Characteristics

plat-forms to gain their desired knowledge. The different social media platplat-forms should also be consid-ered when wanting to use platforms for advertising products, as there is a cultural dimension to it.

Facebook, for example, is blocked by the Chinese government, so a global Facebook campaign would not reach China.

As stated by Wikhamn (2019): "innovations have positive impacts on customer preference, service quality, employee productivity, firms' market value, and share, and customer retention," and new social media platforms and technology have been proven to satisfy consumers. A satisfied con-sumer could be a loyal one.

respon-dent's motives or disclose the full opinions or reasoning behind their replies as it does not reveal their full motives or factors behind their reasoning (Jensen & Knudsen, 2014 p. 14). By using a fixed survey questionnaire, the respondent's cognitive processes kick in, and memories or forgotten motives are easier to retrieve by the respondents.

The attitudes and motives behind the respondent's choices are more precise when using an open question survey, as the respondents can freely formulate their responses and use their own words. Therefore, the survey's results will be compared to the interviews conducted. Even if re-spondents have different motives and reasoning behind their behavior, the interviews will shine a light on some of the motives and can be used as a checkpoint.

In order to not waste either respondent's or researcher's time, the questionnaire starts with background questions. Since the thesis has a geographic location, any respondents not living or hav-ing bought a home in Los Angeles should not reply to the survey. By starthav-ing with the background, the respondents will be more motivated to continue the questionnaire as they are easy and fast. The background questions include gender, age, income, and time of home purchase. (Appendix 6). The nominal scale has been used as scale level for these questions to categorize respondents equally (Jensen & Knudsen, 2014 p. 16). The respondent's persona has now been illustrated, making it easi-er to geneasi-eralize the results and inteasi-erpret the results. Some researcheasi-ers argue that these questions should be saved till last, as some respondents may more comfortably answer the sensitive question as they would feel more anonymous, but it was not chosen to do so as the final interpretation would be easier.

The second part of the questionnaire addresses the respondent's attitudes towards social me-dia and the platforms Facebook and Instagram. Some questions are based on the TAM model, to get an understanding of if they use the different platforms. Within this section, the respondents are ex-plicitly asked to respond based on their perception or attitude - whether or not they have used the platforms before. The third part of the questionnaire is constructed based on the Theory of Buyer Behavior model and the Heuristic-Systematic model to understand the respondent's reasoning be-hind choosing a particular brand or responding to specific content.

The second and third section requires the respondent to utilize a higher cognitive engage-ment level than the background questions, which is why an ordinary 5 point Likert scale is used for categorization: Disagree, partly disagree, neutral, partially agree and agree. Since the respondent cannot reply "do not know", the responder has to express their opinion on every question. By giving the respondent an equal designation of negative and positive statements, the categories are bal-anced, and the data conducted is as objective as possible (Malhotra et al., 2012 p. 427). All ques-tions are curated so the responses can have an opinion, and it is therefore assumed that the respon-dents have one. These assumptions are based on the real estate market being widely known and as-suming the Los Angeles population has knowledge or an attitude towards the market. The response option "Neutral" may influence the data set if respondents do not have an attitude towards a state-ment (Malhotra et al., 2012 p. 428).

As previously described, the advantages of using the Likert scale is that it makes it easier for the researcher to administer the results and construct the questionnaire. One of the disadvantages of using the Likert scale is that the researcher assumes that the respondent understands the underlying issues, and all respondents are rating the questions in the same manner. The number of concepts and outside understanding has, therefore, been attempted minimized, as well as words that can be inter-preted in several ways are not used. When dealing with the consumer's intent, the scale has been switched to an interval scale, also containing five possibilities; 0-6 months, 6 months-1 year, 1-1,5 year, 1,5-2 years, or never. The time perspective will force the respondent to consider their stand and make their intent an actual behavior.

When creating a survey, it is not always a fair representation of either population or sample.

As the survey had a theme on real estate, there is an excellent chance that people who responded either have an interest in real estate or answered based on their relationship with the researcher. The survey was sent out on mail and posted on Facebook. Based on this, it can already be determined that people who do not have access to a computer or internet are excluded from the survey. As the Facebook posted on is owned by the researcher, it is assumed that most of the people who respond-ed are either friends with the researcher, friends of the researcher's friends, or part of the groups the survey had been posted in. These were groups about real estate in Los Angeles on Facebook that the researcher had not created.

Some survey results could look skewed or unrealistic, but with the different groups used to show results that do not have an equal number of respondents, some results could seem rather high, but with a lower number of respondents, the percentage will seem more skewed.