

 history). Question ordering is also very important. Not only should 

 similar and related questions be grouped together into sections, but 

 these sections should be logically related so as to provide for a 

 smooth transition between and among them. Thus, the context of 

 the instrument plays an important, albeit unobtrusive, role in 

 conveying to the respondent that this procedure is worthwhile. In 

 the CETA Follow-Up Evaluation used in the Greenwood-Luloff 

 report, this problem is critical. For example, since the measurement 

 of any improvements in employment history among respondents 

 following CETA participation is a main goal of the survey, it would 

 appear that a question which directly asks for the duration of the 

 respondents employment (or unemployment) during said period 

 would appear critical. 



However, in order to arrive at this(ese), figure(s), it was first 

 necessary to go through a 6-step procedure. The lack of specific 

 questions, especially in the presence of detailed minutia, contributes 

 to the confusion indicated by the respondents, coders and re- 

 searchers in this study. It appears that too much time was spent on 

 seeking specific answers without regard for the overall or "big 

 picture." 



In addition, it should be pointed out that the first few questions 

 of an instrument are the most critical. There are basically three 

 reasons for this: (1) if a response can be obtained to these questions, 

 the likelihood of a subsequent termination prior to completion of the 

 instrument is greatly reduced, (2) these questions can be used to 

 divert objections to the survey by drawing the respondents attention 

 away from other concerns, and (3) the way these questions are asked 

 and answered tends to set the tone for the ensuing interview 

 (Dillman, 1978:218). Therefore, it is necessary to make these first few 

 questions interesting and relevant to the respondents. This is not 

 accomplished by beginning with demographic or background 

 questions (wliich should be the last ones asked in a telephone 

 interview) as in the CETA Follow-Up Survey. 



When detailed information is necessary to track a respondents 

 history with respect to certain variables, a careful and straight 

 forward design is needed. The use of "screening questions" might 

 help here. All questions which refer to all respondents should begin 



8 



