Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

Y3S2 MKTG103 > Data Analysis III: Logistic Regression, Conjoint Analysis (Week 10) > Flashcards

Data Analysis III: Logistic Regression, Conjoint Analysis (Week 10) Flashcards

1
Q

What is logistic regression?

A

Logistic regression: Binary variables for DVs

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What is the equation for logistic regression?

A

Zi = βo + β1 Xi + … + εi

*Coefficients βs affect Z, not response directly

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is the formula relating Z to response?

A

Responsei = e^Zi / (1 + e^Zi)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What are the statistics involved in logistic regression?

A

No F-stat and R^2

1) Hit rate
2) Coefficients - Significance and sign only. No size.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What is hit rate?

A

Hit rate = Overall accuracy of the model

Higher hit rate -> More accurate model

Cutoff value = 0.5 (i.e. Hit rate at least 50%)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What adjustments should we make when responses are unbalanced?

A

We should adjust cutoff value to empirical mean of DV

To increase accuracy of model (not increase hit rate)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What is conjoint analysis?

A
  • Asking consumers to rate bundle of attributes (with certain levels)
  • INFER how impt attributes are based on their ratings
  • For product design
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

How does conjoint analysis reveal the importance of attributes of a product?

A

Customer must make tradeoffs. Conjoint analysis reveals the “value system” behind customer decisions

The more the product rating changes when we vary attribute levels, the more important the attribute

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What are the types of conjoint analysis?

A

1) Rating-based

2) Choice-based

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What is rating-based conjoint analysis?

A

Ask respondents to RATE products, one prdt at a time

Ratings = On a scale

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What is choice-based conjoint analysis?

A

Ask respondents to CHOOSE a prdt, one set at a time

Choice = 0/1

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Which attributes do we include in conjoint analysis?

A
  • Find out which attribs. are IMPT to consumers through focus grps, interviews, business reports
  • Attribs we can control
  • Some attribs are not actionable but impt, e.g. brand
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Which attribute levels do we include in conjoint analysis?

A
  • Which attrib. levels do we consider for our new/revised product?
  • Which levels are AVAILABLE in the mkt?
  • Use concrete levels e.g. 3x3cm vs. small
  • Possible to exclude combinations but advised to minimise no. of exclusions
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What is full factorial design?

A

Present all possible combinations

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

How do you calculate number of combinations for factorial design?

A

No. of levels for Attrib. 1 x No. of levels for Attrib. 2 x …

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What is fractional factorial design?

A

Present a FRACTION of all combinations

MINIMUM no. of combinations that allows for testing of all main effects

17
Q

What are the advantages and disadvantages of fractional factorial design?

A

Adv: Able to capture all main effects
Disadv: Won’t be able to capture interaction effect

18
Q

What are part-worths?

A

Regression coefficients of conjoint analysis

19
Q

How do we interpret part-worths?

A

How much UTILITY does black give (as opposed to base red)?

BASE product has utility of ZERO

20
Q

How do we interpret attribute importance from part-worths?

A
  1. Calculate Range(= Max utility - Min utility) for each attribute
  2. Sum the total range for all attributes
  3. Calculate Importance = Range/Sum x 100%
21
Q

How do we interpret price from part-worths?

A

Price as an attribute in conjoint analysis:

Calculate whether Option 1 has a greater utility than Option 2

22
Q

What are the different choice rules? Which rule applies?

A
  1. First choice
  2. Share of preferences

Depends on context

23
Q

What is the first choice choice rule?

A

Consumer picks product that gives HIGHEST UTILITY

  • Rational decision making
  • High price
  • High involvement
24
Q

What is the share of preferences choice rule?

A

Utility of product over all products

E.g. Share product 1 = Utility prdt 1 / Utility over all prdts

25
Q

What is the process of conjoint analysis?

A
  1. Determine attributes & levels
  2. Create design
  3. Collect ratings from many respondents
  4. Estimate part-worths (regression)
    - Per respondent
    - For groups of respondents
  5. Determine:
    - Attribute importance
    - Ideal prdt (possibly for diff. segments)
    - Market shares

Y3S2 MKTG103 (9 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions