Results

            Based on the ”°Mean Questionnaire Responses”± graph, it can be generally observed that both the indirect and direct avatar gaze conditions are an improvement from chat only conditions, for all four categories of the quality of communication.

            When comparing the direct avatar gaze, and indirect avatar gaze condition, we once again witness the same pattern. However the difference of improvement in quality of communication is less noticeable.

"Mean Questionnaire Responses Graph"


"Mean Questionnaire Responses Table of Values"

            Although the mean questionnaire responses seem to prove the hypothesis true, the null hypothesis must be negated for each category designated to describe the quality of communication.

            For the ”°Face to Face”± category the results negate the null hypothesis and is significant because all four tests result in a p value that is less than 0.05. (F = 23.292, p = 0.000) shows that there is a significant difference between the mean values of the different conditions. This consequently means that at least one average is different from the others.

Face to Face Tests of Within-Subjects Effects

Measure:MEASURE_1

 

 

 

 

 

Source

Type III Sum of Squares

df

Mean Square

F

Sig.

faceface

Sphericity Assumed

34.903

2

17.451

23.292

.000

Greenhouse-Geisser

34.903

1.677

20.817

23.292

.000

Huynh-Feldt

34.903

1.756

19.876

23.292

.000

Lower-bound

34.903

1.000

34.903

23.292

.000

            For the ”°Involvement”± category the results negate the null hypothesis and is significant because the highest p value is 0.013, and that is less than 0.05. The results are also strengthened because this value is for the conservative Lower-bound test. (F = 6.820, p = 0.013) shows that there is a significant difference between the mean values of the different conditions.

Involvement Tests of Within-Subjects Effects

Measure:MEASURE_1

 

 

 

 

 

Source

Type III Sum of Squares

df

Mean Square

F

Sig.

involvement

Sphericity Assumed

15.358

2

7.679

6.820

.002

Greenhouse-Geisser

15.358

1.616

9.504

6.820

.004

Huynh-Feldt

15.358

1.687

9.104

6.820

.004

Lower-bound

15.358

1.000

15.358

6.820

.013

            For the ”°Co-Presence”± category the results negate the null hypothesis and is significant because all four test results in a p value that is less than 0.05. (F = 32.768, p = 0.000) shows that there is a significant difference between the mean values of the different conditions.

Co-Presence Tests of Within-Subjects Effects

Measure:MEASURE_1

 

 

 

 

 

Source

Type III Sum of Squares

df

Mean Square

F

Sig.

copresence

Sphericity Assumed

81.711

2

40.855

32.768

.000

Greenhouse-Geisser

81.711

1.706

47.890

32.768

.000

Huynh-Feldt

81.711

1.790

45.652

32.768

.000

Lower-bound

81.711

1.000

81.711

32.768

.000

            For the ”°Partner Evaluation”± category the results negate the null hypothesis and is significant because the highest p value is 0.005, and this is still less than 0.05. (F = 9.004, p = 0.005) shows that there is a significant difference between the mean values of the different conditions.

Partner Tests of Within-Subjects Effects

Measure:MEASURE_1

 

 

 

 

 

Source

Type III Sum of Squares

df

Mean Square

F

Sig.

partner

Sphericity Assumed

14.398

2

7.199

9.004

.000

Greenhouse-Geisser

14.398

1.216

11.841

9.004

.003

Huynh-Feldt

14.398

1.238

11.632

9.004

.003

Lower-bound

14.398

1.000

14.398

9.004

.005

            Thus all categories have resultantly proved that there is a significant difference between the mean values of the different conditions.

            Now pair wise comparisons must be made to determine whether specific differences between conditions are significant. This means that each condition will be compared to all the other different conditions for significance. After running the pair wise comparisons we find that at a 0.05 significance level, the comparison of direct and indirect conditions are not lower for any of the four different categories of quality of communication. Thus, at the 0.05 significance level we do not have valid evidence to refute the null hypothesis for the comparison of direct and indirect gaze conditions. Thus it is more likely that a significant difference does not exist between these two relationships.

            However the comparison of chat-only to either direct or indirect gaze results in a significance that is less than 0.05 for all four categories. Thus the null hypothesis is negated, and it is very likely that there is a significant difference between the chat and gaze conditions.

***For the following tables 1 represents the chat condition, 2 represents the indirect gaze condition, and 3 represents the direct gaze condition.

Face to Face Pairwise Comparisons

Measure:MEASURE_1

””

””

””

””

(I) faceface

(J) faceface

Mean Difference (I-J)

Std. Error

Sig.a

95% Confidence Interval for Differencea

Lower Bound

Upper Bound

1

2

-1.054*

.165

.000

-1.471

-.637

3

-1.368*

.247

.000

-1.990

-.745

2

1

1.054*

.165

.000

.637

1.471

3

-.314

.210

.432

-.843

.215

3

1

1.368*

.247

.000

.745

1.990

2

.314

.210

.432

-.215

.843

Involvement Pairwise Comparisons

Measure:MEASURE_1

””

””

””

””

(I) involvement

(J) involvement

Mean Difference (I-J)

Std. Error

Sig.a

95% Confidence Interval for Differencea

Lower Bound

Upper Bound

1

2

-.721*

.229

.011

-1.299

-.142

3

-.897*

.314

.022

-1.688

-.106

2

1

.721*

.229

.011

.142

1.299

3

-.176

.218

1.000

-.727

.374

3

1

.897*

.314

.022

.106

1.688

2

.176

.218

1.000

-.374

.727

Co-Presence Pairwise Comparisons

Measure:MEASURE_1

””

””

””

””

(I) copresence

(J) copresence

Mean Difference (I-J)

Std. Error

Sig.a

95% Confidence Interval for Differencea

Lower Bound

Upper Bound

1

2

-1.588*

.238

.000

-2.189

-.987

3

-2.103*

.322

.000

-2.915

-1.291

2

1

1.588*

.238

.000

.987

2.189

3

-.515

.244

.127

-1.130

.100

3

1

2.103*

.322

.000

1.291

2.915

2

.515

.244

.127

-.100

1.130

Partner Pairwise Comparisons

Measure:MEASURE_1

””

””

””

””

(I) partner

(J) partner

Mean Difference (I-J)

Std. Error

Sig.a

95% Confidence Interval for Differencea

Lower Bound

Upper Bound

1

2

-.419*

.115

.003

-.709

-.129

3

-.919*

.280

.007

-1.626

-.212

2

1

.419*

.115

.003

.129

.709

3

-.500

.222

.094

-1.060

.060

3

1

.919*

.280

.007

.212

1.626

2

.500

.222

.094

-.060

1.060