Results

Descriptive Statistics
Prior to testing the hypotheses, preliminary descriptive statistics were computed for categorical, demographic variables.  Ranges, means, and standard deviations were calculated for continuous variables.  The only data missing from the participants was related to a few demographic questions (age, education, gender, and ethnicity).  In the cases where demographic information was used, those cases will be excluded.  Due to a technical error in setup, for a few days of data collection, income was not a requirement to participate in the survey; when the error was discovered it was fixed.  The United States residency and age requirements were still in place for all data collected.  In total, 250 completed surveys were collected and 198 of those met the income requirement.  All results with p values between .05 and .001 were statistically significant during analysis.

Table 3.1 shows descriptive statistics for the participant demographic characteristics.  The sample contained 198 individuals including 126 males and 68 females.  Almost half of the participants fell into the 45-60 age group (48.5%).  The other age groups included 18-29 (7.6%), 30-44 (14.6%), over 60 (27.3%), and 2% were unreported.  The sample population reported their highest level of completed education levels as:  9.6% had obtained GED/high school diploma, 7.6% had obtained associate’s degree, 39.9% had obtained bachelor’s degree, 29.3% had obtained master’s degree, and 13.6% reported a doctorate degree.  Participants reported household incomes were $100,065-$150,000 (45.5%), $150,000-$200,000 (29.8%) and 24.7% earned more than $200,000.  The sample was predominantly Caucasian (90.4%).

t31
The descriptive statistics that focused more on technology, the Internet, and technology used are shown in Table 3.2.  The majority of participants reported using the Internet for seven or more years (98%).  The participants were asked to self-report their Internet skill level and 3.5% reported below average, 35.9% average skill, 45.5% above average skill, and 15.2% reported an expert skill level.  Google was the top search engine used by participants (86.9%).  Participants reported primarily using the Internet at home (81.8%), at the office (14.6%), and on mobile devices (3.5%).  The participants were also asked to report the device that they use when searching for product information on the Internet.  Participants reported they used laptops (46%), desktops (44.4%), cell phones (2.5%), and tablet computers (7.1%).

t32

Table 3.3 shows the descriptive statistics for the composite variables used in the study.  The composite variables are attitude toward Internet advertising, commercial Internet usage, attitude toward the website, attitude toward the brand, and behavior intention.  The attitude toward Internet advertising was calculated by adding the composite scores for perceived informativeness, perceived irritation, perceived entertainment, and perceived credibility.  Some of the questions had to be reverse-coded.  The reverse coded questions were Likert questions mostly related to perceived irritation.  The values needed to be reverse scored so that Internet advertising attitude could be accurately calculated as a high score equaling a more positive attitude.  Attitude toward Internet advertising was divided into two nominal groups at the midpoint; one with positive Internet advertising attitudes and one with less positive Internet advertising attitudes.  The scores for using the Internet to gather information related to commercial products and services and shopping/e-commerce were added to create a new variable.  This variable was divided into two groups at the midpoint; one with high commercial usage and one with low.  The questions for attitude toward the website, attitude toward the brand, and behavior intention were added to create three new variables for attitude toward the website, attitude toward the brand, and behavior intention.  In analysis, no questions were found to affect the reliability of any of the scales.

t33

Manipulation Check for Texts
An independent-sample t-test was conducted to check for manipulation with the symbolic and instrumental text.  One question asked participants to select “which statement about the website message you believe is more accurate” with response options being either (a) the message emphasized features and benefits of automobiles or (b) the messages appealed to a buyer’s image and style of social status.  There was a statistically significant difference between the symbolic (N=93, M=1.88, SD=.324) and instrumental text conditions (N=101, M=1.08, SD=.271), t (194) =18.73, p < .05.  By looking at a cross tabulation of the presented text (symbolic or instrumental) and the responses by the participants of the question above, it confirmed that participants exposed to a symbolic message perceive it to be more symbolic and participants exposed to an instrumental message perceive it to be more instrumental in nature.

Analysis and Hypotheses
Two research questions were used as foundation for this study, including:
RQ1: Do attitudes about Internet advertising and Internet motives affect consumers’ attitude toward a commercial website?
RQ2: Do Internet marketing messages such as symbolic and instrumental texts, on a website’s homepage, affect consumers’ attitudes toward a commercial website?

Pearson correlations were calculated for variables in the study.  The dependent variables were analyzed for correlations.  There was a strong positive relationship between attitude toward the website and behavior intention, r = .805, n = 188, p =.000.  There was a positive correlation between attitude toward the website and brand attitude, r = .716, n = 187, p =.000.  There was a positive correlation between behavior intention and brand attitude, r = .637, n = 186, p =.000.

The Pearson correlation between website text and attitude toward the brand, attitude toward the website, and behavior intention showed a weak relationship between the variables.  There was a positive correlation between website text and attitude toward the brand, r = .202, n = 189, p = .005.  There was a positive correlation between website text and attitude toward the website, r = .270, n = 192, p = .000.  There was a positive correlation between website text and behavior intention, r = .191, n = 190, p = .008.
The next correlations relate to whether Internet advertising attitude and commercial Internet usage relate to attitude toward the brand, attitude toward the website, and behavior intention.  There was a positive correlation between online advertising attitude and attitude toward the brand, r = .283, n = 175, p = .000.  There was a positive correlation between online advertising attitude and attitude toward the website, r = .345, n = 178, p = .000.  There was a positive correlation between online advertising attitude and behavior intention, r = .344, n = 176, p = .000.  There was not a correlation between commercial Internet usage and attitude toward the brand, r = .122, n = 187, p = .095.  There was not a correlation between commercial Internet usage and attitude toward the website, r = .026, n = 190, p = .718.  There was not a correlation between commercial Internet usage and behavior intention, r = .031, n = 188, p = .669.

Hypothesis 1
H1a:  Participants who report high positive attitudes towards Internet advertising (as measured by perceived informativeness, perceived entertainment, and perceived credibility) will show a preference for symbolic website texts, as measured by attitudes towards the brand, attitudes toward the website, and behavioral intention.

Using select cases, participants who were in the high positive attitude towards Internet advertising group were selected.  Analysis of Variance was used to test the hypothesis.  There was not a significant effect of the message text on attitude toward the website at the p <.05 level [F (1, 86) = 3.558, p = 0.063].  There was not a significant effect of the message text on attitude toward the brand at the p <.05 level [F (1, 86) = 1.82, p = 0.181].  There was not a significant effect of the message text on behavior intention at the p <.05 level [F (1, 86) = 2.84, p = 0.093].  Based on the analysis in this study, the researcher did not find empirical evidence indicating a relationship between Internet advertising attitudes and website texts and attitudes towards the brand, attitudes toward the website, and behavioral intention.

Hypothesis 2
H2a:  Participants with negative attitudes toward Internet advertising will prefer the instrumental website text over the symbolic text, as measured by attitude toward the brand, attitude toward the website, and behavioral intention.
Using select cases, participants who were in the less positive attitude towards Internet advertising group were chosen.  Analysis of Variance was used to test the hypothesis.  There was not a significant effect of the message text on attitude toward the brand at the p <.05 level [F (1, 85) = 3.514, p = 0.064].  There was not a significant effect of the message text on behavior intention at the p <.05 level [F (1, 86) = 0.450, p = 0.504].  Based on the analysis in this study, the researcher did not find empirical evidence indicating a relationship between Internet advertising attitudes and website texts and attitudes towards the brand and behavioral intention.
However, there was a significant effect of the message text on attitude toward the website at the p <.05 level [F (1, 86) = 5.415, p = 0.022].  Participants who had less positive attitudes towards Internet advertising and were shown the symbolic text had the following scores for attitude towards the website (M=10.857, SD = 3.76).  Participants who had less positive attitudes towards Internet advertising and were shown the instrumental text had the following scores for attitude towards the website (M=12.63, SD = 3.411).  These results indicated empirical support from participants with less positive attitudes towards Internet advertising have a better attitude towards a website when an instrumental text is used.

Hypothesis 3
H3a:  Participants that report a high usage of shopping and gathering information related to commercial products will show a more positive attitudinal response, as measured by attitude toward the brand, attitude toward the website, and behavioral intention, towards instrumental website texts.

Using select cases, participants in the high commercial Internet usage group were chosen.  Analysis of Variance was used to test the hypothesis.  Among higher commercial Internet users there was not a significant effect of the message text on attitude toward the brand at the p <.05 level [F (1, 101) = 5.806, p = 0.018].  Among higher commercial Internet users there was a significant effect of the message text on attitude toward the website at the p <.05 level [F (1, 101) = 3.566, p = 0.066].  Among higher commercial Internet users there was a significant effect of the message text on behavior intention at the p <.05 level [F (1, 102) = 6.126, p = 0.015].  Based on the analysis in this study, the researcher found empirical evidence indicating a relationship between commercial Internet usage and website texts, and attitudes towards the brand and behavioral intention.

Further analysis on the relationship with brand attitude was conducted.  Participants with high commercial Internet usage who were shown the symbolic text have the following scores for attitude towards the brand (M=13.790, SD = 5.612).  Participants with high commercial Internet usage who were shown the instrumental text have the following scores for attitude towards the brand (M=16.467, SD = 5.519).  This supported the hypothesis that participants that use the Internet for commercial activity prefer the instrumental website text.
Further analysis of the relationship on behavior intention was conducted.  Participants with high commercial Internet usage who were shown the symbolic text have the following scores for behavior intention (M=14.705, SD = 4.889).  Participants with high commercial Internet usage who were shown the instrumental text have the following scores for behavior intention (M=17.250, SD = 5.385).  These results supported the hypothesis that participants that use the Internet for commercial activity prefer the instrumental website text.

Hypothesis 4
H4a:  Participants that report a high usage of shopping and gathering information related to commercial products and a high rating for positive attitudes towards Internet advertising will show the strongest preference for instrumental website texts.
A three-way multivariate analysis of commercial Internet usage, website text, and attitudes of online advertising was conducted.  Box’s Test of Equality of Covariance indicated the results are significant at the p <.05 level, [F (42, 29,778) = 1.44, p = .035].  Wilks’ Lambda was calculated for the independent variables of website text [A= .933, F (3, 159) = 3.817, p = .011], Internet advertising attitude [A = .893, F (3,159) = 6.325, p =.105, and commercial Internet usage [.980, F (3,159) = 1.089, p = .356].  The website text was found to be the only significant variable.
There was some support for this hypothesis, but the results were not significant.  The results can be seen in Table 3.4.  For attitude toward the website, the highest mean value for instrumental message preference was from participants that had low commercial Internet usage and positive views of Internet advertising (M=14.90, SD = 3.56).  For attitude toward the brand, the highest mean value supported the hypothesis that high commercial users with positive Internet advertising attitudes preferred instrumental texts (M = 17.41, SD=5.41).  For behavior intention, the highest mean value supported the hypothesis that high commercial users with positive Internet advertising attitudes preferred instrumental texts (M = 19.00, SD=4.89).

t3

Summary
Chapter 4 reported the results from analyses performed to address the research questions and hypotheses of this study.  A summary of the results can be seen in Table 3.5.  For Hypothesis 1, based on the analysis in this study, the researcher did not find empirical evidence indicating a statistically significant relationship between Internet advertising attitudes and website texts and attitudes towards the brand, attitudes toward the website, and behavioral intention.  For Hypothesis 2, the results indicated empirical support that participants with less positive attitudes towards Internet advertising have a better attitude toward a website when an instrumental text is used.  For Hypothesis 3, the results supported the hypothesis that participants that used the Internet for commercial activity prefer the instrumental website text.  For Hypothesis 4, the results did not find statistical significance in the relationship of Internet commercial usage, online advertising attitude, and website texts with brand attitudes, website attitudes, and behavior intention.  The next chapter discusses these findings as well as recommendations for business practice, theory, and future research.

t4