Types of Data (Measurement) AMA Collegiate Marketing Research Certificate Program.

Post on 03-Jan-2016

217 views 0 download

Tags:

Transcript of Types of Data (Measurement) AMA Collegiate Marketing Research Certificate Program.

Types of Data (Measurement)

AMA Collegiate Marketing Research Certificate Program

Module Objectives

• To introduce the four types of data and their characteristics

• Provide examples• Traits and measurement

strengths

Types of Data/Measurement Scales

These scales differ on what they measure and how the data may be analyzed and interpreted

• When collecting or gathering data we collect data from individual cases on particular variables

• A variable is a unit of data collection whose value can vary

• Variables can be defined into types according to the level of mathematical scaling that can be carried out on the data

• There are four types of data or levels of measurement: Nominal, Ordinal, Interval and Ratio data

Introduction: Types of Data

Nominal Data

Nominal Data

Data comprised of categories that cannot be rank ordered – each category is just different

– Male or female– Own Home vs. Rent vs. other – Bought car in last 12 months/didn’t buy– Own a smart phone/don’t own

Nominal Data

Merely measure the presence or absence of something

Nominal Data: An Example

Nominal categories aren’t hierarchical, one category isn’t “better” or “higher” than another

Marital status

□ Single (never married)

□ Married

□ Divorced

□ Widow/Widower

Nominal Data

• Assignment of numbers to the categories has no mathematical meaning

• Male could be coded “0” and female “1” or maybe “1” vs. “2”

What is your gender?

Nominal Data

Must ensure that each category is mutually exclusive and the system of measurement needs to be exhaustive

–18-24 years old, 24-34 (oops)–18-24, 25-34, 35-44, 45+ (correct)

*Assuming no minors in study

Nominal Data• Nominal data are usually

represented “descriptively” • Graphic representations include

tables, bar graphs, pie charts

Other Examples

• Home ownership status• Race/ethnicity• Any kind of behavior

that is yes/no (have you been to the movies in last 30 days)

• Children in household• Employment status• Media usage• Purchase channels• Activities/hobbies

Measurement Traits/Strength

• Weakest type of data• Frequencies, mode• Cross-tabulations• Common to use as

grouping variable for other analyses (gender and average satisfaction)

Ordinal Data

Ordinal data are data comprised of categories that can be rank ordered (i.e., categories can be ranked above or below each other) = hierarchical data

Ordinal Data

Less than $40,000 $60,000 - $79,999 $100,000 - $119,999 $40,000 - $59,999 $80,000 - $99,999 $120,000 or More     

What is your annual HOUSEHOLD income?

Ordinal Data: Example

• Most Viewed Website• Second Most• Third Most• Forth Most

Note: If you also had number of visitors you would have a ratio scale as well

Ranking Scale is Ordinal

• Favorite Movie Type?– Drama, Action,

Comedy, Romance, Mystery, Horror

– 2nd favorite, 3rd favorite…..• Most Visited Website?

– 2nd most– 3rd most

Can also create a Ranking Score (i.e., like NCAA sports). Example, 3 points for each most important, 2 points for 2nd and 1 point for 3rd.

WEBSITE# Most Visited

# 2nd Most

# 3rd Most

Ranking Score

Facebook 134 130 147 809Google 110 90 77 587Twitter 100 100 67 567Yahoo 77 98 88 515Four-square 9 12 51 102

Other Examples (Ordered Responses)

• Favorite restaurants• Most important buying

attributes• Most preferred buying

channels (retail, online, catalog)

• Most pressing concerns• Highest quality, 2nd…• Most watched, 2nd…• Most visited, 2nd….

Measurement Traits/Strength

• Frequencies, mode, median• Ordering, preference, importance ranking - - offers

additional insight beyond nominal data• But can’t measure distance (only asked to rank

order importance - - so don’t know how much more important between 1st and 2nd, 2nd and 3rd, etc.)

Interval Data

• Interval data are measured on a continuous scale with no true zero point (a complete absence of the phenomenon being measured)

• Equal distance between interval points on scale

Interval Data

Interval Data

–IQ tests • A person can’t have zero

intelligence or zero self esteem

• 120 IQ not twice as intelligent as 60

Semantic Differential

ImportanceQuality

Some Examples of Interval Scales

Measurement Traits/Strength

• Frequencies, mode, median, PLUS mean/average and standard deviation–Mean agreement of 4.5 on 5-point scale

• Ordering: Mean satisfaction of 4.5 > 4.3• But can’t measure comparative distance

–Satisfaction average of 4.0 isn’t twice as high as 2.0–Why: no absolute zero point

• Agreement scale can be 0-4, 1-5, -2 to +2, etc.

Ratio Data

Ratio data measured on a continuous scale and does have a true zero point

• Examples: Exact….

• Age

• Weight

• Height

Ratio Data

Other Examples (Exact Responses)• Number of times dined out

last month (could be none)• Hours spent on Internet

each day (could be none)• Last price paid for dinner

(could have been free)

Remember, must have a true zero point!

Measurement Traits/Strength

• Frequencies, mode, median, mean/average/standard deviation

• PLUS, also allows for ABSOLUTE comparisons

If Jimmy goes to two movies per week and Scott sees four movies, then Scott sees twice as many movies as Jimmy (2:1 Ratio)

The levels of measurement can be placed in hierarchical order

Hierarchical Data Order

• Nominal data are the least complex and indicate whether objects are the same or different

• Ordinal data maintain the principles of nominal data but add a measure of order to what is being observed

• Interval data build on ordinal by adding more information on the range between each observation by allowing us to measure the distance between objects

• Ratio data add to interval with including an absolute zero

Hierarchical Data Order Summary

The data type or level of measurement influences the type of statistical analysis techniques that can be used when analyzing data

It is possible to recode or adjust certain types of data into others

Important?

Application?

ALWAYS employ the highest level/strength of measurement available given (1) response rate and (2) ease of answering/remembering

“We are only as strong as we are united, as weak as we are divided.” ― J.K. Rowling, Harry Potter and the Goblet of Fire