7 Frequency Distributions Cross Tabulations

Embed Size (px)

Citation preview

  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    1/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Applied Marketing

    (Market Research Methods)

    Topic 7:

    Frequency distributions andcross-tabulations

    Dr James Abdey

    http://find/http://goback/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    2/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Overview

    Once data have been collected and prepared foranalysis, basic analysis should be conducted

    Typical forms include frequency distributions and

    cross-tabulations

    Often market research projects do not go beyond

    basic data analysis

    However, such univariate and bivariate analyses

    guide multivariate analysis

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    3/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Frequency distribution

    In a frequency distribution, one variable is

    considered at a time, i.e. a univariate analysis

    A frequency distribution for a variable produces a

    table of frequency counts, percentages and

    cumulative percentages for all the values

    associated with that variable

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    4/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Frequency distribution of familiarity

    with the Internet

    Frequency Percentage Cumulative

    (Count) percentage

    Very unfamiliar (1) 2 1 1Unfamiliar (2) 14 7 8

    Neutral (3) 22 11 19

    Familiar (4) 56 28 47

    Very familiar (5) 106 53 100

    Total 200 100

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    5/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    OverviewFrequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Desciptive statistics associated with

    frequency distributions

    Usual descriptive statistics can be calculated

    These include measures of location (mean,

    median, mode) and measures of dispersion

    (variance, standard deviation, etc.)

    The coefficient of variation is the ratio of the

    standard deviation to the mean expressed as a

    percentage, and is a unitless measure of relativevariability

    CV =S

    X

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    6/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    OverviewFrequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Types of Variable

    Later in the course we will deal with measurable

    variables

    For now we restrict attention to categorical

    variables

    Examples include:

    (i) the number of vegetarians who support Party X(ii) the number of diners at a restaurant who were

    dissatisfied / indifferent / satisfied with the service

    If we can arrange things in a sensible order then we

    have ordinal (categorical) variables

    If not, then we have nominal (categorical) variables

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    7/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    OverviewFrequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Correlation and association

    To determine whether two measurable variables are

    related, we discuss correlation

    Correlation measures the strength of the (linear)

    relationship between two variables

    However, for categorical variables, we discuss

    association

    This will require frequency data

    Here we examine the existence of association rather

    than its strength

    http://find/http://goback/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    8/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    OverviewFrequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Cross-tabulation

    While a frequency distribution describes one variable

    at a time, a cross-tabulation describes two or more

    variables simultaneously

    Cross-tabulation results in tables that reflect the joint

    distribution of two or more variables with a limited

    number of categories or distinct values

    http://find/http://goback/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    9/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    OverviewFrequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Gender and Internet usage

    Internet usage Gender Row total

    Male Female

    Light (1) 50 100 150Heavy (2) 100 50 150

    Column total 150 150

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    10/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    OverviewFrequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Two variables cross-tabulation

    Since two variables have been cross-classified,

    percentages could be computed either

    column-wise, based on column totals, or row-wise,

    based on row totals

    The general rule is to compute the percentages in

    the direction of the independent variable, across the

    dependent variable

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    11/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    OverviewFrequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Internet usage by gender

    Internet usage Gender

    Male Female

    Light (1) 33.3% 66.7%Heavy (2) 66.7% 33.3%

    Column total 100% 100%

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    12/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    OverviewFrequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Gender by Internet usage

    Internet usage Gender Row total

    Male Female

    Light (1) 33.3% 66.7% 100%

    Heavy (2) 66.7% 33.3% 100%

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    13/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    OverviewFrequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    The introduction of a third variable in

    cross-tabulation Using just two (categorical) variables we would either

    conclude that there is no association, or that there issome association between them

    However, by introducing a third variable we may

    reach any of the following conclusions:

    Refine the association between the originalvariables

    Determine that the original association is in factspurious

    Detect suppressed association by introducing athird variable

    Make no change in our original conclusions

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    14/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    OverviewFrequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Purchase of designer clothing by

    relationship status

    Amount of designer Marital status

    clothing purchased Married SingleA lot 30% 55%

    Not a lot 70% 45%

    F

    http://find/http://goback/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    15/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    OverviewFrequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Refining the association with a third

    variable

    55% of single respondents fell in the A lot category,

    as opposed to 30% of the married respondents

    Before concluding that single respondents purchase

    more designer clothing than those who are married,

    a third variable, the buyers gender, was introduced

    into the analysis

    FP h f d i l hi b

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    16/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Purchase of designer clothing by

    marital status and gender

    Amount of designer Gender

    clothing purchased Male FemaleMarried Single Married Single

    A lot 35% 40% 20% 65%

    Not a lot 65% 60% 80% 35%

    FrequencyR fi i h i i i h hi d

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    17/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Refining the association with a third

    variable

    For females, 65% of single women fall in the A lot

    category, as compared to 20% of those who are

    married

    However, the percentages are much closer for

    males, with 40% of single men and 35% of marriedmen falling in the A lot category

    Hence, the introduction of gender (third variable) has

    refined the relationship between marital status and

    purchase of designer clothing (original variables)

    Single respondents are more likely to fall in the A lot

    category than married respondents, and this effect is

    much more pronounced for females than for males

    FrequencyO hi f t b

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    18/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Ownership of sports cars by

    education level

    Own sports car Education

    Graduate Non-graduateYes 30% 20%

    No 70% 80%

    FrequencyO hi f t b

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    19/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Ownership of sports cars by

    education level

    This shows that 30% of graduates own a sports car,

    compared with 20% of non-graduates

    Realising that income may also be a factor, the

    researcher decided to re-examine the relationship

    between education and ownership of sports cars

    taking income into account

    FrequencyO nership of sports cars b

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    20/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Ownership of sports cars by

    education and income levels

    Own Income level

    sports Low Highcar Graduate Non-graduate Graduate Non-graduate

    Yes 20% 20% 40% 40%

    No 80% 80% 60% 60%

    FrequencyUncovering spurious association

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    21/26

    Frequencydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Uncovering spurious association

    The percentages of graduates and non-graduates

    who own sports cars are the same for each of the

    income groups

    When the data for the high-income and low-income

    groups are examined separately, the association

    between education and ownership of sports cars

    disappears, indicating that the initial relationship

    observed between these two variables was spurious

    FrequencyUsing three variables to reveal

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    22/26

    q ydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Using three variables to reveal

    suppressed association

    Want to travel abroad Age

    Under 50 50 and above

    Yes 50% 50%

    No 50% 50%

    We see no association between wanting to travel

    abroad and age

    However, by introducing gender as a third variable,we obtain the following table

    FrequencyWanting to travel abroad by age and

    http://goforward/http://find/http://goback/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    23/26

    q ydistributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Wanting to travel abroad by age and

    gender

    Want to Gender

    travel Male Femaleabroad Under 50 50 and above Under 50 50 and above

    Yes 70% 30% 20% 80%

    No 30% 70% 80% 20%

    FrequencyUsing three variables to reveal

    http://goforward/http://find/http://goback/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    24/26

    distributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Using three variables to reveal

    suppressed association

    Among men, 70% of those under 50 indicated they

    want to travel abroad, compared with 30% of those50 or older

    The pattern was reversed for women, where 20% of

    those under 50 indicated they want to travel abroad

    as opposed to 80% of those 50 or older Since the association between wanting to travel

    abroad and age runs in the opposite direction for

    males and females, the relationship between these

    two variables is masked when the data are

    aggregated across gender

    But when the effect of gender is controlled, the

    suppressed association between desire to travel

    abroad and age is revealed for the separate

    categories of males and females

    Frequencydi ib i dEating frequently in fast-food

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    25/26

    distributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Eating frequently in fast-food

    restaurants by family size

    Consider the cross-tabulation of family size and the

    tendency to eat out frequently in fast-food

    restaurants no association is observed

    Visit fast-food restaurants Family size

    frequently Small Large

    Yes 60% 60%

    No 40% 40%

    Frequencydi t ib ti dEating frequently in fast-food

    http://find/
  • 7/27/2019 7 Frequency Distributions Cross Tabulations

    26/26

    distributions andcross-tabulations

    Dr James Abdey

    Overview

    Frequency distribution

    Types of Variable

    Correlation and association

    Cross-tabulation

    Eating frequently in fast food

    restaurants by family size and income

    Introducing income as a third variable in the analysis;

    again, no association is observed

    No change in initial relationship

    Visit fast-food Income

    restaurants Low High

    frequently Small Large Small Large

    Yes 60% 60% 60% 40%No 40% 40% 40% 40%

    http://find/