2008 INFORMS Marketing Science Conference June 12 14

2008 INFORMS Marketing Science Conference June 12 - 14, 2008 Sheraton Vancouver Wall Centre Hotel, Vancouver, British Columbia, CANADA Cluster: Contributed Session Thursday June 12, TA 05 – Grand Ballroom C, 13: 00 - 14: 30 Issues in New Product Adoption One’s Innovativeness and Adoption Time of an Innovation* V. 1. 4 Masataka Yamada The Graduate School of Management Kyoto Sangyo University Motoyama, Kamigamo, Kita-ku, Kyoto-shi 603 -8555, Japan myamada@cc. kyoto-su. ac. jp * This study has been supported by Scientific Research (C) #19530394 of the Grant-in-Aid for Scientific Research, JSPS

Problem • In diffusion theory in marketing, there is no significant study on the concept of one’s innovativeness since Midgley and Dowling’s thorough work (1978). • They considered the innovativeness as one of personality traits instead of relative time of adoption which had been and still is believed as innovativeness. • They proposed intervening variables between the construct of innate innovativeness and the relative time of adoption. 2020/10/3 (C) Masataka Yamada 2

Problem (continued) • The intervening variables are “interest in product, ” “communicated experience, ” and “situational effects” (Rogers 1962). • Even though they suggested cross sectional method to measure the innate innovativeness, they did not present any measure. • Also, they did not propose any measure for the product category specific innovativeness. 2020/10/3 (C) Masataka Yamada 3

Problem (continued) • Fortunately, Goldsmith and Hofacker (1991) developed a scale for the product category specific innovativeness. • But they did not link their measure to forecast individual’s time of adoption. • Therefore, this study tries to investigate the innate innovativeness and the product category specific innovativeness and to forecast individual’s time of aoption. 2020/10/3 (C) Masataka Yamada 4

Order of Presentation 1. Objectives of Study 2. Previous Relevant Studies 3. Theoretical Framework　 4. Data 5. Results and Analyses 6. Conclusions 2020/10/3 (C) Masataka Yamada 5

1. The Objectives of This Study This is an exploratory study rather than a hypothesis-based study. • 1. To measure innovativeness as a construct instead of actualized adoption time. • 2. To find the relationship between innate innovativeness and product category specific innovativeness. • 3. To develop mathematical models to forecast one’s adoption time. 2020/10/3 (C) Masataka Yamada 6

2. Previous Relevant Studies • • • 2. 1 Midgley and Dowling (1978) 2. 2 Gatignon and Robertson(1985) skip 2. 3 Goldsmith and Hofacker (1991) 2. 4 Yamada and Zhu (2005) 2. 5 Diffusion patterns other than normal distribution skip 2020/10/3 (C) Masataka Yamada 7

2. 1. Midgley and Dowling (1978) • There was a thorough conceptual study done by Midgley and Dowling (1978). They defined that innovativeness is a personality trait possessed, to a greater or lesser degree, by all members of a society. • Innovativeness is the degree to which an individual makes innovation decisions independently of the communicated experience of others. 2020/10/3 (C) Masataka Yamada 8

2. 1. Midgley and Dowling (1978) (continued) • Time of adoption is some function of interest in product, communicated experience, and situational effects. They also separated innovativeness into conceptual innovativeness and actualized innovativeness. 2020/10/3 (C) Masataka Yamada 9

Figure B The New Model of Innovativeness Such as empathy, dogmatism, achievement motivation, self-monitoring, intelligence, etc. Such as social participation, social integration, cosmopolitism, social character, etc. This is included to indicate that most individuals do not have all-encompassing interests or activities, especially with respect to the purchase of goods or services. It is also to be expected that interest would be dependent on psychological, sociological, and demographic factors. Subsumes a variety of situation-specific and person-specific factors. 2020/10/3 (C) Masataka Yamada A shorthand way of denoting the network of interpersonal messages relating to the product and the effects of 10 these messages on individuals.

2. 3. Goldsmith and Hofacker (1991) • Goldsmith and Hofacker aimed at product category specific innovativeness. • They developed a six-item, self-report scale to measure innovativeness within a specific domain of interest familiar to the consumer. This six-item, self-report scale is aimed to be adaptable across product categories. 2020/10/3 (C) Masataka Yamada 11

Goldsmith and Hofacker (1991) continued • Although they seem to have attained this objective, they left the link between the construct and time of adoption untouched. • They found only the correlations between the construct and behavior: Rock Music Innovator (Record awareness, purchase, FM listening), Fashion Innovator, Cologne or Scent Innovator. 2020/10/3 (C) Masataka Yamada 12

Goldsmith and Hofacker’s sixitem, self-report scale Negative-direction Q 1: I am the least interested in among my friends. Q 2: Usually I am the last person who becomes aware of recent. Q 3: When a new appeared, often times, I was the last person who bought the new model among my friends. 2020/10/3 (C) Masataka Yamada 13

Goldsmith and Hofacker’s sixitem, self-report scale Positive-direction Q 4: If I were allowed to buy a new , then I would buy it immediately. Q 5: I am a kind of a person who buy a new without testing it by myself. Q 6: I know the launching date of a new before other people know it. 2020/10/3 (C) Masataka Yamada 14

Goldsmith and Hofacker’s sixitem, self-report scale • For all questions, respondents must choose one of the following answers: – I strongly agree with the statement. – I rather agree with the statement. – I am indifference. – I rather disagree with the statement. – I strongly disagree with the statement. 2020/10/3 (C) Masataka Yamada 15

2. 4. Yamada and Zhu (2005) • Donnelly (1973) : we can expect an individual’s social character to fall somewhere between the two extremes of complete inner- or other-direction. • Based on the above, they adopted Kiuchi’s Independent-Interdependent Scale because the number of items are smaller than Donnelly’s. • This construal of self measure is higher level of abstraction than Midgley and Dowling’s cross sectional approach. 2020/10/3 (C) Masataka Yamada 16

2. 4. Yamada and Zhu (2005) continued Kiuchi’s Independent-Interdependent Scale • Kiuchi’s Independent-Interdependent Scale (1995) is based on the construal of self of Markus and Kitayama (1991). • Kiuchi’s Independent-Interdependent Scale is suitable to measure the innate innovativeness defined as above. 2020/10/3 (C) Masataka Yamada 17

Plot of LTIME　vs. 　SCORE Legend: A = 1 obs, B = 2 obs, etc. Adoption time LTIME | 3 + | | | ABAE BC | FCB 2 + C | B LTIME=Ln(TIME) C CA A A FCB IJECLBADHCA A AC | A A I FFAFAA LCA FAGABC LCOC CC IGGACFF B B Pearson Correlation Coefficients, N = 759 Prob > |r| under H 0: Rho=0 B LIIIIFCFFMKGFHIBHECDCIE A | 1 + AC C BBDFBCJIHDBICKBBFFFEL FC LTIME SCORE LTIME 1. 00000 0. 71943 <. 0001 SCORE 0. 71943 <. 0001 1. 00000 | | CCFICCAH HLACNMDCMCM A A AA | | 0 + CCCABC LGBC EC C CF F | --+---------+---------+---------+---------10 20 Independent 2020/10/3 30 40 50 60 70 Interdependent SCORE (C) Masataka Yamada 18

Details of Classical Model 1 Results The REG Procedure Model: MODEL 1 Dependent Variable: LTIME PROC REG; MODEL LTIME=SCORE SEX 1 PRICE/STB; RUN; Analysis of Variance Sum of Mean DF Squares Square Source Model 3 138. 26937 46. 08979 Error 755 123. 74883 0. 16391 Corrected Total 758 262. 01820 F Value 281. 20 Pr > F <. 0001 Root MSE 0. 40485 R-Square 0. 5277 Dependent Mean 1. 27080 Adj R-Sq 0. 5258 Coeff Var 31. 85799 Parameter Estimates Variable Intercept SCORE SEX 1 PRICE 2020/10/3 DF Parameter Estimate 1 -0. 62456 1 0. 04625 1 -0. 24924 1 0. 00001530 Standard Error 0. 06968 0. 00163 0. 08468 0. 00000543 t Value -8. 96 28. 31 -2. 94 2. 82 Standardized Pr > |t| Estimate <. 0001 0. 0033 0. 0050 (C) Masataka Yamada The influence of SCORE is ten times stronger than other variables. 0 0. 71923 -0. 07423 0. 07103 19

Findings • Personality scores measured by Independent. Interdependent Scale are approximately normally distributed between the two extremes. • Personality scores have a strong relationship to adopters’ actualized innovativeness or their adoption times. • Other independent variables are relatively less significant. Because cellphone is regarded as an extremely necessary and fashion good so that price, income and demography become less relevant. 2020/10/3 (C) Masataka Yamada 20

3. Theoretical Framework 3. 1. The Structure of Our Study Necessary goods: Yamada and Zhu (2005) 2020/10/3 (C) Masataka Yamada 21

3. 2. Definition of Innovativeness • Midgley and Dowling’s Definition: Innovativeness is the degree to which an individual makes innovation decisions independently of the communicated experience of others. • Our Definition： Innovativeness is the degree to which an individual makes innovation adoption decisions independently of the communicated experience of others. 2020/10/3 (C) Masataka Yamada 22

4. Data • Data was collected on the college students at KSU from June 22 to July 28, 2007 by the internet questionnaire. • The products chosen for this study are game machines, specifically Nintendo’s DS, DS LITE and Wii. We believe that the college students are one of the most important segments for this product category. 2020/10/3 (C) Masataka Yamada 23

Dates of New Product Launch • DS: December 2, 2004 • DS LITE: March 2, 2006 • Wii: December 2, 2006 Data was collected on the college students at KSU from June 22 to July 28, 2007 by the internet questionnaire. 2020/10/3 (C) Masataka Yamada 24

Structure of Questionnaire • Q 1~Q 16: Kiuchi’s 16 -item construal-of-self test for innate innovativeness (see appendix 1). Q 17~Q 20: Criterion • Q 17: # of models aware Variables for Categoryspecific Innovativeness • Q 18: # of models adopted • Q 19: Frequency of media access • Q 20: Frequency of store visit • Q 21~Q 26: Goldsmith & Hofacker’s six-item test for category-specific innovativeness (see appendix 2). • Q 27~Q 31: Demographics 2020/10/3 (C) Masataka Yamada 25

Structure of Questionnaire (continued) • Q 33~Q 37: Criterion variables for innate innovativeness • Q 39: DS Time of adoption • Q 40: DS Time of adoption decision making • Q 41: Previous model owned before DS • Q 42: Cues for DS adoption • Q 43~Q 46: Same items as Q 39~Q 42 for DS LITE • Q 47~Q 50: Same items as Q 39~Q 42 for Wii 2020/10/3 (C) Masataka Yamada 26

5. Results and Analyses Adopter Segments 　　 DS 4 1 7 DS LITE 2 　　　　　 6 5 3 Wii 8 Non-adopters 2020/10/3 (C) Masataka Yamada 27

5. Results and Analyses (continued) 2020/10/3 (C) Masataka Yamada 28

5. Results and Analyses (continued) 2020/10/3 (C) Masataka Yamada 29

5. Results and Analyses (continued) 2020/10/3 (C) Masataka Yamada 30

5. 1. Results for Innovativeness Scores 2020/10/3 (C) Masataka Yamada 31

5. Results and Analyses (continued) PSCORE=Kiuchi’s independent-interdependent test Q 1~Q 16: Cronbach’s α The results are same level as ones of Yamada and Zhu (2005). Kiuchi’s Test is reliable. Variables α --------------Raw 0. 820190 Standardized 0. 824794 ISCORE=Goldsmith and Hofacker’s category-specific test Q 21~Q 26: Cronbach’s α Variables α --------------Raw 0. 780817 Standardized 0. 782199 2020/10/3 (C) Masataka Yamada 32

5. Results and Analyses (continued) The GLM Procedure Tukey's Studentized Range (HSD) Test for PSCORE NOTE: This test controls the Type I experimentwise error rate, but it generally has a higher Type II error rate than REGWQ. Alpha 0. 05 Error Degrees of Freedom 1069 Error Mean Square 45. 66905 Critical Value of Studentized Range 2. 77495 Minimum Significant Difference 0. 8678 Harmonic Mean of Cell Sizes 467. 0177 NOTE: Cell sizes are not equal. Means with the same letter are not significantly different. Tukey Grouping Mean N Q 30 A 37. 7056 727 2 B 36. 6599 344 1 Mean of male’s PSCORE is statistically significantly higher than mean of female’s PSCORE. 2020/10/3 (C) Masataka Yamada 33

5. Results and Analyses (continued) The GLM Procedure Tukey's Studentized Range (HSD) Test for ISCORE NOTE: This test controls the Type I experimentwise error rate, but it generally has a higher Type II error rate than REGWQ. Alpha 0. 05 Error Degrees of Freedom 1127 Error Mean Square 28. 94394 Critical Value of Studentized Range 2. 77479 Minimum Significant Difference 0. 6721 Harmonic Mean of Cell Sizes 493. 2861 NOTE: Cell sizes are not equal. Means with the same letter are not significantly different. Tukey Grouping Mean N Q 30 A 14. 9974 765 2 B 13. 5934 364 1 Mean of male’s ISCORE is statistically significantly higher than mean of female’s ISCORE. 2020/10/3 (C) Masataka Yamada 34

The GLM Procedure Level of CAT 1 2 3 4 5 6 7 8 2020/10/3 N 69 204 11 16 13 9 9 714 ------PSCORE-----Mean Std Dev 37. 2753623 36. 7647059 35. 2727273 36. 1875000 33. 0000000 40. 2222222 34. 5555556 37. 6750700 (C) Masataka Yamada 5. 96776650 6. 54002677 7. 64317878 6. 46239120 5. 75905085 8. 13599683 8. 66185764 6. 90149243 35

The GLM Procedure Dunnett's t Tests for PSCORE NOTE: This test controls the Type I experimentwise error for comparisons of all treatments against a control. Alpha 0. 05 Error Degrees of Freedom 1037 Error Mean Square 46. 09787 Critical Value of Dunnett's t 2. 68734 Comparisons significant at the 0. 05 level are indicated by ***. CAT Comparison 6 1 2 4 3 7 5 2020/10/3 - 8 8 8 8 Difference Between Means 2. 5472 -0. 3997 -0. 9104 -1. 4876 -2. 4023 -3. 1195 -4. 6751 Simultaneous 95% Confidence Limits -3. 5730 -2. 6999 -2. 3589 -6. 0998 -7. 9459 -9. 2397 -9. 7814 (C) Masataka Yamada 8. 6673 1. 9005 0. 5381 3. 1247 3. 1412 3. 0006 0. 4313 36

5. Results and Analyses (continued) The UNIVARIATE Variable: PSCORE= Procedure Innate innovativeness Schematic Plots | 65 + | | | 0 60 + 0 | | 0 0 | 0 55 + | 0 | | 0 0 | | 50 + 0 | | | | +-----+ | | | | 45 + | | | | | +-----+ | | | | | 40 + | | | +-----+ | | | | +-----+ | | | +-----+ *--+--* | | *--+--* +-----+ *-----* | | | | *--+--* *-----* | | +-----+ 35 + | | | | +--+--+ | | | *-----* | | +-----+ | | | +-----+ | +-----+ *-----* | | | 30 + +-----+ | | | | | | | | 25 + | | | | 0 | | | 20 + | | 0 | 0 * 15 + ------+-----------+-----------+-----------------+------+-----7 8 1 2 3 4 5 6 CAT | 65 + | | | 60 + | | | 55 + | | | 50 + | | | 45 + | | | 40 + | | | 35 + | | | 30 + | | | 25 + | | | 20 + | | | 15 + CAT There are no significant differences among the means. Innate innovativeness is approximately equal for each 0 segment. 0 2020/10/3 (C) Masataka Yamada 37

The GLM Procedure Level of CAT 1 2 3 4 5 6 7 8 2020/10/3 N 74 211 12 17 15 8 10 756 ------ISCORE-----Mean Std Dev 16. 2567568 17. 0094787 15. 8333333 17. 8235294 16. 8666667 18. 8750000 19. 0000000 13. 3306878 (C) Masataka Yamada 5. 42460354 5. 19522722 5. 89041337 6. 93933799 6. 30041571 4. 58062691 7. 16472842 5. 02382773 38

The GLM Procedure Dunnett's t Tests for ISCORE NOTE: This test controls the Type I experimentwise error for comparisons of all treatments against a control. Alpha 0. 05 Error Degrees of Freedom 1095 Error Mean Square 26. 65587 Critical Value of Dunnett's t 2. 68707 Comparisons significant at the 0. 05 level are indicated by ***. CAT Comparison 7 6 4 2 5 1 3 2020/10/3 - 8 8 8 8 Difference Between Means 5. 6693 5. 5443 4. 4928 3. 6788 3. 5360 2. 9261 2. 5026 Simultaneous 95% Confidence Limits 1. 2533 0. 6135 1. 0905 2. 5986 -0. 0814 1. 2363 -1. 5339 (C) Masataka Yamada 10. 0853 10. 4751 7. 8952 4. 7589 7. 1534 4. 6159 6. 5391 *** *** *** 39

Mean in each segment is statistically significantly Variable: ISCORE=Product Category higher than one in #8 except ones in #3 and #5. Specific Innovativeness Though not statistically significant, ISCORE Schematic Plots is increasing | | as the number of machines 30 + 0 | | | 30 + owned increases | | | The UNIVARIATE Procedure 27. 5 25 22. 5 20 17. 5 15 12. 5 10 7. 5 5 | | + | | | + | | | + | | | + 0 | | | | +-----+ | | | *--+--* | | | | +-----+ | | | | | | | +-----+ | | | | *--+--* | | | | | +-----+ | | | | | | | +-----+ | | | + | | | | *-----* | | +-----+ | | | | +-----+ | | | | | *-----* | + | | | | | | | | | | +-----+ | | | | | *-----* | + | | | | +-----+ | | | | | 27. 5 | | | +-----+ | | | | *--+--* +-----+ | | | 25 22. 5 20 17. 5 15 12. 5 0 10 7. 5 2020/10/3 (C) Masataka Yamada ------+-----------+-----------+------+-----CAT 1 2 3 4 5 6 5 CAT | | + | | | + | | | + | | | +-----+ | | | | *--+--* | | | +-----+ | | | | | | | | | +-----+ | | | | | *--+--* | | | +-----+ | | ------+------+-----7 8 40

5. 2. Results for the Relationship between Innate Innovativeness and Product Category Specific Innovativeness 2020/10/3 (C) Masataka Yamada 41

#7 The Highest Interest in Product Category Plot of PSCORE*ISCORE. Only segment #7 indicates some correlation between the two scores. Legend: A = 1 obs, B = 2 obs, etc. PSCORE | 50 + A | | | A | | 40 + | | A A A | 30 + A | | | | A 20 + --+-----------+-----------+------+-5 10 15 20 25 30 Pearson Correlation Coefficients Prob > |r| under H 0: Rho=0 Number of Observations PSCORE ISCORE 1. 00000 0. 60355 0. 0853 9 9 ISCORE 0. 60355 0. 0853 9 1. 00000 10 NOTE: 1 obs had missing values. 2020/10/3 (C) Masataka Yamada 42

Correlation between PSCORE and ISCORE • No segment indicates statistically significant correlation between the two scores except segment #7. • We may say that the correlation between innate innovativeness and product specific innovativeness becomes stronger as the number of innovations adopted increases. • But there is no strong relationship between the two constructs in general (Midgley and Dowling 1978; Goldsmith and Hofacker 1999).

Summary of Findings for Innovativeness • We had eight segments based on the brand number of game machines owned. • The innate innovativeness score (PSCORE) and the product category specific innovativeness score (ISCORE) are approximately one dimensional and normally distributed (uni-modal). • Mean of male’s PSCORE is statistically significantly higher than mean of female’s PSCORE. 2020/10/3 (C) Masataka Yamada 44

Summary of Findings for Innovativeness (continued) • Mean of male’s ISCORE is statistically significantly higher than mean of female’s ISCORE. • Mean of PSCOREs in each segment is approximately equal. • Mean of ISCOREs in each segment is statistically significantly higher than one in #8 except ones in #3 and #5. 2020/10/3 (C) Masataka Yamada 45

Summary of Findings for Innovativeness (continued) • Though not statistically significant, ISCORE is increasing as the number of machines owned increases • It is confirmed that there is no strong correlation between the two constructs in general. • But we found some correlation on #7. 2020/10/3 (C) Masataka Yamada 46

5. 3. Results for Adoption Time 2020/10/3 (C) Masataka Yamada 47

The GLM Procedure Level of CAT 1 4 6 7 2020/10/3 N 74 17 9 10 -------DST------Mean Std Dev 13. 3918919 11. 7058824 8. 4444444 11. 2000000 (C) Masataka Yamada 8. 7629631 10. 3184586 8. 0017359 12. 3899601 48

The GLM Procedure Dunnett's t Tests for DST NOTE: This test controls the Type I experimentwise error for comparisons of all treatments against a control. Alpha 0. 05 Error Degrees of Freedom 106 Error Mean Square 86. 82063 Critical Value of Dunnett's t 2. 34335 Comparisons significant at the 0. 05 level are indicated by ***. CAT Comparison 1 4 6 2020/10/3 - 7 - 7 Difference Between Means 2. 192 0. 506 -2. 756 (C) Masataka Yamada Simultaneous 95% Confidence Limits -5. 165 -8. 196 -12. 788 9. 548 9. 208 7. 277 49

The GLM Procedure Level of CAT 2 4 5 7 2020/10/3 N 212 17 15 10 -------DSLT------Mean Std Dev 9. 26415094 8. 41176471 7. 20000000 5. 80000000 (C) Masataka Yamada 4. 32846171 5. 73431364 5. 15751878 4. 51663592 51

The GLM Procedure Dunnett's t Tests for DSLT NOTE: This test controls the Type I experimentwise error for comparisons of all treatments against a control. Alpha Error Degrees of Freedom Error Mean Square Critical Value of Dunnett's t 0. 05 250 20. 1413 2. 30072 Comparisons significant at the 0. 05 level are indicated by ***. CAT Comparison 2 4 5 2020/10/3 - 7 - 7 Difference Between Means 3. 4642 2. 6118 1. 4000 Simultaneous 95% Confidence Limits 0. 1229 -1. 5032 -2. 8153 (C) Masataka Yamada 6. 8054 *** 6. 7267 5. 6153 52

The GLM Procedure Level of CAT 3 5 6 7 2020/10/3 N 12 15 9 10 -------Wii. T------Mean Std Dev 2. 83333333 2. 77777778 1. 70000000 (C) Masataka Yamada 1. 94624736 1. 49602648 2. 22361068 1. 25166556 54

Dunnett's t Tests for Wii. T NOTE: This test controls the Type I experimentwise error for comparisons of all treatments against a control. Alpha 0. 05 Error Degrees of Freedom 42 Error Mean Square 3. 015608 Critical Value of Dunnett's t 2. 42832 Comparisons significant at the 0. 05 level are indicated by ***. CAT Comparison 3 6 5 2020/10/3 - 7 - 7 Difference Between Means 1. 1333 1. 0778 0. 6333 Simultaneous 95% Confidence Limits -0. 6722 -0. 8598 -1. 0882 (C) Masataka Yamada 2. 9389 3. 0153 2. 3549 55

DS Adoption time 2020/10/3 DSLite Adoption time (C) Masataka Yamada Wii Adoption time 57

Summary of Findings for Adoption Time • Mean of adoption times becomes faster as the number of game machines owned, namely, level of interest in product increases. 2020/10/3 (C) Masataka Yamada 58

Summary of Findings for Adoption Time (continued) • This corresponds to the fact that mean of ISCOREs increases as the number of game machines owned increases. • This means that adoption time becomes faster as product category specific innovativeness increases. 2020/10/3 (C) Masataka Yamada 59

5. 4. Mathematical Models to forecast one’s adoption time 2020/10/3 (C) Masataka Yamada 60

Segment #7 The REG Procedure Model: MODEL 1 Dependent Variable: LDST Analysis of Variance DF Sum of Squares Mean Square 1 7 8 7. 50600 12. 06198 19. 56799 7. 50600 1. 72314 Root MSE Dependent Mean Coeff Var 1. 31268 1. 25161 104. 87991 Source Model Error Corrected Total R-Square Adj R-Sq F Value Pr > F 4. 36 0. 0753 0. 3836 0. 2955 Parameter Estimates Variable Label Intercept PSCORE 2020/10/3 DF Parameter Estimate Standard Error t Value Pr > |t| Standardized Estimate 1 1 5. 11587 -0. 11183 1. 90250 0. 05358 2. 69 -2. 09 0. 0311 0. 0753 0 -0. 61934 (C) Masataka Yamada 61

Segment #7 The REG Procedure Model: MODEL 4 Dependent Variable: LDST Analysis of Variance DF Sum of Squares Mean Square 1 7 8 7. 27070 12. 29729 19. 56799 7. 27070 1. 75676 Root MSE Dependent Mean Coeff Var 1. 32543 1. 25161 105. 89796 Source Model Error Corrected Total R-Square Adj R-Sq F Value Pr > F 4. 14 0. 0814 0. 3716 0. 2818 Parameter Estimates Variable Label Intercept ISCORE 2020/10/3 DF Parameter Estimate Standard Error t Value Pr > |t| Standardized Estimate 1 1 3. 65196 -0. 12560 1. 25990 0. 06174 2. 90 -2. 03 0. 0230 0. 0814 0 -0. 60956 (C) Masataka Yamada 62

Segment #7 The REG Procedure Model: MODEL 7 Dependent Variable: LDST Analysis of Variance DF Sum of Squares Mean Square 2 6 8 9. 33599 10. 23200 19. 56799 4. 66799 1. 70533 Root MSE Dependent Mean Coeff Var 1. 30588 1. 25161 104. 33657 Source Model Error Corrected Total R-Square Adj R-Sq F Value Pr > F 2. 74 0. 1430 0. 4771 0. 3028 Parameter Estimates Variable Label Intercept PSCORE ADPT* Intercept PSCORE ADPT * DF Parameter Estimate Standard Error t Value Pr > |t| Standardized Estimate 1 1 1 5. 59367 -0. 09641 -0. 10219 1. 94803 0. 05534 0. 09865 2. 87 -1. 74 -1. 04 0. 0284 0. 1321 0. 3402 0 -0. 53395 -0. 31751 ADPT=# of game machines owned 2020/10/3 (C) Masataka Yamada 63

Segment #7 The REG Procedure Model: MODEL 10 Dependent Variable: LDST Analysis of Variance DF Sum of Squares Mean Square 2 6 8 7. 37384 12. 19415 19. 56799 3. 68692 2. 03236 Root MSE Dependent Mean Coeff Var 1. 42561 1. 25161 113. 90218 Source Model Error Corrected Total R-Square Adj R-Sq F Value Pr > F 1. 81 0. 2420 0. 3768 0. 1691 Parameter Estimates Variable Label Intercept ISCORE ADPT 2020/10/3 DF Parameter Estimate Standard Error t Value Pr > |t| Standardized Estimate 1 1 1 3. 65988 -0. 14735 0. 04124 1. 35558 0. 11719 0. 18305 2. 70 -1. 26 0. 23 0. 0356 0. 2553 0. 8292 0 -0. 71512 0. 12812 (C) Masataka Yamada 64

Segment #7 The REG Procedure Model: MODEL 8 Dependent Variable: LDSLT Analysis of Variance DF Sum of Squares Mean Square 2 6 8 2. 43984 5. 70512 8. 14496 1. 21992 0. 95085 Root MSE Dependent Mean Coeff Var 0. 97512 1. 49554 65. 20180 Source Model Error Corrected Total R-Square Adj R-Sq F Value Pr > F 1. 28 0. 3437 0. 2996 0. 0661 Parameter Estimates Variable Label Intercept PSCORE ADPT 2020/10/3 DF Parameter Estimate Standard Error t Value Pr > |t| Standardized Estimate 1 1 1 3. 58595 -0. 06575 0. 01837 1. 45462 0. 04132 0. 07366 2. 47 -1. 59 0. 25 0. 0488 0. 1627 0. 8114 0 -0. 56443 0. 08845 (C) Masataka Yamada 65

Segment #7 　　　　　　　　The REG Procedure Model: MODEL 11 Dependent Variable: LDSLT Analysis of Variance DF Sum of Squares Mean Square 2 6 8 5. 24213 2. 90283 8. 14496 2. 62106 0. 48380 Root MSE Dependent Mean Coeff Var 0. 69556 1. 49554 46. 50909 Source Model Error Corrected Total R-Square Adj R-Sq F Value Pr > F 5. 42 0. 0453 0. 6436 0. 5248 Parameter Estimates Variable Label Intercept ISCORE ADPT 2020/10/3 DF Parameter Estimate Standard Error t Value Pr > |t| Standardized Estimate 1 1 1 2. 82330 -0. 18762 0. 22832 0. 66140 0. 05718 0. 08931 4. 27 -3. 28 2. 56 0. 0053 0. 0168 0. 0431 0 -1. 41134 1. 09953 (C) Masataka Yamada 66

Segment #7 The REG Procedure Model: MODEL 9 　　　　　Dependent Variable: LWii. T Analysis of Variance DF Sum of Squares Mean Square 2 6 8 1. 42079 1. 56870 2. 98949 0. 71039 0. 26145 Root MSE Dependent Mean Coeff Var 0. 51132 0. 38508 132. 78248 Source Model Error Corrected Total R-Square Adj R-Sq F Value Pr > F 2. 72 0. 1445 0. 4753 0. 3003 Parameter Estimates Variable Label Intercept PSCORE ADPT 2020/10/3 DF Parameter Estimate Standard Error t Value Pr > |t| Standardized Estimate 1 1. 64734 -0. 04910 0. 04393 0. 76276 0. 02167 0. 03863 2. 16 -2. 27 1. 14 0. 0741 0. 0640 0. 2988 0 -0. 69571 0. 34915 (C) Masataka Yamada 67

Segment #7 The REG Procedure Model: MODEL 12 Dependent Variable: LWii. T Analysis of Variance DF Sum of Squares Mean Square 2 6 8 2. 06011 0. 92937 2. 98949 1. 03006 0. 15490 Root MSE Dependent Mean Coeff Var 0. 39357 0. 38508 102. 20365 Source Model Error Corrected Total R-Square Adj R-Sq F Value Pr > F 6. 65 0. 0300 0. 6891 0. 5855 Parameter Estimates Variable Label Intercept ISCORE ADPT 2020/10/3 DF Parameter Estimate Standard Error t Value Pr > |t| Standardized Estimate 1 1 1 0. 92215 -0. 11572 0. 16932 0. 37424 0. 03235 0. 05054 2. 46 -3. 58 3. 35 0. 0488 0. 0117 0. 0154 0 -1. 43678 1. 34590 (C) Masataka Yamada 68

Segment #6 The REG Procedure Model: MODEL 7 Dependent Variable: LDST Analysis of Variance DF Sum of Squares Mean Square 2 5 7 8. 25078 3. 93279 12. 18357 4. 12539 0. 78656 Root MSE Dependent Mean Coeff Var 0. 88688 1. 45576 60. 92232 Source Model Error Corrected Total R-Square Adj R-Sq F Value Pr > F 5. 24 0. 0592 0. 6772 0. 5481 Parameter Estimates Variable Label Intercept ISCORE ADPT 2020/10/3 DF Parameter Estimate Standard Error t Value Pr > |t| Standardized Estimate 1 1 1 2. 19470 -0. 16390 0. 24786 1. 65109 0. 07336 0. 09939 1. 33 -2. 23 2. 49 0. 2412 0. 0758 0. 0549 0 -0. 56906 0. 63513 (C) Masataka Yamada 69

Sammary of Findings of Mathematical Models to forecast one’s adoption time We could developed some models on DS LITE and Wii only for segments #6 and #7 but not on DS. Segments #6 and #7 are higher level of interest in product category. PSCORE does not contribute to the forecating. 2020/10/3 (C) Masataka Yamada 70

6. Conclusions • Although a strong relationship between adoption times and Personality scores (=PSCORE) considered as innate innovativeness on the cell phone adoption study (Yamada and Zhu 2005) is found, we do not find such a strong relationship on this game machine category (hobby). • Rather we found that either ISCORE has a relatively weak relationship with adoption time only when the level of the interest in product category is higher in such case as #6 and # 7. 2020/10/3 (C) Masataka Yamada 71

6. Conclusions (continued) • PSCORE can be suitable to the case where an innovation is extremely attractive or necessary so that only innovative personality becomes relevant for the adoption over situational and interaction influences like in the cell phone adoption case. • PSCORE does not contribute to the forecasting in case of non necessary good. Rather ISCORE contributes with additional variables, 2020/10/3 (C) Masataka Yamada 72

Innate Innovativeness Necessary goods Category-specific Innovativeness Hobby products Finally, only individuals whose level of interest in product category is high can have strong product category specific innovativeness and innate innovativeness. 2020/10/3 (C) Masataka Yamada 73

References Donnelly, James H. (1970), “Social Character and Acceptance of New Products, ” Journal of Marketing Research, 7 (February), 111 -113. Goldsmith, Ronald E. , Charles F. Hofacker. 1991. Measuring consumer innovativeness. Journal of the Academy of Marketing Science, Summer 19 (3) 209 -221. Mahajan, Vijay, Eitan Muller, Rajendra K. Srivastava. 1990. Determination of adopter categories by using innovation diffusion models. Journal of Marketing Research. 27 (February) 37 -50. Markus, Hazel Rose and Shinobu Kitayama (1991), “Culture and the self: Implication for cognition, emotion, and motivation, ” Psychological Review, 98, 224 -253. Midgley, David F. , Grahame R. Dowling. 1978. Innovativeness: The concept and its measurement. Journal of Consumer Research. 4 (March) 229 -242. Moe, Wendy, Peter Fader. 2002. Using advanced purchase orders to forecast new product sales. Marketing Science. 21 (3) 347 -364. Van den Bulte, Christophe, Yogesh V. Joshi. 2007. New product diffusion with influentials and imitators. Marketing Science. Forthcoming. 　 2020/10/3 (C) Masataka Yamada 74

References continued Yamada, Masataka, Ryuji Furukawa, Hiroshi Kato 2001. New product and eagerly wanted product adoption and diffusion processes: A conceptual model, " Review of Marketing Science Working Papers, 1. The School of Management at The University of Texas at Dallas. Yamada, Masataka and Xiaoying Zhu 2005. A basic model for identifying adopter’s category (adoption time) with personality score, ” Marketing Science Conference 2005. Emory University, Atlanta Georgia, INFORMS, Society for Marketing Science. 木内亜紀（1995）「相互独立・相互依存的自己理解尺度の作成及び信頼性・妥当性 の検討」『心理学研究』，66，100 -106。 2020/10/3 (C) Masataka Yamada 75

Appendix 1 Kiuchi’s 16 -item Test Q 1 A: I generally agree with the opinions of other people. B: I always express my own opinion. Q 2 A: I show my individuality. B: I cooperate with other people. Q 3 A: In order to meet the expectations of other people, I usually conform to their ways of thinking. B: Despite receiving criticism from other people, I rarely change my way of thinking. Q 4 A: I usually express my feelings honestly. B: I usually try to conform to others. Q 5 A: When I have to do something, I usually think first about how other people expect me to act. B: When I have to do something, I usually think first about how I can make the best use of my abilities. 2020/10/3 (C) Masataka Yamada 76

Kiuchi’s 16 -item Test continued Q 6 A: I usually do what I want to do despite opposition from other people. B: I usually give up doing what I want to do, if other people do not want to do it. Q 7 A: I usually accomplish my goals despite opposition from other people. B: I usually give up trying to accomplish my goals, if I meet with opposition from other people. Q 8 A: I express my individuality rather than behaving the way other people want me to behave. B: I behave the way other people want me to behave. Q 9 A: I behave the way other people want me to behave rather than making the most of my abilities. B: I make the most of my abilities. 2020/10/3 (C) Masataka Yamada 77

Kiuchi’s 16 -item Test continued Q 10 A: When I have to do something, I usually think first about how to please other people. B: When I have to do something, I usually think first about how I can make the best of my abilities. Q 11 A: I usually avoid conflicts of interest. B: I usually make my interests and desires clear to other people. Q 12 A: In expressing my opinion, I usually consider how other people think. B: I usually have confidence in my opinion, and therefore, I express it frankly. Q 13 A: In acting, I usually consider the values of other people. B: I usually act according to my own values. 2020/10/3 (C) Masataka Yamada 78

Kiuchi’s 16 -item Test continued Q 14 A: Whenever I do something, I usually make concessions to other people. B: Whenever I do something, I rarely make concessions to other people. Q 15 A: I usually make a decision based on my own judgment, and I take responsibility for the decision. B: I usually make a decision after consulting other people. Q 16 A: At a meeting with other people, I usually speak without reservation. B: At a meeting with other people, I am usually reserved. 2020/10/3 (C) Masataka Yamada 79

Kiuchi’s 16 -item Test continued For all questions, respondents must choose one of the following answers: l I strongly agree with A. l I agree with A if I have to choose from A or B. l I agree with B if I have to choose from A or B. l I strongly agree with B. return 2020/10/3 (C) Masataka Yamada 80

Appendix 2 Goldsmith and Hofacker’s Six-item Test　for Category-specific Innovativeness Negative-direction Q 21: I am the least interested in game machines among my friends. Q 22: Usually I am the last person who becomes aware of recent game machines. Q 23: When a new game machine appeared, often times, I was the last person who bought the new model among my friends. 2020/10/3 (C) Masataka Yamada 81

Appendix 2 continued Positive-direction Q 24: If I were allowed to buy a new game machine, then I would buy it immediately. Q 25: I am a kind of a person who buy a new machine without testing it by myself. Q 26: I know the launching date of a new game machine before other people know it. 2020/10/3 (C) Masataka Yamada 82

Appendix 2 continued For all questions, respondents must choose one of the following answers: l I strongly agree with the statement. l I rather agree with the statement. l I am indifference. l I rather disagree with the statement. l I strongly disagree with the statement. Return 2020/10/3 (C) Masataka Yamada 83