Predict Consumer Decisions with Choice-Based Conjoint

A course showing managers and researchers how to run Choice Based Conjoint experiments to predict people's decisions

Created byLuke Greenacre

Last updated 1/2021

English

What you'll learn

Understand the capabilities of Choice-Based Conjoint and Discrete Choice Experiments
Recognise the requirements of running a CBC and DCE research project
Be able to develop and run your own small scale CBC and DCE project
Practice your design and analysis skills with real data sets and extra examples

Course content

9 sections • 34 lectures • 4h 17m total length

The lesson plan for this course2:55
Welcome and Introduction to the course.

00:00 - Title

00:10 - Welcome

01:03 - The lessons

02:11 - What you should come away with
How to progress your learning (for students only)2:49
An overview of how you can best learn Choice Based Conjoint.
00:00 - Title
00:23 - This course's learning approach
01:13 - Understanding learning styles
02:09 - Your learning objective

The aim of all conjoint tasks: Measuring utility5:16
Explains how 'utility' is the basis of choice based conjoint.
00:00 - Title
00:14 - Conjoint and utility
00:49 - Utility as a dependant variable
02:43 - (Academic) utility - Random Utility Theory
03:40 - What utility does....
04:44 - Conjoint is about measuring utility
The ways of measuring utility9:30
Describes the different ways to measure utility, highlighting why measuring using 'Choice' offers the biggest advantages.

00:00 - Title
00:10 - Utility recap
00:33 - The different ways to measure utility
02:16 - Rating scale measurement
03:48 - Ranking based conjoint measurement
06:00 - Choice based conjoint measurement
07:55 - A side note on combined measurement

The adaptive conjoint link from the last slide in the lecture:
http://www.sawtoothsoftware.com/products/acbc/
The look and feel of Choice Based Conjoint3:36
Explains each of the elements of a Choice Based Conjoint experiment.

00:00 - Title
00:21 - Experiencing a survey
00:45 - An 'Alternative'
00:59 - An 'Attribute'
01:24 - An 'Attibute level'
02:40 - 'Choice sets'
03:21 - The whole Choice Based Conjoint
The capabilities of Choice Based Conjoint9:07
Provides examples of what Choice Based Conjoint can be used for.

00:00 - Title
00:31 - What problems does CBC solve?
01:37 - Some examples
02:02 - Cost management in health example
04:53 - Brand equity measurement example
07:07 - Product design example
An example of a Choice Based Conjoint study9:33
An example of how a Choice Based Conjoint study is developed, run, and analysed.

00:00 - Title
00:36 - Digital camera context
00:59 - Target features to investigate
02:03 - Designing the experiment
02:59 - The survey
03:41 - The data collected
04:34 - Analysing the data
05:36 - The regression output
06:54 - The results
07:56 - Using the results

Note: the final slide has has an error. Those market shares were calculated as if Camera D had a 7x zoom, when in fact it should have been calculated as having a 4x zoom.

The types of experimental designs11:38
Explains the three main types of experimental design.

00:00 - Title
00:30 - The challenge of design
01:23 - The purpose of design
02:01 - The types of design
03:19 - Alternative design
05:23 - Choice set design
07:55 - Information/context design
09:43 - The reality of design
10:36 - The benefits of design
Alternative design: factorial designs19:44
Explains how to use factorial designs to generate the alternatives in the Choice Based Conjoint experiment.

00:00 - Title
00:13 - Alternative design
00:43 - Factorials (definition)
02:01 - Full factorial (creation)
05:43 - Full factorial (size and interactions)
07:20 - Fractional factorials (definitions)
10:10 - Fractional factorials (creation)
15:56 - Fractional factorials (correlation test)
19:02 - Factorial designs recap

Note: I explain one mechanism here to create fractional factorial designs. There are other ways of generating fractionals that can produce different results. With further study in this area you can start to learn these different ways and anticipate when one design method may offer advantages over another.

A website that examines different ways to automate the creation of factorial designs:
http://harvest.nps.edu/software.html#DOE

A good overview of fractional factorials can be found here:
http://methodology.psu.edu/ra/most/fefaq#t41n270
Alternative design: main effects designs8:08
Explains how to create and Orthogonal Main Effects Plan (OMEP) to generate the alternatives in your CBC.

00:00 - Title
00:13 - OMEPs
01:33 - Advanced OMEPs
02:34 - Generating OMEPs in SPSS

If you need any help with creating an OMEP in SPSS this help website can talk you through. It also provides sample coding:
http://publib.boulder.ibm.com/infocenter/spssstat/v20r0m0/index.jsp?topic=%2Fcom.ibm.spss.statistics.help%2Fidh_orth.htm
Alternative design: natural designs5:20
Introduces an alterantive to orthogonal designs (Full/Fractional factorials and OMEPs); natural designs.

00:00 - Title
00:10 - Orthogonal designs
00:45 - Natural designs
01:26 - Some effects to design in...
04:17 - Natural designs (conclusion)
Choice set design: random designs12:33
Explains how to construct a choice set using a random design.

00:00 - Title
00:13 - Choice set design
00:38 - Random design generation
08:24 - The quality of random design
Choice set design: combinatorials8:07
Details how to use a combinatorial design to create choice sets.

00:00 - Title
00:04 - Combinatorial Designs
01:33 - Generating them in Excel
03:59 - Generating them online
06:53 - Design Quality
Choice set design: balanced incomplete block designs4:59
Explains how to construct choice sets using BIBDs, and highlights other potential designs available.

00:00 - Title
00:07 - BIBDs
01:04 - BIBD generation
02:27 - Design quality
03:31 - Other block designs

I have attached an Excel file containing some BIBDs to help you start building your library of designs.

Layout of the choice sets4:44
Details the main issues to consider when choosing how to present your choice sets to participants (the survey layout).

00:00 - Title
00:13 - Survey layout
00:51 - Classic matrix
01:29 - Shop layout
02:29 - Horizontal vs vertical
03:25 - Text vs visual
03:44 - Interaction
04:06 - Choice set layout
Giving effective instructions5:11
Gives some helpful hints and tips about how to give instructions in your survey that will improve your data quality.

00:00 - Title
00:08 - Instructions
00:27 - Survey instructions
03:21 - Choice set instructions
Software for running the survey10:42
Introduces some of the software packages that can be used to design your CBC experiment.

00:00 - Title
00:36 - Packages: Advanced and DIY
01:23 - Advanced: Confirmit
03:01 - DIY: SurveyGizmo and similar
04:44 - A look at SurveyGizmo

SurveyGizmo: http://www.surveygizmo.com/
Confirmit: http://www.confirmit.com/

Determining your analysis options10:59
Introduces some of the major issues to consider when choosing an analysis approach for your CBC project.

00:00 - Title
00:34 - Analysis = Discrete DV models
01:45 - Why are there different models?
05:24 - Discrete DV models
06:54 - The link to design
07:39 - Software

Stata: www.stata.com
Regression Models for Categorical Dependent Variables Using Stata, 2nd Edition, by J. Scott Long and Jeremy Freese
Simple analysis methods10:39
Provides an overview of the main analysis methods available to CBC researchers.

00:00 - Title
00:18 - What is simple?
01:07 - Linear regression
03:27 - Linear probability models
05:41 - Logit and probit
07:40 - (Conditional) Multinomial logit model
09:54 - Many more
More complex analysis methods7:48
Gives a top line overview of some of the more elaborate analysis methods available to CBC researchers.

00:00 - Title
00:30 - What is complex?
01:19 - The big three
02:11 - Mother logit
03:49 - Latent class
05:32 - Bayesian estimation
06:40 - A warning

For a technical introduction to some of the more complex analysis methods try this website:
http://elsa.berkeley.edu/~train/distant.html

A brief overview3:19
Gives the background to the research problem being addressed in this example of regression analysis.

00:00 - Title
00:13 - Research background
00:43 - Keeping it simple
01:20 - Alternatives
01:41 - Experimental design
02:48 - The data
Data layout needed5:39
Describes the way to structure the data in Excel so that the analysis can be completed.

00:00 - Title
00:05 - The survey data
00:33 - What it needs to look like
01:56 - Excel
Running the analysis3:50
Gives step by step instructions on how to run the regression in Excel.

00:00 - Title
00:13 - MS Excel
Interpreting the output6:37
Gives an overview of how to interpret the regression model to be able to predict choice frequency.

00:00 - Title
00:58 - Summary
02:55 - ANOVA
04:30 - Coefficients
05:55 - Regression equation
Calculating choice probabilities and using them10:03
Explains how to calculate the choice probabilities for the alternatives that you are interested in. These choice probabilities are analagous to market shares and let you predict the market.

00:00 - Title
00:11 - Using the regression equation
00:30 - Calculating the choice probabilities in Excel

A brief overview4:21
Introduces the conditional Multinomial Logit (cMNL)model for analysing CBC data. Gives you some background on the data we will be using for this exercise.

00:00 - Title

01:03 - Research background

01:40 - Keeping it simple

02:12 - Alternatives

02:37 - Experimental design

03:51 - The data
Data layout needed11:57
Describes how to prepare the data set so you can run the cMNL analysis. This is a very important stage as most of the difficulty with this analysis is in the data setup.

00:00 - Title

00:18 - The survey data

01:17 - What the data set needs

02:02 - The data in Excel
Running the analysis3:53
Gives step by step instructions on how to run the analysis in SPSS. A written version of these instructions can be found in word file attached to the "Brief Overview" lecture at the start of this section.

00:00 - Title

00:08 - Analysing in SPSS
Interpreting the output8:54
Goes through each of the major elements of the output of the analysis and explains how to interpet them meaningfully.

00:00 - Title

00:25 - Case processing summary

01:13 - Stratum status

02:01 - Omnibus tests

02:58 - McFadden's R-square

05:10 - Variables in the equation

07:30 - Regression equation
Calculating choice probabilities and using them10:00
Explains how to use the results to calculate choice probabilities, which are generally much more meaningful for decision makers.

00:00 - Title

00:12 - Using the regression equation

00:41 - Calculating in Excel

An overview of Decisions Support Systems3:38
Decision Support System help decision makers make sense of results and can be a great way to add value to a CBC project. This lecture explains the basic elements of a DSS.

00:00 - Title

00:11 - Definition

01:07 - Aims

02:44 - Reasons for having one
The basics of building a DSS6:38
Talks through a simple example of a DSS. The sky is the limit with how elaborate you wish to make a DSS, but even the most simple one can make a lot of difference when helping a decision maker understand CBC results.

00:00 - Title

00:10 - The example

Further learning for CBC7:15
Lists some of the resources available to help you move into more advanced forms of CBC.

00:00 - Title

00:12 - The challenges

01:31 - What you know

02:03 - Experimental design

02:49 - Analysis methods

04:08 - Presentation of results

05:21 - State of the art

06:02 - Hunt around and build your resources
Some suppliers that are handy to know about8:20
Provides the details of some of the key supplies in the industry that can help you with your CBC projects.

00:00 - Title

00:21 - Survey software

01:04 - Panel providers

05:30 - Data analysis

06:10 - Complete solutions

As a heads up the Centre for the Study of Choice (CenSoC) has now moved and is called the Institute for Choice. You can find their details here:

https://www.unisa.edu.au/Research/Institute-for-Choice/

Requirements

There is no assumed knowledge for the early topics in this course.
For those progressing to the advanced topics in this course an understanding of introductory level statistics is recommended.
Access to MS Excel for running analysis is essential for those progressing to advanced topics. Additional topics covering analysis using SPSS are also provided but they are not essential.

Description

Choice-Based Conjoint, also often called Discrete Choice Experimentation, is a powerful research and management tool that allow us to understand and predict people's preferences. Whether it is a manager wanting to predict product preferences, a health researcher wanting to explore the treatment preferences of patients, or a transport engineer examining people's choices of public transport this tool can provide the insight needed.

This course starts with an introduction to the capabilities and applications of Choice Based Conjoint that is suitable for all audiences. It explains the basic requirements of a CBC research project, and details the outputs that can be obtained. The course then continues on to more advanced topics where you will get training and hands on experience developing and running a CBC project. This includes design, data collection, analysis and reporting of results.

This course is suitable for managers, marketing/business researchers, and academic researchers interested in building an understanding of CBC.

If you are a PhD student I am happy to provide a discounted rate for this course. Please contact me through the Udemy messaging service. Provide your university email address, your name, and a link to your supervisor's profile on your university website.

Who this course is for:

Managers wanting to understand the capabilities of CBC and DCEs for measuring people's preferences
Researchers wanting to learn how to implement CBC and DCE projects

Predict Consumer Decisions with Choice-Based Conjoint

What you'll learn

Explore related topics

Course content

Introduction and Overview2 lectures • 6min

An Overview of Choice Based Conjoint (An ideal summary for Managers)5 lectures • 37min

Designing a Choice Based Conjoint Experiment7 lectures • 1hr 10min

Laying Out The Survey3 lectures • 21min

A Brief Introduction to Analysing Your Data3 lectures • 29min

Analysing Your Data: Regression using MS Excel5 lectures • 29min

Analysing Your Data: The (c)MNL in SPSS5 lectures • 39min

Building Decision Support Systems with the Results2 lectures • 10min

Thank You and Good Night2 lectures • 16min

Requirements

Description

Who this course is for: