Discussion: View Thread

SAMPLE SIZE question

  • 1.  SAMPLE SIZE question

    Posted 06-15-2014 16:09


    - Sample size question -

    On a showroom, set 3 cars with the same characteristics, 
    but from different companies. 

    Each car has on their windshield a paper with the letter A, B or C.
    Previously, copies were made of a CD with 1.5 minutes of music:
    30 seconds from each of 3 music selections.
    Insert copies of same CD in each of the 3 cars.
    Invite all visitors to the showroom to sit for 1.5 minutes in each of the 3 cars.

    Give participants paper and pencil.
    The paper has 3 lines, labeled car A, car B, and car C.
    The lines are continuous and labeled from 1 to 10.
    Participants are asked to put an X any where on the line, 
    as they end listening to the 1.5 minutes of music, from each car.
    The objective is for the participants to rate each of the 3 cars
    in terms of how much they liked listening to the music in each car.

    NOTE: This is all my customer offered. There is no other information!

    The question is HOW MANY PARTICIPANTS WOULD BE NEEDED ?
    Is it required to: Have the 3 cars in the showroom for only 1 week [?], longer? shorter?
    Taking the same 3 cars to various showrooms, 1 week at a time [?]
    HOW MANY SHOWROOMS WOULD BE NEEDED ?

    PLEASE COMMENT.
    Thanks for your time.



    -------------------------------------------
    Anamaria Kazanis, PStat
    -------------------------------------------


  • 2.  RE: SAMPLE SIZE question

    Posted 06-15-2014 16:35
    Whew, the client is asking for nothing of a statistical nature...

       The objective is for the participants to rate each of the 3 cars
        in terms of how much they liked listening to the music in each car.

    He should get as many (as few?) participants that he can afford.

    Ok, now that that's out of my system, here's some questions that pop into mind, when thinking of a design...

    How many showrooms does the client have access to?
    How many customers on average visit a particular showroom in a given week?
    How much time will this survey take (is there a sales pitch going on as well?), and will this be administered by sales staff or the consultant?

    Guiding your client, what is that is really wanted?  The "right choice" of music to put into the car to make the sale, with some guarantee of success?

    -Mark
    -------------------------------------------
    Mark Lancaster
    George Mason University
    -------------------------------------------








  • 3.  RE: SAMPLE SIZE question

    Posted 06-15-2014 16:46
    Suggest your client goes to a mall or auto audio store to gauge likely buyers preferences. Doing it as is, selectivity bias (why are they in THAT showroom) and consumer confusion. . . ------------------------------------------- Daniel Slottje FTI Consulting -------------------------------------------


  • 4.  RE: SAMPLE SIZE question

    Posted 06-15-2014 16:58
    THANKS for your comments.
    ALL those are the type of question I am asking of my client.
    Lets see if they pay attention to the statistician...

    -------------------------------------------
    Anamaria Kazanis, PStat
    -------------------------------------------








  • 5.  RE: SAMPLE SIZE question

    Posted 06-15-2014 17:17
    You have to know the goal.  This is a terrible test!  Are they trying to see who likes which car?  which stereo system?  Which music?   which showroom?  which salesperson?  That is Likert scale data so isn't worth a lot and you are limited with what you can do with it so to make any meaningful comparisons you need a lot of it.  

    -------------------------------------------
    Bill Bentley
    -------------------------------------------








  • 6.  RE: SAMPLE SIZE question

    Posted 06-15-2014 18:48
    And don't forget about the effect size! That is where most of my clients have problems. They want me to power a study to see a significant effect between groups etc. but they have no idea about how large of a difference, correlation, etc. is meaningful to them.  There is quite a bit of collaboration needed in this one. Some researchers don't want their questions answered with more questions...but we statisticians are great at doing just that. And it is often needed to get the best design and results possible for the client.

    Typical Scneario:

    Researcher: I need to know if there is an association between three types of cars and three types of  music. How big of a sample do I need?
    Statistician: That depends....

    And a study is never perfect, so adjustments and allowances need to be made here and there to work with the client's time and financial constraints.

    So get the researcher to do some work on his/her end and have fun collaborating to put it together. And tell yourself that you gotta love this stuff, because it happens a lot!

    Best,

    Elaine

    -------------------------------------------
    Elaine Eisenbeisz
    Owner and Principal Statistician
    Omega Statistics
    -------------------------------------------








  • 7.  RE: SAMPLE SIZE question

    Posted 06-15-2014 22:02
    THANKS for all the very interesting comments.
    Regretfully, this is not the first time I encounter this type of requests from clients.
    I find myself in a role as an educator to clients, more often than as actual analyst.
    But, ... this is why they pay the 'big bucks'!

    Thanks again.

    -------------------------------------------
    Anamaria Kazanis, PStat








  • 8.  RE: SAMPLE SIZE question

    Posted 06-15-2014 23:13
    The issue of effect size and sample size gets a lot of attention in Gerald van Belle's book, "Statistical Rules of Thumb", including different ways to phrase the "effect size" question.

    -------------------------------------------
    Michael Kruger
    Information Resources Inc
    -------------------------------------------








  • 9.  RE: SAMPLE SIZE question

    Posted 06-16-2014 07:36
    You definitely need to help your client move from the presenting question to a researchable question.  This often the most critical part of any methodological consulting.

    IFF your client is really interested if finding out whether directly competing models with the same equipment are differentially associated with particular songs, artists, decades of origin, rhythm, genres, string, other instrumental, choral, tempos, a mix, etc.  any effects are going to be very subtle.

    Also a single study might not be sufficient.  A single method might not be sufficient. Methods could vary from psychophysical experiments; other experiments; quasi-experiments; focus groups or other group techniques of idea building;etc.

    -------------------------------------------
    Arthur Kendall
    Social Research Consultants
    -------------------------------------------








  • 10.  RE: SAMPLE SIZE question

    Posted 06-15-2014 20:05
    Apparently the rater is not even told if 1 on the 1-10 scale is "like it better" or if 10 is "like it better".

    -------------------------------------------
    Robert Riffenburgh
    Naval Medical Center
    -------------------------------------------








  • 11.  RE: SAMPLE SIZE question

    Posted 06-15-2014 21:58
    Actually, they were told to place the X taking under consideration that 10 meant they liked listening to the music in that car.

    -------------------------------------------
    Anamaria Kazanis, PStat
    -------------------------------------------








  • 12.  RE: SAMPLE SIZE question

    Posted 06-16-2014 00:20
    Well, the first thing that your client needs is to state a research hypothesis. You and I could probably guess what that hypothesis would be, based on this description, but do try to get the client to articulate their research hypothesis clearly..

    Then you need the variability of your outcome measure. They might not be able to do this, but the fact that your outcome variable is bounded will also place an upper bound on the standard deviation. It's pretty much impossible to have a standard deviation of 500 for a variable that is bounded between 1 and 10.

    It's a crossover design, so you need an estimate of variation within AND between customers. Finally, it is a multi-center study, so you need some sort of measure of variation between and within centers (showrooms). You may have to put in some SWAGs for some of these values. SWAG stands for Scientific Wild something Guess.

    Then, as several people have mentioned, they need to decide how much of a difference is clinically significant.

    There's a time element here, and they need to give an estimate of the number of customers per week who would be likely to volunteer. If there is a cost from moving from showroom to showroom, perhaps this needs to be factored in.



    -------------------------------------------
    Stephen Simon
    Independent Statistical Consultant
    P. Mean Consulting
    -------------------------------------------








  • 13.  RE: SAMPLE SIZE question

    Posted 06-16-2014 11:22
    As everyone has mentioned, the objective of the study needs to be clarified and a proper hypothesis formulated, including the Type I and Type II error requirements.

    Assuming, which may not be the case, that the consumers are evaluating the sound systems of each car, show rooms would matter if there are different demographics for the kinds of customers the client is interested in reaching.

    -------------------------------------------
    Patrick Spagon
    -------------------------------------------








  • 14.  RE: SAMPLE SIZE question

    Posted 06-17-2014 16:12


    -------------------------------------------
    W. Vogt
    Professor
    -------------------------------------------

    The sample size matters less here than the fact that the car companies and music selections are confounded.






  • 15.  RE: SAMPLE SIZE question

    Posted 06-18-2014 09:08
    I didn't perceive the confound. My understanding is that the only thing varying is the vehicle...same CD in each vehicle. CD contains all 3 music selections.

    My thoughts:
    1) to estimate sample size, you need to specify the effect size you'd like to be able to detect for test statistic you anticipate using
    2) Analysis Plan
    a) if participants place Xs largely in the middle, you may be able to get away with ANOVA, as your data may approximate Gaussian. This isn't too improbably given the psychophysics, which if memory serves will push people away from extremes.
    b) alternatively, you could use a binomial model, recoding the location of X as a proportion of the distance of the line...or use the success/trials notation using distance from left and total distance.
    c)  in either case would probably want to employ sandwich estimation
    3) For sample size, it won't be perfect, but you could estimate using power for ANOVA (effect size f)...but may want to conceptualize the differences initially using Cohen's d (mean difference/standard deviation of a group)
    4) If someone else has a suggestion for power anticipating a generalized linear model for binomial data, I'd go with that...but I suspect you will be "close enough" using ANOVA
    5) If not already available, you will probably need some pilot data at least to estimate the standard deviation...nice if get some sense of effect size, though I'd recommend targeting a minimally relevant difference rather than assuming the pilot effect size is a great estimate of the "real" one.
    6) Design
    a) order of music may matter...consider counterbalancing car x music order (3 x 3 =9 combinations).
    b) the potential for order effects are compounded since ratings take place after all 3 music selections have been heard, rather than following each...consider ratings after each if feasible. Order still may matter, but primacy/recency effects will be lessened
    c) The question of showrooms is more complicated...more epidemiology... An education or psychology researcher may use a hierarchical model (random effects) with showroom as the unit of analysis, rather than potential customer. This will appropriately nest observations and provide you with a closer answer to the fundamental question of which are better than which, somewhat robust to showroom.
    d) On the other hand, if there are features of showroom that MODERATE the effect of the differences between combinations, you're missing them. This would seem relevant, particularly if the showrooms vary by region and/or demographic, and particularly as one considers music.
    - Different regions/demographics may favor particular genres
    - Different genres may interact with the psychoacoustics (a system may sound great for folk vocals and terrible for hard rock)
    7) The acoustical environment of a car is very challenging. This smells like an easy project to miss something in the details if there isn't enough convergence of knowledge. I suspect your customers understand this complexity, which is why they're employing you. That raises the bar. You should educate yourself on the various issues of psychoacoustics and the environment of a vehicle...there is a rich literature on psychoacoustics and the best in the car audio business are very well informed.
    8) What a fun project!!!! I'm jealous. Enjoy.





    -------------------------------------------
    Jason T. Machan
    Director, Lifespan Biostatistics Core,
    Lifespan Hospital System
    Research Scientist, Biostatistics, Research
    Rhode Island Hospital
    Assistant Professor, Departments of Orthopaedics and Surgery
    The Warren Alpert Medical School, Brown University
    Director Biostatistics Externship, Adjunct Assistant Professor, Department of Psychology
    University of Rhode Island
    -------------------------------------------