Survey Instrument Analysis

CHAPTER 6

Survey Instrument: Development and Validation

6.1 Introduction

In this chapter we describe a survey instrument designed to investigate broadband transformational power, usage and impact within the Indian household. We also describe the development and validation of so designed survey instrument.

The following three stages led to the development of a reliable instrument:

(1) To explain broadband awareness and transformational power behavior, some initial factors were identified from the literature and then a decision upon how to determine them in an exploratory survey approach;

(2) Content validation was performed on results from the exploratory survey. The purpose of this step was to confirm that items were typical representative to a particular construct domain; and

(3) A pre-test and a pilot test were conducted

After the development of the survey instrument, the reliability of the measure was confirmed with the help of a content validation test.

The next section briefly re-introduces the conceptual model and provides a list of the constructs included in the various stages of the validation process. The instrument testing process that includes the pre and pilot test is described in Section 4.6. Finally, the summary and conclusions are presented in Section 5.7.

6.2 CONCEPTUAL MODEL OF CONSTRUCTS

A brief account of the constructs is provided in this section. The constructs included in this study were adapted from the Model of the Transformational power of Technology in Households (MATH) (utilitarian outcomes, knowledge and)hedonic outcomes) ( by Browns and Venkatesh, 2001), diffusion of Transformational power, Usage and Impact of Broadband : India Households innovations (relative advantage) (Rogers, 1...

... middle of paper ...

...aviour (Ajzen, 1991).

2. Need for New Items

The two constructs, 'secondary influence' and 'self-efficacy' were composed of only one item each (Appendix 6.2). This caused problems when calculating the reliability of the construct. Therefore, this limitation was also considered during the content validity test and one more additional item was added to each construct (Table 6.7).

3. Problem of Low Reliability

From the exploratory study findings and the aforementioned discussion it was found that although the estimated mean value of many constructs (e. g. relative advantage) was high, the reliability (alpha) was low (Appendix 6.2). On the other hand, the estimated mean value of the social influence construct was low, but its reliability value was higher than any other construct. These variations in estimated values necessitated a further validation of the instrument