CHAPTER 5. ANALYSIS AND RESULTS
In this chapter, we review our statistical results and discuss the findings of our framework.
5.1 Data Analysis
First, we measured the gender preference of the posts by the number of likes, comments and shares according to gender. We found far more female than male fans on Company A’s Facebook Fan page; nearly all posts had more likes, comments and shares from females. Therefore, we calculated the Z-score for the like, comment and share numbers from males and females to eliminate the difference in proportion. We then compared the number to obtain our gender category in each post.
We used Emotional and Rational Proportion Scores (E/R Scores) for each post to determine the post’s emotional and rational proportion. The E/R Scores were calculated from the number of times that the coders chose emotional and rational descriptions for each post in SET-A, SET-B and SET-C. We also calculated the Z-score to standardize the times that emotional and rational descriptions were chosen to eliminate the bias because the emotional descriptions outnumbered the rational descriptions. For example, one post contained three sets of emotional and rational descriptions: SET-A-Emotional: 6 times; SET-A-Rational: 3 times; SET-B-Emotional: 8 times; SET-B-Rational: 2 times; SET-C-Emotional: 3 times;
SET-C-Rational: 1 time. We then calculated the Z-score for each post: SET-A-Emotional:
2.440728879; SET-A-Rational: 1.305651651; Emotional: 3.031139572; SET-B-Rational: 0.70337728; SET-C-Emotional: 0.510301378; SET-C-SET-B-Rational: -0.024324511. We list these data in Table 5-1.
Table 5-1 Post’s E/R Score Example
Times Z-score Abbreviation
SET-A-Emotional 6 2.440728879 SA-E
SET-A-Rational 3 1.305651651 SA-R
SET-B-Emotional 8 3.031139572 SB-E
SET-B-Rational 2 0.70337728 SB-R
SET-C-Emotional 3 0.510301378 SC-E
SET-C-Rational 1 -0.024324511 SC-R
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
We defined the E/R Score as
E/R Score = (SA-E + SB-E + SC-E) / (SA-E + SA-R + SB-E + SB-R + SC-E + SC-R)
Therefore, in our example, the post E/R Score is (2.440728879+3.031139572+
0.510301378) / (2.440728879+1.305651651+3.031139572+0.70337728+0.510301378+-0.024324511) = 0.750880413325324. A higher E/R Score means that the coders chose more Emotional descriptions on that post. We calculated the E/R Score for all posts and then calculated the descriptive statistics by E/R Score for each post. The results are shown in Table 5-2. The results show that the mean of the males’ E/R Score is 0.20125, which is smaller than the mean of the females’ E/R Score, at 0.84414. Because the females’ E/R Score has higher means, we can find that the females’ preferred posts might be more emotional than the males’ posts. In other words, males’ preferred posts might be more rational than those of females.
Table 5-2 E/R Score Group Statistics
Gender N Mean Std. Deviation Std. Error Mean E/R Score Male 288 .20124907350 5.130019689679 .302289309183
Female 261 .84414150427 3.849709022151 .238291006251
We use sensation-seeking and sensation-avoiding scores (SS/SA Scores) for each tourist product to determine each tourist product’s sensation-seeking and sensation-avoiding proportion. The SS/SA Scores were calculated from sensation-seeking points for each tourist product in SET-A and SET-B by coder. We also calculated the Z-score for sensation-seeking points to eliminate possible bias from two coders having different standards. For example, one tourist product had two sets of sensation-seeking points: SET-A: 2,1,1,1,1, with an average of 1.2; SET-B: 2,2,2,2,2, with an average of 2. Then, we calculated the Z-scores for SET-A and SET-B, which were -2.142010702 and -1.563288266, respectively. We list these data in Table 5-3.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table 5-3 Tourist Product SS/SA Score Example
Points Average Z-Score Abbreviation
SET-A-SS-Points 2,1,1,1,1 1.2 -2.142010702 SA-SP
SET-B-SS-Points 2,2,2,2,2 2 -1.563288266 SB-SP
We defined the SS/SA Score as
SS/SA Score = (SA-SP + SB-SP) / 2
Therefore, in our example, the tourist product SS/SA Score is (2.142010702 + -1.563288266) / 2 = -1.852649484. A higher SS/SA Score means that the coders chose higher sensation-seeking points for that tourist product. We calculated the SS/SA Score for each tourist product and then calculated the descriptive statistics according to SS/SA Score for each tourist product. The results are shown in Table 5-4. We also calculated the number of times that each tourist product appeared in our data set, and we determined the tourist products’ gender category by calculating the appearance of the posts’ gender preferences.
The results show that the mean of the males’ SS/SA Score, 3.356, is higher than the mean of the females’ E/R Score, 3.17143. The SS/SA Score has higher means, so we can find that males’ preferred tourist products might be more sensation seeking than females’. In other words, females’ tourist product preferences might be more sensation avoiding than males’.
Table 5-4 SS/SA Score Group Statistics
Gender N Mean Std. Deviation Std. Error Mean SS/SA
Score
Male 75 3.35600000000 .697144755995 .080499342508 Female 56 3.17142857143 .823013152599 .109979758637
After determining the posts’ E/R Score and SS/SA Score, we used these two scores minus their means to classify our posts, and we used the scores to draw an E/R Score and SS/SA Score four-cell chart, Chart 5-1.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Chart 5-1 E/R Score and SS/SA Score four-cell chart
In Chart 5-1, although many of the data are concentrated at the center, we can still see that more male preference posts seem to be concentrated on Sensation seeking and Rational, whereas more female preference posts seem to be concentrated on Sensation avoiding and Emotional. We now extract the posts with performances higher than one standard deviation and draw a four-cell chart, Chart 5-2. We can observe that these data seem to be concentrated at the upper-left and bottom-right. Thus, by matching gender preferences, the performances of Facebook fan pages might improve.
Sensation seeking
Emotional
Rational
Sensation avoiding
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Chart 5-2 E/R Score and SS/SA Score four-cell chart for better performance posts
After data judgment and descriptive statistics analysis, our hypothesis seems to be confirmed. We now proceed with a two-sample t-test to support our hypothesis.