LECTURE 1 - INTRODUCTION IMR HBBA CHAPTER 1+2
Marketing research problems:
I. Targeting and segmentation
II. Positioning
III. Guide marketing mix decisions
→ Use statistics to find solutions to these issues
→ For each type of problem, we will be able to find a type of statistical method that fits to solve
that problem.
Multivariate analysis
Refers to all statistical methods that simultaneously analyze multiple measurements on each
individual or object under investigation - any simultaneous analysis of more than two variables -
examining relationships between/among more than two variables.
The variate
Building block of multivariate analysis - a linear nation of variables with empirically determined
weights determined to meet a specific objective - captures the multivariate character of the
analysis.
Measurement scales:
1. Non-metric scales (qualitative)
> Nominal
Assigns numbers to label or identify subjects or objects - these numbers have no
quantitative meaning beyond indicating the presence or absence of the attribute
- Lowest measurement scale
- Unique definition, identification classification (SNR/ANR/gender)
- Methods of analysis: %, mode (value that most often occurs in dataset), chi
square test
> Ordinal
- Indicate ‘order’, sequence (level of education, preference ranking)
- Methods of analysis: percentiles, median, rank correlation (+ all previous
statistics)
2. Metric scales (quantitative)
Variables can be ordered in relation to the amount of the attribute possessed - every
subject can be compared with another in terms of a ‘’greater than’’ or ‘’less then’’
Relationship - highest level measurement precision
> Interval
- Arbitrary origin (attribute scores/price index)
- Methods of analysis: arithmetic average, range, standard deviation,
product-moment correlation (+ all previous methods)
> Ratio
- Unique origin (age, cost, # customers)
- Methods of analysis: geometric average, coefficient of variation (+ all previous
, methods)
- Highest form of measurement precision
> Real difference: interval scales use an arbitrary zero point (Celcius), ratio include
absolute zero point
→ Should be able to determine what scale is used
Measurement error:
The degree to which the observed values are not representative of the ‘’true’’ values.
→ When variables with measurement error are used to compute correlations or means,
the ‘’true’’ effect is partially masked by the measurement error, causing the
correlations to weaken and the means to be less precise.
Errors:
1. Reliability
Degree to which the observed variables measures the ‘’true’’ value and is ‘’‘error free’’: thus, it is
the opposite of measurement error - If the same measure is asked repeatedly, more reliable
measures will show greater consistency than less reliable measures. Is the measure consistent,
correctly registered?
2. Validity
The degree to which a measure accurately represents what it is supposed to do - does the
mesure capture the concept it is supposed to measure?
Statistical significance and power
Interpreting statistical inferences
requires the researcher to specify the
acceptable levels of statistical error
that result from using a sampling
(sampling error)
Type I error (α):
- Probability of test showing significance when it is NOT present; occurs if you measure a
difference in your analysis, but in reality there is no difference (so measurement error)
- Probability of incorrectly rejecting the null hypothesis - in most cases, it means saying a
difference or correlation exists when it actually does not.
→ false positive
, Type II error (β):
- Probability of incorrectly failing to reject the null hypothesis - not rejecting when it is actually
false
- The chance of not finding a correlation or mean difference when it does exists
→ Extension is power (1-β)
! Value of 1-Type II error (1-B) = power
Power:
The probability of correctly rejecting the null hypothesis when it should be rejected - the
probability that statistical significance will be indicated if it is present
Limit type I error → use stricter decision rule
Cutoff: set it in such way that alpha (Type I error) is smaller than 5% (0.05)
→ Alpha < 0.05 means there is a difference
Power determined by:
1. Alpha (+) positive relationship
As alpha becomes more restrictive, power decreases - as researcher reduces the
chance of incorrectly saying an effect is significant when it is not, the probability of
correctly finding an effect decreases
→ If you are willing to accept a higher type I error, it increases the power
2. Effect size (+)
Probability of achieving statistical significance is based not only on statistical
considerations, but also on the actual size of the effect
→ If the actual effect size is bigger, the power increases
3. Sample size n (+)
Increased sample sizes always produce greater power for the statistical test - as sample
size increases, researchers must decide if the power is too high - increasing sample
size, smaller and smaller effects (correlations) will be found to be statistically significant
until at very large sample sizes almost any effect is significant.
Implications:
- Anticipate consequences of alpha, effect, and n
- Assess/incorporate power when interpreting results
Marketing research problems:
I. Targeting and segmentation
II. Positioning
III. Guide marketing mix decisions
→ Use statistics to find solutions to these issues
→ For each type of problem, we will be able to find a type of statistical method that fits to solve
that problem.
Multivariate analysis
Refers to all statistical methods that simultaneously analyze multiple measurements on each
individual or object under investigation - any simultaneous analysis of more than two variables -
examining relationships between/among more than two variables.
The variate
Building block of multivariate analysis - a linear nation of variables with empirically determined
weights determined to meet a specific objective - captures the multivariate character of the
analysis.
Measurement scales:
1. Non-metric scales (qualitative)
> Nominal
Assigns numbers to label or identify subjects or objects - these numbers have no
quantitative meaning beyond indicating the presence or absence of the attribute
- Lowest measurement scale
- Unique definition, identification classification (SNR/ANR/gender)
- Methods of analysis: %, mode (value that most often occurs in dataset), chi
square test
> Ordinal
- Indicate ‘order’, sequence (level of education, preference ranking)
- Methods of analysis: percentiles, median, rank correlation (+ all previous
statistics)
2. Metric scales (quantitative)
Variables can be ordered in relation to the amount of the attribute possessed - every
subject can be compared with another in terms of a ‘’greater than’’ or ‘’less then’’
Relationship - highest level measurement precision
> Interval
- Arbitrary origin (attribute scores/price index)
- Methods of analysis: arithmetic average, range, standard deviation,
product-moment correlation (+ all previous methods)
> Ratio
- Unique origin (age, cost, # customers)
- Methods of analysis: geometric average, coefficient of variation (+ all previous
, methods)
- Highest form of measurement precision
> Real difference: interval scales use an arbitrary zero point (Celcius), ratio include
absolute zero point
→ Should be able to determine what scale is used
Measurement error:
The degree to which the observed values are not representative of the ‘’true’’ values.
→ When variables with measurement error are used to compute correlations or means,
the ‘’true’’ effect is partially masked by the measurement error, causing the
correlations to weaken and the means to be less precise.
Errors:
1. Reliability
Degree to which the observed variables measures the ‘’true’’ value and is ‘’‘error free’’: thus, it is
the opposite of measurement error - If the same measure is asked repeatedly, more reliable
measures will show greater consistency than less reliable measures. Is the measure consistent,
correctly registered?
2. Validity
The degree to which a measure accurately represents what it is supposed to do - does the
mesure capture the concept it is supposed to measure?
Statistical significance and power
Interpreting statistical inferences
requires the researcher to specify the
acceptable levels of statistical error
that result from using a sampling
(sampling error)
Type I error (α):
- Probability of test showing significance when it is NOT present; occurs if you measure a
difference in your analysis, but in reality there is no difference (so measurement error)
- Probability of incorrectly rejecting the null hypothesis - in most cases, it means saying a
difference or correlation exists when it actually does not.
→ false positive
, Type II error (β):
- Probability of incorrectly failing to reject the null hypothesis - not rejecting when it is actually
false
- The chance of not finding a correlation or mean difference when it does exists
→ Extension is power (1-β)
! Value of 1-Type II error (1-B) = power
Power:
The probability of correctly rejecting the null hypothesis when it should be rejected - the
probability that statistical significance will be indicated if it is present
Limit type I error → use stricter decision rule
Cutoff: set it in such way that alpha (Type I error) is smaller than 5% (0.05)
→ Alpha < 0.05 means there is a difference
Power determined by:
1. Alpha (+) positive relationship
As alpha becomes more restrictive, power decreases - as researcher reduces the
chance of incorrectly saying an effect is significant when it is not, the probability of
correctly finding an effect decreases
→ If you are willing to accept a higher type I error, it increases the power
2. Effect size (+)
Probability of achieving statistical significance is based not only on statistical
considerations, but also on the actual size of the effect
→ If the actual effect size is bigger, the power increases
3. Sample size n (+)
Increased sample sizes always produce greater power for the statistical test - as sample
size increases, researchers must decide if the power is too high - increasing sample
size, smaller and smaller effects (correlations) will be found to be statistically significant
until at very large sample sizes almost any effect is significant.
Implications:
- Anticipate consequences of alpha, effect, and n
- Assess/incorporate power when interpreting results
