Conjoint Analysis Simulation - weird output

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

fabio1989
Posts: 4
Joined: Sun Jan 04, 2015 9:45 pm

Conjoint Analysis Simulation - weird output

Postby fabio1989 » Mon Jan 05, 2015 8:13 pm

Hi everyone,

I am doing a traditional conjoint study with 3 attributes (brand,price and applicator) each with three levels.
I asked respondents to rate the profiles from 1 to 9.

Now I run 2 simulations with 9 profiles each:

1st simulation:

1) Nivea Spray 2,99€
2) Rexona Spray 2,99€
3) Dove Spray 2,99€
4) Nivea Rollon 2,99€
5) Rexona Rollon 2,99€
6) Dove Rollon 2,99€
7) Nivea Stick 2,99€
8) Rexona Stick 2,99€
9) Dove Stick 2,99€


2nd simulation:

1) Nivea Spray 2,99€
2) Rexona Spray 3,99€
3) Dove Spray 2,99€
4) Nivea Rollon 2,99€
5) Rexona Rollon 3,99€
6) Dove Rollon 2,99€
7) Nivea Stick 2,99€
8) Rexona Stick 3,99€
9) Dove Stick 2,99€

As you can see in the second simulation Rexona has an higher price compared to the 1st simulation.

Since from the output of conjoint I can easily see that the higher the price lowest the utility I do not understand why in the simulation outputs Rexona share of preference increases if I increase its price.
This happen with logit and first choice for 2 of the 3 Rexona products.

Can you please help me to understand why or tell me what I am doing wrong?

PS: I use the same subjects, the same data, I already checked the two plan files for the simulation and they are correct.

Let me know if you need further info.

Thank you very much,

Fabio
statman
Administrator
Posts: 2721
Joined: Tue Jun 12, 2007 12:08 pm
Location: Florida, USA

Re: Conjoint Analysis Simulation - weird output

Postby statman » Mon Jan 05, 2015 9:25 pm

Could be several reasons since these are simulations but need more like the design syntax and how the attribute/levels were measures. Can you post the syntax and ...
See the note below

NOTE: Please read the Posting Guidelines and always tell us your OS, the SPSS version and information about your study and data!

Statman
Statistical Services
fabio1989
Posts: 4
Joined: Sun Jan 04, 2015 9:45 pm

Re: Conjoint Analysis Simulation - weird output

Postby fabio1989 » Mon Jan 05, 2015 10:29 pm

Below some more details.
Before I would like to say that some of the profiles present in my simulation were present also in my survey hence they have been rated by subject. I don't know if this can be an issue somehow. Moreover, in the survey I asked to rate the profiles from 1 to 9. then I needed to add +10 to each score because when computing simulation some profiles were getting negative predicted score. Therefore some respondents were taken out from the analysis.

Syntax Simulation:

CONJOINT PLAN='C:\Users\gino\Desktop\SERIOTESI\Femalemarketrexonapremium.sav'
/DATA=*
/SCORE=Q100_1 to Q100_9 /SUBJECT=progressivo /FACTORS=Brand 'Marca' ( 1 'Nivea' 2 'Rexona' 3 'Dove' ) Applicator 'Applicatore' ( 1 'Spray' 2 'Rollon' 3 'Stick' ) Price 'Prezzo' ( 1 '1,99' 2 '2,99' 3 '3,99' )
/PRINT=all

The summary utility table for price says:
Price 1,99€: +1,91
Price 2,99€: - 0,07
Price 3,99€: - 0,184


The summary preference score of simulation NUMBER ONE (rexona 2,99€) says:

1) Nivea Spray 2,99€ 14,290
2) Rexona Spray 2,99€ 13,088
3) Dove Spray 2,99€ 14,496
4) Nivea Rollon 2,99€ 14,401
5) Rexona Rollon 2,99€ 13,200
6) Dove Rollon 2,99€ 14,607
7) Nivea Stick 2,99€ 13,926
8) Rexona Stick 2,99€ 12,725
9) Dove Stick 2,99€ 14,132

The preference probability for simulation NUMBER 1 (Rexona 2,99€) says:

MAXIMUM UTILITY LOGIT
1) Nivea Spray 2,99€ 18,5% 15,8%
2) Rexona Spray 2,99€ 3,1% 6,4%
3) Dove Spray 2,99€ 18,4% 17,0%
4) Nivea Rollon 2,99€ 17,0% 14,4%
5) Rexona Rollon 2,99€ 3,7% 5,9%
6) Dove Rollon 2,99€ 19,3% 16,0%
7) Nivea Stick 2,99€ 5,4% 8,3%
8) Rexona Stick 2,99€ 2,2% 4,5%
9) Dove Stick 2,99€ 12,4% 11,6%


While the summary preference score of simulation NUMBER 2 (Rexona 3,99€) says:

1) Nivea Spray 2,99€ 14,290
2) Rexona Spray 3,99€ 12,912
3) Dove Spray 2,99€ 14,496
4) Nivea Rollon 2,99€ 14,401
5) Rexona Rollon 3,99€ 13,023
6) Dove Rollon 2,99€ 14,607
7) Nivea Stick 2,99€ 13,926
8) Rexona Stick 3,99€ 12,548
9) Dove Stick 2,99€ 14,132

The preference probability for simulation NUMBER 2 says:

MAXIMUM UTILITY LOGIT
1) Nivea Spray 2,99€ 18,3% 15,3%
2) Rexona Spray 3,99€ 5,5% 8,0%
3) Dove Spray 2,99€ 16,2% 16,0%
4) Nivea Rollon 2,99€ 17,1% 14,9%
5) Rexona Rollon 3,99€ 3,8% 5,3%
6) Dove Rollon 2,99€ 19,1% 16,1%
7) Nivea Stick 2,99€ 5,0% 8,3%
8) Rexona Stick 3,99€ 3,7% 4,6%
9) Dove Stick 2,99€ 11,3% 11,6%

My question is: if price summary says "the lower the better" why profiles 2) and 8) increases their shares when their price is higher? whatever the reason...why is this not happening with profile 5)?

Notice that Preference score for 2) and 8) are higher in the first scenario (Rexona with a lower price). This makes sense to me but I don't understand why shares go in the other direction?

Thank you very much,

Please let me know if you need any further detail.

Fabio

Who is online

Users browsing this forum: No registered users and 2 guests

cron