Introduction (~1 points)

Introduce the research problem briefly (literature review is not required)

The research question proposed is to understand the main factors influencing the perceived usefulness of Wikipedia within teaching. Wikipedia can become a great learning tool within the classroom and provides little to no barriers for engagement. Wikipedia could provide the perceived usefulness and ease of use accomplishing all required elements for th Technology Acceptance Model. This SEM model will discover which factors influences the perceived usefulness, perceived ease, and behavioral intention leading to use behavior.

Data (~2 points)

Describe the data and variables

The codebook is listed blow. The variables are separated under the constructs measured. Job Relevance (JR1) My university promotes the use of open collaborative environments in the Internet (JR2) My university considers the use of open collaborative environments in the Internet as a teaching merit

Sharing Attitude (SA1) It is important to share academic content in open platforms (SA2) It is important to publish research results in other media than academic journals or books (SA3) It is important that students become familiar with online collaborative environments

Social Image (IM1) The use of Wikipedia is well considered among colleagues (IM2) My colleagues use Wikipedia

Quality (QU1) Articles in Wikipedia are reliable (QU2) Articles in Wikipedia are updated (QU3) Articles in Wikipedia are comprehensive

Perceived Enjoyment
(ENJ1) The use of Wikipedia stimulates curiosity (ENJ2) The use of Wikipedia is entertaining

Perceived Usefulness (PU1) The use of Wikipedia makes it easier for students to develop new skills (PU2) The use of Wikipedia improves student’s learning (PU3) Wikipedia is useful for teaching Perceived ease of use

Perceived ease of Use (PEU1) Wikipedia is user-friendly (PEU2) It is easy to find in Wikipedia the information you seek

Behavioral Intention (BI1) In the future I will recommend the use of Wikipedia to my colleagues and students (BI2) In the future I will use Wikipedia in my teaching activity

Use Behavior (USE1) I use Wikipedia to develop my teaching materials (USE2) I use Wikipedia as a platform to develop educational activities with students (USE3) I recommend my students to use Wikipedia (USE4) I recommend my colleagues to use Wikipedia

Describe the data cleaning and/or preparation process, if any (e.g., check for non-normality, data transformation, handling of missing values, etc.)

The data was imported through R and the researcher identified missing data within the dataframe. The correlation and covaraince matrix was created through pairwise deletion. Pairwise deletion was used in order for the incomplete cases to still be used in other analysis across the matrix. If listwise deletion was use, each case would be dropped from the analysis due to one missing value. That would result in < 200 cases instead of 850 cases.

Collinearity

All eigenvalues are positive and no out-of-bounds correlation/covariance is identified. The determinant of the covariance matrix is close to zero (1.175676e-11). With all variables measured, the skewness and kurtosis ranges between -2 and +2 providing acceptable normality asymmetry.

Provide relevant descriptive statistics

A correlation plot and descriptive statistics are provided below. The darker the shade, the higher the correlation between the two variables. The descriptive statistics provide the skewness and kurtosis values as mentioned above.

Table 1. Descriptive Statistics of the Variables
	n	Mean	Std.Dev	Median	Min	Max	25th	75th	Skew	Kurtosis
JR1	700	3.731429	1.0734863	4	1	5	3	5	-0.6229416	-0.3060293
JR2	700	3.148571	1.2035644	3	1	5	2	4	-0.1832194	-0.7897979
SA1	700	4.210000	0.8218493	4	1	5	4	5	-0.8516367	0.2900886
SA2	700	4.138571	0.9520391	4	1	5	4	5	-1.0433857	0.5875612
SA3	700	4.402857	0.7699602	5	1	5	4	5	-1.1848639	0.9350976
IM1	700	2.488571	0.9782386	2	1	5	2	3	0.3844644	-0.2017729
IM2	700	3.307143	1.0295799	3	1	5	3	4	-0.2068502	-0.6664320
PU1	700	3.162857	1.0045920	3	1	5	2	4	0.0342706	-0.5995319
PU2	700	3.145714	0.9769424	3	1	5	2	4	0.0278848	-0.5188323
PU3	700	3.455714	1.0506672	3	1	5	3	4	-0.1233836	-0.7211264
QU1	700	3.188571	0.8517495	3	1	5	3	4	-0.2301999	-0.2802205
QU2	700	3.427143	0.8417608	3	1	5	3	4	-0.2304733	-0.1491299
QU3	700	2.970000	0.8578260	3	1	5	2	4	0.0980200	-0.2736499
ENJ1	700	3.824286	0.9443660	4	1	5	3	4	-0.6523444	0.1154758
ENJ2	700	3.855714	0.8612025	4	1	5	3	4	-0.4305264	-0.2968724
PEU1	700	4.342857	0.7804059	5	1	5	4	5	-1.1303995	1.0896155
PEU2	700	4.067143	0.8163602	4	1	5	4	5	-0.6119256	-0.0570434
BI1	700	2.964286	1.0331467	3	1	5	2	4	0.0245769	-0.4420980
BI2	700	2.998571	1.0744787	3	1	5	2	4	0.0028313	-0.5916774
USE1	700	2.084286	1.0617738	2	1	5	1	3	0.7415204	-0.2102942
USE2	700	1.835714	1.0496318	1	1	5	1	2	1.1089226	0.3821330
USE3	700	2.658571	1.1703285	3	1	5	2	4	0.1564164	-0.9260192
USE4	700	2.552857	1.1602897	3	1	5	2	3	0.2340526	-0.9019232

Methods (~8 points)

• State the research hypotheses (Note. This section typically goes in Introduction and then you collect the data. But since you are not collecting data for this project, come up with research hypotheses based on the available data)

o In SEM, each effect of interest (e.g., a direct effect or a mediation effect) can be stated as a hypothesis

Hypotheses

Hypothesis 1 Job relevance of Wikipedia determines the social image.

Hypothesis 2 Job relevance will impact the sharing attitude of using Wikipedia

Hypothesis 3 Social image will determine the behavioral intention of using Wikipedia

Hypothesis 4 The sharing attitude of using Wikipedia will impact the behavioral intention

Hypothesis 5 The social image (H5a) and behavioral intention (H5b) will determine the quality of Wikipedia use

Hypothesis 6 The sharing attitude (H6a) and behavioral intention (H6b) will determine the perceived enjoyment

Hypothesis 7 Behavioral intention of using Wikipedia will impact the perceived usefulness

Hypothesis 8 The quality of Wikipedia will impact the perceived ease of use

Hypothesis 9 The perceived enjoyment of using Wikipedia will affect the use behavior

Hypothesis 10 The perceived usefulness of Wikipedia will directly impact the use behavior

Hypothesis 11 The perceived ease of use will determine the use behavior of using Wikipedia

Model A (based on the research hypotheses)

o Specify the model based on the research hypotheses
▪ Note that your constructs can be either formative or reflective (provide justification) o Provide the equation form
▪ Make sure to introduce each notation in the text or in a table. Do not forget to include the covariance structure and to mention the parameters that are constrained (e.g., fixed to 1) o Provide a well-labeled path diagram of the model o Check for identification (show steps)

Equation form:

\(SI = \beta_{jr/si}H1 + \zeta_{si}\)

\(SA = \beta_{jr/sa}H2 + \zeta_{sa}\)

\(BI = \beta_{si/bi}H3 + \beta_{sa/bi}H4 + \zeta_{bi}\)

\(QU = \beta_{bi/qu}H5b + \beta_{si/qu}H5a + \zeta_{qu}\)

\(PE = \beta_{sa/pe}H6a + \beta_{bi/pe}H6b + \zeta_{pe}\)

\(PU = \beta_{bi/pu}H7 + \zeta_{pu}\)

\(PEU = \beta_{qu/peu}H8 + \zeta_{peu}\)

\(UB = \beta_{peu/ub}H11 + \beta_{pu/ub}H9 + \beta_{pe/ub}H10 + \zeta_{ub}\)

Identification

Data observed : 9 variables p = 9*(9+1)/2 = 45

Number of models: 22 = q q = # of direct effects + # variances to estimate + # covariances to estimate q = 13 + 9 + 0 = 22

Model degrees of freedom = 45 - 22 = 23

The model degrees of freedom is 23 and is identified since it meets the > 0 cutoff.

First Model Fit Indices

The Chi-square test is not significant (p = 0.00 < 0.05), supporting that the residual covariance matrix is not statistically significantly different than 0. We will tentatively reject the model.

The RMSEA (0.08) is greater than 0.05 with a 90% confidence interval of 0.076-0.085. The model does not meet the suggested cutoff of <.05.

The SRMR (0.083) was found less than 0.10, supporting that differences between the observed and predicted correlations are relatively small. The model is accepted with the SRMR score.

The incremental fit indices, CFI (0.885 < 0.95) and TLI (0.866 < 0.90), are not within the model cutoffs. The model meets only one of the five fit indices providing the acceptable ranges. Respecification will occur.

The AIC (36976.020) and BIC (37244.533). This will be compared to Model B.

Model B (alternative model)

o Specify an alternative model to test if Model A has a poor fit ▪ State the changes in research hypotheses corresponding to this model o Provide a well-labeled path diagram of the model o Check for identification (show steps) • Describe the methods that you plan to use to test the reliability and validity of the latent constructs and to evaluate the model fit of the SEM. Provide the reference value for each measure.

Equation form:

\(SI = \beta_{jr/si}H1 + \zeta_{si}\)

\(SA = \beta_{jr/sa}H2 + \beta_{si/sa}H12 + \zeta_{sa}\)

\(BI = \beta_{si/bi}H3 + \beta_{sa/bi}H4 + \beta_{jr/bi}H13 + \zeta_{bi}\)

\(QU = \beta_{bi/qu}H5b + \beta_{si/qu}H5a + \zeta_{qu}\)

\(PE = \beta_{sa/pe}H6a + \beta_{bi/pe}H6b + \zeta_{pe}\)

\(PU = \beta_{bi/pu}H7 + \beta_{qu/pu}H15 + \zeta_{pu}\)

\(PEU = \beta_{qu/peu}H8 + \beta_{pu/peu}H14 + \zeta_{peu}\)

\(UB = \beta_{peu/ub}H11 + \beta_{pu/ub}H9 + \beta_{pe/ub}H10 + \zeta_{ub}\)

Alternative Hypothsis for Respecification

To respecify, the alternative hypotheses will be added to the model.

Hypothesis 12 The social image will impact the sharing attitude of Wikipedia use

Hypothesis 13 The job relevance will affect the behavior intention

Hypothesis 14 The behavioral intention will influence the perceived ease of use Wikipedia provides

Hypothesis 15 The quality will impact the perceived usefulness of Wikipedia use

Indentification

Data observed : 9 variables p = 9*(9+1)/2 = 45

Number of models: 22 = q q = # of direct effects + # variances to estimate + # covariances to estimate q = 17 + 9 + 0 = 26

Model degrees of freedom = 45 - 26 = 17

The model degrees of freedom is 17 and is identified since it meets the > 0 cutoff.

Second model

The Chi-square test is not significant (p = 0.00 < 0.05), supporting that the residual covariance matrix is not statistically significantly different than 0. We will tentatively reject the model.

The RMSEA (0.079) is greater than 0.05 with a 90% confidence interval of 0.074-0.083. The model does not meet the suggested cutoff of <.05. This model is rejected under this fit test.

The SRMR (0.077) was found less than 0.10, supporting that differences between the observed and predicted correlations are relatively small. The model is accepted with the SRMR score.

The incremental fit indices include CFI (0.89 < 0.95) and TLI (0.87 < 0.90). The CFI is not within the model cutoff > .95 nor does TLI with the cutoff of > .90. The model meets only one of the five fit indices providing the acceptable ranges.

The AIC (36915.066) and BIC (37201.784). This will be compared to the first model.

Results (~4 points)

• For both Model A and Model B: o Discuss the reliability and validity of the latent constructs (measurement model; this could be same for both the models depending on your alternative model)
o Estimate the model and discuss fit indices
• Select one model to present the results (provide justification) • Provide estimation results of the structural model of the selected model

Reliability and Validity.

All eigenvalues are positive and no out-of-bounds correlation/covariance is identified. The determinant of the covariance matrix is close to zero (1.175676e-11). Skewness and kurtosis range between -2 and +2. Cronbach’s Alpha for the data is 0.90 supporting that the data is acceptable. The eigenvalues, determinant of the covaraince matrix, skewness, kurtosis, and Cronbach’s Alpha support that the data is acceptable.

The two models were compared with the fit indices as referenced in the above table. The second model with the alternative hypotheses will be accepted as it is a more acceptable fit. Model B is supported by the RMSEA (.079), AIC(36915.066), and BIC (37201.784) scores. The standardized estimation results were used in the model below.

Discussion & Conclusions (~3 points)

• Interpret the results in the context of research hypotheses • Propose one new item (question/indicator) for each latent construct that you think would increase the reliability and validity of the construct

Overall, 13 hypotheses passed while two were not supported. In the results, H1, H2, H3, H4, H5, H6, H7, H8, H9, H11, H12, H13, and H15 passed while H10 and H14 failed. The researchers found that Social Image (H3 .4) has a larger impact on Behavioral Intention compared to other direct affects from Job Relevance (H13 .11) and Sharing Attitude (H4 .31). The factor of Perceived Usefulness had a larger impact from Behavioral Intention (H7 .69) compared to Quality (H15 .29). Applying the theory of Technology Acceptance Model and the three factors of activation, Perceived Usefulness, Perceived Enjoyment, and Perceived Ease of Use, the researchers found that Perceived Usefulness (H9 .85) had a positive significant impact on use behavior. Furthermore, the Perceived Ease of Use (H11 -.09) and Perceived Enjoyment (H10 -.03)

Passed Hypotheses