Hypotheses Test

After assuring the manipulation, reliability and validity of the analyzed data, the study proceeded to conduct ANOVA, General Linear Model and Independent-Sample T Test to test the hypotheses.

4.4.1 Reason Provided and Satisfaction

Reason provided was given into three groups respectively: zero, one, and five reasons provided. ANOVA was used to test if satisfaction differed when the number of reasons provided differed. In hypothesis 1, the study speculated that impulsive consumers with more buying reasons were more satisfied than those with fewer reasons, and more satisfied than those with no reason.

Table 4.5 showed the descriptive statistics of the analysis and Table 4.6 presented ANOVA results which indicated that the main effect was significant (p<0.1). After multiple comparisons with least square deviation (LSD), the research found that statistically significant differences

existed between zero reasons and five reasons (p<0.1). No significant differences existed among groups of zero to one and one to five reasons provided (Table 4.7). Thus, H1a was not supported, suggesting that impulsive consumers with more reasons for buying were more satisfied than those with fewer reasons. H1b was not absolute. There was no significant difference between zero and one reason provided, however, statistical differences occurred between zero and five reasons. The result indicated that H1b was supported when sufficient buying reasons were provided

Table 4- 5 Descriptive Statistics of Reason Provided

Reason N Mean Std. Deviation

Zero 75 4.6519 .71528

One 78 4.8162 .68466

Five 74 4.9159 .75108

(Dependent Variable: Satisfaction)

Table 4- 6 Tests of Mood, Tendency to Regret and Reason Provided

Source Type III Sum

of Squares df Mean

Square F Sig.

Reason 2.886 2 1.443 2.862 .059

Mood .139 1 .139 .275 .600

Tendency to Regret .916 1 .916 1.816 .179

Reason * Mood .084 2 .042 .083 .920

Reason * Tendency to

Regret 2.354 2 1.177 2.334 .099

Mood * Tendency to

Regret 2.003 1 2.003 3.973 .048

Reason * Mood *

Tendency to Regret 1.007 2 .504 .999 .370

R Squared = .080 (Adjusted R Squared = .032)

(Dependent Variable: Satisfaction)

Table 4- 7 Multiple Comparisons (LSD) (I) Reason

Provided

(J) Reason

Provided Mean Difference (I-J) Std. Error p-value

Zero One -.1644 .11484 .154

Zero Five -.2641 .11636 .024

One Five -.0997 .11524 .388

4.4.2 Mood and Satisfaction

H2 was to check whether mood prior to shopping influenced impulse buyers’ satisfaction.

The papers discussed before stated that consumers in a good mood before shopping would have higher satisfaction than those in a bad mood.

Stimulant pictures classified participants into two groups, a positive mood and a negative mood respectively. The hypothesis was tested by comparing the means of satisfaction for each group; thus, an independent-sample T test was used to determine the difference in satisfaction between the two mood classifications. Eventually, the study found out that mood states prior to purchase was indifferent to the degree of satisfaction (Table 4.8). Hypothesis 2 was not supported (p>0.1).

Table 4- 8 Descriptive Statistics of Mood

Mood N Mean Std. Deviation

Positive Mood 116 4.7692 .71306

Negative Mood 111 4.8208 .73339

t-value = -0.538, p-value = 0.591

(Dependent Variable: Satisfaction)

4.4.3 Mood with Reason Provided and Satisfaction

H3a and H3b predicted that as buying reasons were provided, impulsive buyers in a negative mood would have more significant increased post-purchase satisfaction than those in a positive mood. However, according to H2 presented above, mood prior to purchase was indifferent to satisfaction, and Table 4.6 suggested that there was no interaction effect between mood and reason provided (p>0.1). As a whole, the study found that impulse buyers would not change their satisfaction depending on mood prior to purchase, and mood prior to purchase would not change the association between reason provided and satisfaction. Although there were some directional differences in the mean values shown in Table 4.9 and Figure 4.1, statistically no interaction effect existed between moods prior to purchase and reason provided. H3a and H3b were not supported.

Mood prior to purchase and tendency to regret showed interaction effect in Table 4.6 (Figure 4.2).

When respondents were in a positive mood, there were no differences between people with a high tendency to regret and with a low tendency to regret (p>0.1); however, there were significant difference when they were in a negative mood (p<0.1).

Table 4- 9 Descriptive Statistics of Mood and Reason Provided Reason Provided

Mean (Std. Deviation)

Zero Reason One Reason Five Reasons Positive Mood 4.5916 (.69194) 4.8307 (.65493) 4.8769 (.77988) Negative Mood 4.7105 (.74178) 4.7994 (.72682) 4.9550 (.72976)

Figure 4- 1 Interactions between Mood Prior to Purchase and Reason Provided

Figure 4- 2 Interactions between Mood Prior to Purchase and Reason Provided

4.4

4.4.4 Tendency to Regret with Reason Provided and Satisfaction

According to the papers discussed before, the study perceived people with higher tendency to regret would be more sensitive to reason provided after impulse buying. Median value (median value =4.6667) was used to classify participants into two groups. Thus, there were 114

respondents in high tendency to regret and 113 respondents in low tendency to regret.

When the study used tendency to regret as an independent variable, there was no difference in satisfaction between the two tendency groups. The main effect was not significant (Table 4.6).

However, there was some interaction between tendency to regret and reason provided (p<0.1), suggesting that the relationship between reason provided and satisfaction would be influenced by tendency to regret. Table 4.6 and Figure 4.2 showed the interaction between tendency to regret and reason provided. Multiple comparisons were used to test the inference. As a result, when reason provided was from zero to one reason, satisfaction of consumers with a high tendency to regret had significantly increase than those with low tendency to regret’s increase, supporting H4a. But the dissimilarity of satisfaction between zero to one and one to five reasons provided were indifferent to people with a high tendency to regret and a low tendency to regret. H4b was not supported (Table 4.10). For the three groups of reason provided, low tendency to regret was indifferent to satisfaction (p>0.1) but significantly different for high tendency to regret (p<0.05).

For low tendency to regret, zero to one and zero to five reasons provided a significant increase in satisfaction did not exist; for one to five reasons provided a significant increase existed. For a high tendency to regret, for one to five reasons provided a significant increase in satisfaction did not exist; for zero to one and zero to five reasons provided a significant increase existed (Table 4.12 & Table 4.13).

Table 4- 10 Contrast Tests

Contrast Estimations Value of Contrast

Std.

Error t df p-value

1. 5-4>2-1 .4411 .23043 1.914 221 .057 2. 6-5>3-2 -.3119 .23138 -1.348 221 .179 3. 6-4>3-1 .1292 .23302 .554 221 .580

Table 4- 11 Descriptive Statistics of Tendency to Regret and Reason Provided Reason Provided

Mean (Std. Deviation)

Zero Reason One Reason Five Reasons Low Tendency to Regret 4.8348(.77849) 4.7804(.54135) 5.0256(.69035) High Tendency to Regret 4.4737(.60621) 4.8603(.83424) 4.7937(.80582)

Total numbers =227

Table 4- 12 Multiple Comparisons of Tendency to Regret and Reason Provided Reason Provided

*The mean difference is significant at the 0.1 level.

Table 4- 13 Summary of Contrast Analysis Reason Provided

O: Satisfaction increased significantly. X: Satisfaction did not change significantly.

Figure 4- 3 Interactions between Tendency to Regret and Reason Provided

4.8348

4.7804

5.0256

4.4737

4.8603

4.7937

4.4 4.5 4.6 4.7 4.8 4.9 5 5.1

Zero One Five

Reason Provided

Satisfaction Low Tendency to

Regret

High Tendency to Regret