Table III.

Scale development process (four-step)

StepsMethodsDataResults
Step 1: Definition of negative online brand engagement, Item generationActivity 1Literature review on (a) negative and (b) positive consumer engagementFor the period 2000–2020 all papers on negative engagement (31 articles) and 314 articles on positive engagement were selected using a systematic approach (Table IV)Construct definition, dimensionality and dimensions definitions (Table VI) Generation of the initial set of 171 items
Activity 2Online observation (Study 1)A total of 654 Amazon reviews (73,110 words).
Activity 3Semi-structured interviews with moderators of online anti-brand communities (Study 2)10 moderators produced 40,231 words of transcription (Table V)
Activity 4Semi-structured interviews with members of online anti-brand communities (Study 3)15 community members produced 54,506 words of transcription (Table V)
Step 2: Item purificationActivity 1Face validityAsk six anti-brand community members if the item pool is relevant to, can be useful and appropriate for what it intends to measure.Reduced the 171 to 160 items
Activity 2Research team meetingThirteen 60-min-long face-to-face meetings between co-authors.Reduced the 160 to 61 items
Activity 3Academic experts panel (Study 4)29 academic researchers in brandingReduced the 61 to 32 items
Step 3: Reliability and validityActivity 1Pre-test (Study 5)20 marketing researchers at the UK-based UniversityAll 32 items were retained
Activity 2Pilot study (Study 6)A convenience sample (author's network and snowballing) generated 41 usable responsesNo reliability or any other issues concerning administration or response
Activity 3Item reduction – Exploratory Factor Analysis (Study 7)Calibration sample collected from Social Media anti-brand communities (N = 205) (Table VII)Reduced the 32 to 27 items (Table VIII)
Activity 4Item reduction – Confirmatory factor analysis (Study 8)Both Calibration sample collected from Social Media anti-brand communities (N = 205) and Validation sample collected via snowballing in Social Media (N = 205) (Table VII)Reduced the 27 to 17 items The 17-item scale exhibited a good fit and properties (Tables IX and X)
Step 4: Nomological network, Discriminant validityActivity 1Negative online brand engagement relationships with brand disloyalty and happiness (Study 9)Validation sample collected from snowballing in Social Media (N = 205) (Table VII)Nomological validity supported indicating reliable and validity with good fit and properties (Tables XI)

Source: Authors' own work

or Create an Account

Close Modal
Close Modal