You are on page 1of 20

Social Media Credibility SOCIAL MEDIA CREDIBILITY: WHO DO YOU TRUST MORE?

Abstract Media has shifted from traditional to new media, which includes many social media sites. Due to this fact many people are changing the outlets where they search for news. 94 participants from the Pennsylvania State University (12 men and 82 women) were randomly assigned to two different stimuli. One group viewed a Facebook stimulus and the other a Twitter stimulus. The groups were asked pretest questions on their media use and their attitude about Twitter and Facebook. Their responses were measured using a 10-point Likert scale. After being exposed to the stimulus, they were asked how they viewed the stimulus, either Onward States Twitter page or Onward States Facebook page, and their attitudes were rated on a 10-point likert scale. Results showed there was no significant difference between Facebook and Twitter credibility as a News source.
Hyangsook KIm 12/15/10 11:25 AM Comment: Should be a better idea to include significant findings instead of insignificant one. Hyangsook KIm 12/15/10 11:24 AM Deleted: l Hyangsook KIm 12/15/10 11:23 AM Comment: I appreciate your effort to include this abstract in your final report.

Introduction???

Hyangsook KIm 12/15/10 11:38 AM Comment: Where is it?? You had in the research proposal draft. You should have been more careful to the final report! Since you included the abstract, I did not subtract many point for this missing part, though. Sounds fair?
0.8 out of 1 point.

Literature Review We analyzed and critiqued five articles pertinent to our research question. We plan to research how the effects of different online media effects perceived credibility. Each article relates to various aspects of our research determining the most credible source of news. Johnson and Kaye (2004) investigated how traditional media and Internet reliance influenced the way blog users perceived the credibility of weblogs. The independent variable for the study was the type of different news outlets and the dependent variable was how the weblog users viewed the credibility of the different outlets. Johnson and Kaye (2004) defined credibility for the study as worthiness of being believed, accuracy, fairness, and depth of information. A moderating variable was the participants prior reliance on the different news outlets. This relates to our study, understanding how different users of media sources view credibility and what prior experiences can shape their view of the news outlet. They administered a survey to Weblog readers, which was posted for a month in 2003. The survey was linked to 131 different Weblogs, 14 Weblogoriented bulletin boards, sent the survey to Weblog chat rooms and bloggers for them to post themselves. This relates to our study in understanding how different users of media sources view credibility and what prior experiences can shape their view of the news outlet. The major finding of this study (Johnson & Kaye, 2004) was that a participants reliance on Weblog use was a strong and positive predictor for the perception of Weblog credibility. Most Weblog readers viewed the medium as moderately to highly credible. Also the participants viewed the Weblogs as more credible than any other news source.

Hyangsook KIm 12/15/10 11:34 AM Comment: Very good!


3 out of 3 points.

This guides us to believe that the more a person uses a source the more likely they are to view the source as credible, which could change how the outcome of the experiment is viewed if this variable is not accounted for in some way. We will be able to use the findings from Johnson and Kayes study (2004) to better account for and prevent confounding variables for our experiment. With our next article, researchers Hu and Sundar (2010) conducted a study, which showed comparisons that relate to our own research. The basis of this study is to examine the direct and combined influences of original sources (doctors vs. laypersons) and selecting sources (web sites vs. bulletin boards vs. blogs vs. personal home pages vs. Internet) on perceived credibility of-and behavioral intentions toward health information (Hu & Sundar, 2010). There are 3 IVs for the study which are first, original sources are doctors vs. laypersons so you will either have a Chris Park, M.D. or a layperson (someone who is not titled a doctor etc) Chris Park, the second is having a similar layout for the Internet (control variable) and the third is level of controversy on health issues for example milk vs. skin cancer. This is a 2 x 2 x 5 factorial design and the independent variable is simply the sources - whether web site, bulletin boards, personal home pages, Internet and blogs on perceived credibility. Also in the experiment, it was very important to do the two pretests for accurate results in the main experiment. We can use this study (Hu & Sundar, 2010) to formulate an appropriate study for testing social media and believability. We can emulate their Likert-type scale since it is more accurate in assessing peoples feelings relating to the media. Within this study we can see both the major findings and potential hazards with this research that is pertinent in our upcoming research. Two potential factors that may
Hyangsook KIm 12/15/10 11:28 AM Deleted:

have affected their research is if the particular people who claimed it was more believable were just less knowledgeable on the truth of the actual illness, etc. and if the person has family with issue (ex. Cancer) they are more likely than patients to contribute. To rule out alternative explanations it is necessary to control for confounding variables. Statistically you can control this factor by entering data of personal relevance with the issues of milk or skin cancer to the data analysis. The major findings of this study were that there was no significant main effect for selecting source on perceived credibility of information but it did show a significant effect of selecting source on behavioral intentions (which means websites highest and then bulletin board, blog, Home page and then lowest Internet). There were significant results supporting that three-way interaction between messages, original source and selecting source on perceived credibility. This may have to do with the participants perceived relevance to themselves (sunscreen & milk specifically). The main point of this research is that the person receiving the message has a certain psychological response of perceived credibility (accuracy and believability), which is dependent and impacted by the type of selecting source venues. In our study we can utilize this knowledge and apply the research and experimental design done by Hu and Sundar (2010). Another article we analyzed detailed the research conducted by Sutton, Palen, and Shklovski (2008). This study was conducted by first operationalizing the various social media outlets used, including photo sharing websites, personal blogs, web discussions and forums, and mobile phones. They then gathered information from respondents via an online questionnaire about information and communications technology use before and during disaster situations, and the perceived legitimacy of the social media and

backchannel media outlets. The independent variables for this study (Sutton, Palen, & Shklovski, 2008) were the respondents use of backchannel communication and the use of other social media outlets. The dependent variables were the prominence of social media and backchannel media outlets and the perceived legitimacy of the information gained using these types of media, both of which were measured using open-ended response questions on the online questionnaire. Similar to our own research question, this study focused on the perceived credibility of various social media outlets. The overall results of this study (Sutton, Palen, & Shklovski, 2008) showed that the use of backchannel media outlets and social media channels, including web forums, blogs, photo sharing websites, etc. is increasingly popular during times of disaster. The accessibility of these outlets encourages citizens to use and rely on information gained from backchannel media sources. The backchannel media outlets seem both credible and reliable in disaster situations because the information is current and personal. Respondents complaining of incorrect information being broadcast through typical, larger media channels were appeased by the close to home nature of backchannel media. Backchannel media appears to be a credible source for spreading information during disaster situations, and may be an important tool in spreading news and other information quickly to a large number of people in various situations. This study (Sutton, Palen, & Shklovski, 2008) provides a basis for our future research, in that it indicates that backchannel and social media outlets are perceived to be reliable by the general population. They also appear to be an accessible and current source for breaking news and updates which can be an explanation for the reason as to why people may perceive social media as a credible source for news which is pertinent to our study.

The research conducted by Kiousis (2001) is similarly related to the research that our group will be experimenting. This study (Kiousis, 2001) was tested on a sample population of randomly selected residents in Austin, Texas. The research was conducted to measure respondents attitudes regarded media channels of news from various sources including online, television, and newspaper. The method used was a 4-point Likert scale with the options of "strongly agreed", "agreed", "disagreed", or "strongly disagreed" were listed as options of opinion of credibility related to each given source of news. The dependent variables included perceptions of news credibility, channel credibility, and interpersonal communication. The independent variables were the three given media channels, which included newspaper, television, and Internet or online sources. This study (Kiousis, 2001) is almost directly related to the research that we will be conducting and is an excellent source of secondary research. The major findings of this study (Kiousis, 2001) showed that the newspaper as a source of news was found to be much more credible than the television as a news source. Newspapers were found most credible, followed by news found on the Internet, and television news as the least credible. Interestingly, the use of the media as a source of news was directly related to credibility in both newspaper and online sources, but for television there was no relation found. The research that was obtained through the research question and the researcher's hypothesis did not always match in terms of face validity and often times the proposed hypothesis did not show a direct correlation. When choosing the sample population for this study, the researcher chose a large collegiate population from Austin, Texas that may have negatively impacted this experiment by causing biased results in the conclusion of research. This relates to our study because we

also will be using research participants from a large collegiate based population which means our results may be impacted by this population just as they were in this particular study. Flanagin and Metzger (2007) conducted a 4 X 2 factorial design that was varied by genre of website in order to test the research question which was how does that role of certain factors in a natural environment effect credibility of information. In their research they evaluated specifically of the role of site features, user attributes, and information verification behaviors across different genres of websites in a somewhat natural environment. The dependent variables were the following types of credibility perceptions: sponsor credibility, observed information verification behaviors, message, website, and self-reported verification of website content. While this study (Flanagin & Metzger, 2007) does not focus specifically on the variation of news credibility it does relate to our research through the similarity in the analysis of credibility of media sources. The major findings of this study (Flanagin & Metzger, 2007) showed that the genre of website impacts the perceived credibility of the sponsor. They found that the genre of website also impacts the message found on the website, and the overall content of the website. News organization sites had the highest perceived credibility in comparison to other genres for both sponsor and message credibility. Also, there was no difference in perceived credibility amongst e-commerce and special interest sites while overall special interest sites were seen as more credible than personal sites. Factors that were positive effects of the research are being aware of the diversity of websites, which determines that there, may or may not be an effect. Some factors that may have
Hyangsook KIm 12/15/10 11:32 AM Deleted: . Hyangsook KIm 12/15/10 11:32 AM Comment: Still a bit weird argument. You indicated the fault of the study in terms of response bias possibility but you want to apply this strategy to your study???

negatively impacted the research were presented in Flanagin and Metzgers (2007) study. The fabrication of sites for this study (Flanagin & Metzger, 2007) may have interfered with the results because they were compared to actual and recognizable sites. We can utilize the findings of this study (Flanagin & Metzger, 2007) in our own research as a starting point for our research in news credibility. In our experiment we will be testing for Penn State Students in order to determine which online media outlet has the highest perceived credibility. Our hypothesis is that credibility of the news source will be highest for the networks Twitter page. We believe this will be the result of the study because Twitter presents itself as a more of information driven site than other social driven sites such as Facebook. The independent variables we are using are different online media outlets. Will examine two variations of online media including Twitter, Facebook, measuring the perceived credibility of new source as our dependent variable, and we will use the news networks website as a control to compare the credibility of social media sources. We will control for the news network limiting the source to the American Broadcasting Company (ABC), because it reaches a large audience and is usually viewed as politically neutral. Hypothesis and Research Question Research Question: What is the relationship between the type of social media and the level of perceived credibility?

Hypothesis 1: The more frequently Twitter is used, the higher perceived credibility of Twitter as a source of news

Hypothesis 2: The more frequently Facebook is used, the higher the level of perceived credibility of Facebook as a source of news. Hypothesis 3: The credibility of social media as a news source will be highest for Twitter. Methods Participants For this study the researchers will target the Penn State University main campus undergraduate students. There will be 94 participants from this demographic. The researchers will obtain our participants for the survey by sending the survey link out over Twitter and Facebook and ask that only Penn State main campus undergraduate students participate. Our sampling method will be a probability sample. Variables Independent Variables The independent variables are the social media sites, Facebook and Twitter. To operationalize our independent variable the researchers will take separate screen shots of both a manipulated Facebook and Twitter page. They will then include a screen shot for the participants to see in order to answer follow up questions about each source. They will manipulate each screen shot so they have the same source (The Daily Collegian) and the same 160-word headline. This will be done prior to providing the picture in order to get rid of any confounding variables that could be caused by difference in article information. The researchers broke the independent variable of social media into two separate categories. The first category is Facebook and the second is Twitter. The
Hyangsook KIm 12/15/10 12:43 PM Comment: Because now everything has been conducted at the moment when you were writing this section, it should have been past tense for all verbs. This rule is applied to all of the method section.
Except some issues found in the specific comments below, no major problems were detected. Good job. 3.5 out of 4 points

researchers did this to see if credibility varies between high and low amounts, in different social media platforms in societys view of credibility today. Dependent Variable The dependent variable is level of perceived credibility. The dependent variables will be measured on a 1 to 10 Likert scales. The scales were determined by the fact that a 10-point Likert provide substantial feedback for our analyzing at the end of the study of participants is conducted. The researchers can emulate the Hu and Sundars study (2010) that we analyzed in our literature review, and model it from that. The questions came loosely from previous studies that we have reviewed. For example, the researchers were able to sample questions from the journals that we analyzed in each of our Paper Critique. The questions that they have formulated to ask the participants are the ones that they believe will give them the most crucial and beneficial evidence in order to analyze the study through the means of analytical testing. Procedure The stimulus materials are representing different types of social media platforms one being a Facebook screen shot and the other will be a Twitter screen shot. The researchers are going to take a screen shot of the OnwardState Facebook page and Twitter page. Onwardstate is a Penn State online news source that reaches out to Penn State students with information about things happening in the State College area and Penn State network. The screen shot will be a photo of the actual websites so it will be a realistic representation of both of the websites. Once the researchers take the screen shots, they will manipulate the shot so followers, founding date, and article are the same on each page to get rid of any confounding variables that could be caused by differences
Hyangsook KIm 12/15/10 12:40 PM Comment: In fact, you were supposed to state out what the measures were from the previous studies, not your plan of how to use them. Please refer to the attached done by previous students. : this is the comment for this section, but you did not appropriately address the comment.

Hyangsook KIm 12/15/10 12:30 PM Comment: APA error - 0.2 Hyangsook KIm 12/15/10 12:29 PM Deleted: & Hyangsook KIm 12/15/10 12:29 PM Deleted: 2010

in these variables. They will then add the links to each of these screen shots to the survey so the participant can view these and then answer the questions about how they feel and think about the different sites (Twitter and Facebook). The surveys will be administered through Facebook. They are created through a Grad Student website, which Professor Kim accessed. Professor Kim then sent the links for the surveys to the researchers and then the researchers were able to share the link with participants. The researchers will include the links of our screen shots in the survey, which were created on Photoshop. They will take the links and then post them on both Twitter and Facebook to reach the target demographic. The participants can then take this survey online, on any computer. Professor Kim will store the data for the researchers and allow them to check to see the answers and how many people have taken the survey. The researchers will then combine the data from the different conditions and then analyze the data using the SPSS software program. Results
Hyangsook KIm 12/15/10 1:03 PM Comment: Nice work except one thing. I asked you to include tables and figures at end of the paper even after the reference section before the final presentation. Is there anyone who listened to me? I announced this information twice. Anyone among four of you??? I am asking you this because there are students who listened to me and read the examples and did include these tables and figures at the end of the paper in this class. Sigh..
5.5 out of 6 points

In order to test the relationship between the experiment condition and the perceived credibility of social news, an independent T-test was used. The results of the independent T-test showed that there is not a significant relationship between the experiment condition and the perceived credibility of social news [Condition 1 (Facebook): M = 5.39 SD = 1.84, Condition 2 (Twitter): M = 4.94 SD = 1.63] t(94)=1.21, p=.23. Group Statistics Study Condition 1 2 N 58 38 Mean Std. Deviation 5.3858 1.84193 4.9408 1.62532 Std. Error Mean .24186 .26366

Social News Credibility

In order to test the effects of the experiment condition in conjunction with the frequency of social media use on the perceived credibility of social news, the factorial ANOVA test was used. Two separate ANOVA tests were performed, the first analyzed the frequency of Facebook use and the second analyzed the frequency of Twitter use. The factorial ANOVA test results, for the frequency of Facebook, showed that there is no significant main effect of the experiment condition [F(1,88)=.41, p=.52] and that there is no significant main effect of the number of times per day the respondent used Facebook [F(3,88)=1.48, p=.23]. The results also show that there is no significant interaction effect of the experiment condition with the number of times per day the respondent used Facebook [F(3,88)=.29, p=.83]. Tests of Between-Subjects Effects Dependent Variable:Social News Credibility Type III Sum of Mean Squares Square Source df F Sig. a Corrected Model 28.743 7 4.106 1.354 .235 Intercept 1195.569 1 1195.569 394.150 .000 condition 1.257 1 1.257 .414 .521 FacebookTimes 13.426 3 4.475 1.475 .227 ADay condition * 2.624 3 .875 .288 .834 FacebookTimes ADay Error 266.929 88 3.033 Total 2901.141 96 Corrected Total 295.672 95 The factorial ANOVA test results, for the frequency of Twitter, showed a a. R Squared = .097 (Adjusted R Squared = .025) significant main effect of the number of times per day the respondent used Twitter [F(3,24)=3.75, p<.05]. There was no significant main effect of the experiment condition [F(1,24)=.006, p=.94] and there was no significant interaction effect of the experiment

condition with the number of times per day the respondent used Twitter [F(1,24)=1.81, p=.19]. After finding little significance the researchers condensed the social media usage independent variables (FacebookTimesADay, TwitterTimesADay) from 4 ordinal levels, to 2 nominal levels (FBUsage, TwitterUsage). The independent t-test was again used to measure the relationship between the level of social media usage and the social media credibility. The results of the independent t-test showed significance between social media usage and perceived social media credibility for both experiment conditions (Facebook, Twitter). The first test analyzed Facebook usage [Condition 1 (low): M = 4.67 SD = 1.76, Condition 2

(high): M = 5.46 SD = 1.72] t(94)=-2.07, p < .05. The second test analyzed Twitter usage [Condition 1 (low): M = 4.82 1.70] t(28)=-2.84, p < .05. SD = 1.47, Condition 2 (high): M = 6.53 SD =

Group Statistics FBUsage Social News Credibility 1.00 2.00 N 30 66 Mean 4.6667 5.4564 Std. Deviation 1.76146 1.72238 Std. Error Mean .32160 .21201

Independent Samples Test Levene's Test for Equality of Variances t-test for Equality of Means Facebook Usage 95% Confidence Std. Sig. Mean Error Interval of the (2Differ Differ Difference df tailed) ence ence Lower Upper 94 .041 - .3819 .7897 3 1.548 .0314 7 10 4

F Equal variances assumed .042

Sig. .838

t 2.0 68

Group Statistics TwitterUsage Social News Credibility 1.00 2.00 N 20 10 Mean 4.8188 6.5250 Std. Deviation 1.47260 1.70151 Std. Error Mean .32928 .53806

Independent Samples Test Levene's Test for Equality of Variances t-test for Equality of Means Twitter Usage 95% Confidence Std. Sig. Mean Error Interval of the Difference (2Differ Differ df tailed) ence ence Lower Upper 28 .008 - .6002 1.706 6 2.935 .4766 25 83 7 15. .016 - .6308 944 1.706 3 3.043 .3685 25 92 8

F Social Equal News variances Credibility assumed Equal variances not assumed .519

Sig. .477

t 2.8 43 2.7 05

Two 2x2 factorial ANOVA tests were then used to test the relationship between social media usage and perceived credibility. The factorial ANOVA test results, for the

frequency of Facebook usage, showed that there is no significant main effect of the experiment condition [F(1,92)=.85, p=.36] and that the main effect of the number of times per day the respondent used Facebook is approaching significance, but still fails to establish concrete evidence of significance [F(1,92)=3.668, p=.059]. The results also show that there is no significant interaction effect of the experiment condition with the number of times per day the respondent used Facebook [F(1,92)=.12, p=.73].

Tests of Between-Subjects Effects Dependent Variable:Social News Credibility Type III Sum of Source Corrected Model Intercept condition FBUsage condition * FBUsage Error Total Corrected Total Squares 17.021a 2017.958 2.587 11.095 .356 278.651 2901.141 295.672 df 3 1 1 1 1 92 96 95 Mean Square 5.674 2017.958 2.587 11.095 .356 3.029 F 1.873 666.253 .854 3.663 .118 Sig. .140 .000 .358 .059 .732

a. R Squared = .058 (Adjusted R Squared = .027)

The factorial ANOVA test results, for the frequency of Twitter usage, showed a significant main effect of the number of times per day the respondent used Twitter [F(1,26)=6.80, p<.05]. There was no significant main effect of the experiment condition [F(1,26)=.58, p=.45]. The results for this test also showed a significant interaction effect of the experiment condition with the number of times per day the respondent used Twitter [F(1,26)=4.89, p<.05]. Tests of Between-Subjects Effects Dependent Variable:Social News Credibility Type III Sum of Source Corrected Model Intercept condition TwitterUsage condition * TwitterUsage Error Total Corrected Total Squares 30.129a 831.134 1.258 14.778 10.628 56.538 957.422 86.667 df 3 1 1 1 1 26 30 29 Mean Square 10.043 831.134 1.258 14.778 10.628 2.175 F 4.618 382.211 .578 6.796 4.887 Sig. .010 .000 .454 .015 .036

a. R Squared = .348 (Adjusted R Squared = .272) Discussion


Hyangsook KIm 12/15/10 1:17 PM Comment: It is nice to summarize the results in the beginning of the discussion, but it shouldnt be the whole discussion. I think you should have been able to elaborate this part more than what you have now here, especially in the line with findings from the previous studies that you reviewed. Or at least something discussed in the final presentation.
5.2 out of 6 points.

In both experiment conditions, the participants were shown a screen shot of a news article from a social media site. After viewing the image, participants were asked

to rate the perceived credibility of social media news. The results show that there was no significant relationship between the experiment conditions and the perceived credibility of social media news. Both experiment conditions perceived social media news as moderately credible. This could have been the result of similarities between the different social media forms (Facebook, Twitter) to which participants were exposed. The effects of the experiment conditions in conjunction with the frequency of social media use on the perceived credibility of social media news were then tested. In doing this, the conditions were evaluated separately. The Facebook experiment condition showed no significant main effects of the experiment condition or the frequency of social media usage on perceived credibility. There was also no significant interaction effect of the experiment condition with the frequency of social media usage on perceived credibility. This may have been caused by the separation of social media usage into four levels. The Twitter experiment condition showed no main effect of the experiment condition on perceived credibility, and no interaction effect between the experiment condition with the frequency of social media usage on perceived credibility. The results, however, did show a significant main effect of the frequency of social media usage on perceived credibility. Again, the lack of significance in these results may have been caused by the separation of social media usage on perceived credibility. As a result of the lack of significance in these results, the four ordinal levels of social media usage were combined into two nominal levels, low and high usage. After combining the social media usage levels the researchers re-ran the above tests and as expected, more significance was found among the test results.

The second round of analysis tests showed significance between the level of social media usage and the level of social media usage. The significance was seen between the levels of usage within each condition, and perceived credibility between the two conditions was similar. The relationship between social media usage and perceived credibility was again analyzed with two separate tests. These tests showed more significance between the variables, in that the frequency of Twitter usage proved to be significant in determining perceived credibility, and the frequency of Facebook usage was marginally significant in perceived credibility. This research could be beneficial for news media outlets because the results show that, overall, social media is perceived as a relatively credible source for news information, and that Twitter is used more frequently as a social news media source than Facebook. Companies can also use this research when looking into their target audience in order to predict which social media platforms are used more frequently. From that information they will be able to determine where their message should be disseminated. Limitations The experiment included it own limitations including a possible gender bias that was caused by the fact that the survey was on voluntary basis and there was a higher percentage of responses from women in the study than men. Because the researchers were not able to create a equal percentage of gender response from men and women, there was no way to prevent a possible gender bias from occurring. Also the study is limited to a population that included students at the undergraduate level, because the 100% of the participants were a part of this population. This prevents
Hyangsook KIm 12/15/10 1:10 PM Comment: Is that all???

this survey from being generalized to an older or younger population who were not represented in the study. Lastly, the specific nature of the two independent variables, Twitter and Facebook, may exclude other social media platforms that participants may use for a source of news. This can limit the findings through exclusion of other variables. Future Studies Future studies should look further into researching how the amount of time spent on a social media platform effects perceived credibility. If perceived credibility of social media as a news source depends on the amount of time spent on a person spends on a specific social media platform. Also, future studies could allow for more social media platforms to be included in the study, which would help to reduce limitation of the current study. This would also allow for more specific results and less diverse. Other social media that could be included would be MySpace, Blogs and StockNoise. Total: (0.8+3+3.5+5.5+5.2) 0.2 = 17.8 out of 20 points. References Flanagin, A.J., & Metzger, M. J. (2007). The role of site features, user attributes, and information verification behaviors on the perceived credibility of web-based information. New Media & Society, 9, 319-342.

Hu, Y., & Sundar, S. S. (2010). Effects of online health sources on credibility and behavioral intentions. Communication Research, 37, 105-132.

Johnson, T. J., & Kaye, B. K. (2004). Wag the blog: How reliance on traditional media and the Internet influence credibility perceptions of weblogs among blog users. Journalism and Mass Communication Quarterly, 81, 622-642.

Kiousis, S. (2001). Public trust or mistrust?: Perceptions of media credibility in the information age. Mass Communication and Society, 4, 381-403.

Sutton, J., Palen, A., & Shklovski, I. (2008). Backchannels on the front lines: Emergent uses of social media in the 2007 Southern California wildfires. In Proceedings of the 5th International ISCRAM Conference, Washington, DC.

You might also like