After reviewing the online questionnaire with Likert scales regarding how Instagram Influencers affect consumers’ purchase decisions, I had found that there are several areas for improvement. Firstly, by having 5-point Likert scale in the questionnaire, there is possibility that the respondents will have central tendency bias, which is defined as a response bias that inclines respondents to avoid the extreme ends of the answers and prefer answers that are more towards the mid-point responses (Douven 2018, p. 1203). Therefore, the respondents may simply answer neutrally by clicking the middle point of 3 in every answer, and caused inaccuracy of research method. It is recommended to improve the questionnaire by changing the format of Likert scale into 4-points or 6-points Likert scale, so that the central tendency bias will be eliminated.
Secondly, the invitation to participate in the questionnaire on a large scale may invite the wrong group of samples, such as people who do not use Instagram at all or people who is not too familiar with Instagram’s functions of advertising using Instagram influencers, which will in turn create more inaccuracy in the answers acquired by the researcher. Therefore, it is advised to reduce the randomity of selecting samples by ensuring the suitability of the background of participants, which may be done through judgmental sampling that filters out unsuitable participants from the researcher’s professional judgment, or through the inclusion of a question in the questionnaire that acquires the Instagram username of the participants so that the researcher can judge the validity of the participant’s answers by checking the status and activity of the participant (Ponemon & Wendell 2019, p. 18). Regarding checking the participant’s Instagram, the researcher can observe the posting frequency or the numbers of followers and the numbers of people whom the participant is following.
References
Douven, I 2018, ‘A Bayesian perspective on Likert scales and central tendency’, Psychonomic Bulletin & Review, vol. 25, no. 3, pp. 1203–1211, accessed 14/4/2019,
<https://search-ebscohost-com.ezproxy.uow.edu.au/login.aspx?direct=true&db=mnh&AN=28752379&site=eds-live>.
Ponemon, L & Wendell, J 2019, ‘Judgmental Versus Random Sampling in Auditing – an Experimental Investigation’, Auditing-A Journal Of Practice & Theory, vol. 14, no. 2, pp. 17–34, accessed 15/4/2019,
<https://search-ebscohost-com.ezproxy.uow.edu.au/login.aspx?direct=true&db=edswss&AN=A1995RX14400002&site=eds-live>.
MBA 