doane chapter 02

7/30/2019 Doane Chapter 02

1/82


2/82

Data Collection

Data VocabularyLevel of Measurement

Time Series and Cross-sectional Data

Sampling Concepts

Sampling Methods

Data Sources

Survey Research

Chapter

2


3/82

Data Vocabulary

Data is the plural form of the Latin datum(a givenfact).

McGraw-Hill/Irwin 2007 The McGraw-Hill Companies, Inc. All rights reserved.

In scientific research, data arise

from experiments whose resultsare recorded systematically.

Important decisions may depend on data.

In business, data usually arise from

accounting transactions ormanagement processes.


4/82

Data Vocabulary

Sub jects , Variables, Data Sets

We will refer to Data as plural and data setas aparticular collection of data as a whole.

Observation each data value.

Subject(orindividual) an item for study (e.g., anemployee in your company).

Variable a characteristic about the subject orindividual (e.g., employees income).


5/82

Data Vocabulary

Sub jects , Variables, Data Sets Three types of data sets:

Data Set Variables Typical Tasks

Univariate One Histograms, descriptivestatistics, frequency tallies

Bivariate Two Scatter plots, correlations,simple regression

Multivariate More thantwo

Multiple regression, datamining, econometric modeling


6/82

Data Vocabulary

Sub jects , Variables, Data SetsConsider the multivariate data set with

5 variables 8 subjects 5 x 8 = 40 observations


7/82

Data Vocabulary

Data Types A data set may have a mixture ofdata types.

Types of Data

Attribute(qualitative)

Numerical(quantitative)

Verbal LabelX= economics

(your major)

CodedX= 3

(i.e., economics)

DiscreteX= 2

(your siblings)

ContinuousX= 3.15

(your GPA)


8/82

Data Vocabulary

Attr ibu te Data Also called categorical, nominal or qualitative data.

Values are described by words rather than

numbers. For example,

- Automobile style (e.g.,X= full, midsize,compact, subcompact).

- Mutual fund (e.g.,X= load, no-load).


9/82

Data Vocabulary

Data Coding Codingrefers to using numbers to represent

categories to facilitate statistical analysis.

Coding an attribute as a number does notmakethe data numerical.

For example,

1 = Bachelors, 2 = Masters, 3 = Doctorate Rankings may exist, for example,

1 = Liberal, 2 = Moderate, 3 = Conservative


10/82

Data Vocabulary

B inary Data A binary variable has only two values,

1 = presence, 0 = absence of a characteristic of

interest (codes themselves are arbitrary). For example,

1 = employed, 0 = not employed1 = married, 0 = not married

1 = male, 0 = female1 = female, 0 = male

The coding itself has no numerical value so binaryvariables are attribute data.


11/82

Data Vocabulary

Numerical Data Numericalorquantitative data arise from counting

or some kind of mathematical operation.

For example,- Number of auto insurance claims filed inMarch (e.g.,X= 114 claims).

- Ratio of profit to sales for last quarter

(e.g.,X= 0.0447). Can be broken down into two typesdiscrete or

continuous data.


12/82

Data Vocabulary

Disc rete Data A numerical variable with a countable number of

values that can be represented by an integer (no

fractional values). For example,

- Number of Medicaid patients (e.g.,X= 2).- Number of takeoffs at OHare (e.g.,X= 37).


13/82

Data Vocabulary

Cont inuous Data A numerical variable that can have any value

within an interval (e.g., length, weight, time, sales,

price/earnings ratios). Any continuous interval contains infinitely many

possible values (e.g., 426


14/82

Data Vocabulary

Rounding Ambiguity is introduced when continuous data are

rounded to whole numbers.

Underlying measurement scale is continuous. Precision of measurement depends on instrument.

Sometimes discrete data are treated as

continuous when the range is very large (e.g., SATscores) and small differences (e.g., 604 or 605)arent of much importance.


15/82

Level of Measurement

Fou r levels of measu rement for data:

Level of

Measurement Characteristics Example

Nominal Categories only Eye color (blue, brown,green, hazel)

Ordinal Rank has meaning Bond ratings (Aaa, Aab,C, D, F, etc.)

Interval Distance hasmeaning

Temperature (57oCelsius)

Ratio Meaningful zeroexists

Accounts payable ($21.7million)


16/82


Nom inal Measu rement Nominal data merely identify a category.

Nominal data are qualitative, attribute, categorical

or classification data (e.g., Apple, Compaq, Dell,HP).

Nominal data are usually coded numerically,codes are arbitrary (e.g., 1 = Apple, 2 = Compaq,

3 = Dell, 4 = HP). Only mathematical operations are counting (e.g.,

frequencies) and simple statistics.


17/82


Ordinal Measurement Ordinal data codes can be ranked

(e.g., 1 = Frequently, 2 = Sometimes, 3 = Rarely,

4 = Never). Distance between codes is not meaningful

(e.g., distance between 1 and 2, or between 2 and3, or between 3 and 4 lacks meaning).

Many useful statistical tests exist for ordinal data.Especially useful in social science, marketing andhuman resource research.


18/82


In terval Measu rement Data can not only be ranked, but also have

meaningful intervals between scale points

(e.g., difference between 60

F and 70

F is sameas difference between 20F and 30F). Since intervals between numbers represent

distances, mathematical operations can be

performed (e.g., average). Zero point of interval scales is arbitrary, so ratiosare not meaningful (e.g., 60F is nottwice aswarm as 30F).


19/82


L ikert Scales A special case of interval data frequently used in

survey research.

The coarseness of a Likert scale refers to thenumber of scale points (typically 5 or 7).

College-bound high school students should be required to study aforeign language. (check one)

StronglyAgree

SomewhatAgree

NeitherAgree

NorDisagree

SomewhatDisagree

StronglyDisagree


20/82


L ikert Scales A neutral midpoint(Neither Agree Nor Disagree)

is allowed if an oddnumber of scale points is usedor omitted to force the respondent to lean one

way or the other.

Likert data arecoded numerically

(e.g., 1 to 5) but anyequally spacedvalues will work.

Likert coding:

1 to 5 scaleLikert coding:

-2 to +2 scale

5 = Help a lot

4 = Help a little3 = No effect2 = Hurt a little1 = Hurt a lot

+2 = Help a lot

+1 = Help a little0 = No effect1 = Hurt a little2 = Hurt a lot


21/82


L ikert Scales Careful choice of verbal anchors results in

measurable intervals (e.g., the distance from 1 to2 is the same as the interval, say, from 3 to 4).

Ratios are not meaningful (e.g., here 4 is nottwice 2).

Many statistical calculations can be performed(e.g., averages, correlations, etc.).


22/82


L ikert Scales More variants of Likert scales:

How would you rate your marketing instructor? (check one)

Terrible

Poor

Adequate

Good

Excellent

How would you rate your marketing instructor? (check one)

Very Bad Very Good


23/82


Ambigu i ty Grades are usually coded numerically

(A = 4, B = 3, C= 2, D = 1, F= 0) and are used tocalculate a mean GPA.

Is the intervalfrom 3.0 to 4.0 really the same asthe interval from 1.0 to 2.0?

What is the underlying reality ranging from 0 to 4

that we are measuring? Best to be conservative and limit statistical tests to

those for ordinal data.


24/82


Ratio Measu rement Ratio data have all properties of nominal, ordinal

and interval data types and also possess ameaningful zero (absence of quantity beingmeasured).

Because of this zero point, ratios of data valuesare meaningful (e.g., $20 million profit is twice as

much as $10 million). Zero does not have to be observable in the data,

it is an absolute reference point.


25/82


Use the fol low ing p rocedu re torecognize data types:

Question If Yes

Q1. Is there ameaningful zero point?

Ratio data (all statistical operations areallowed)

Q2. Are intervalsbetween scale points

meaningful?

Interval data (common statistics allowed,e.g., means and standard deviations)

Q3. Do scale pointsrepresent rankings?

Ordinal data (restricted to certain typesof nonparametric statistical tests)

Q4. Are there discrete

categories?

Nominal data (only counting allowed,

e g finding the mode)


26/82


Chang ing Data by Recoding In order to simplify data or when exact data

magnitude is of little interest, ratio data can berecoded downwardinto ordinal or nominalmeasurements (but not conversely).

For example, recode systolic blood pressure asnormal (under 130), elevated (130 to 140), or

high (over 140). The above recoded data are ordinal (ranking is

preserved) but intervals are unequal and someinformation is lost.


27/82


Time Series Data Each observation in the sample represents a

different equally spaced point in time (e.g., years,months, days).

Periodicitymay be annual, quarterly, monthly,weekly, daily, hourly, etc.

We are interested in trends and patterns over time(e.g., annual growth inconsumer debit card usefrom 1999 to 2006).


28/82


Cross-sect ional Data Each observation represents a different individual

unit (e.g., person) at the same point in time(e.g., monthly VISA balances).

We are interested in- variation among observations or in- relationships.

We can combine the two data types to getpooledcross-sectional and time series data.


29/82

Sampling Concepts

Sample or Census? A sample involves looking only at some items

selected from the population.

A census is an examination of all items in adefined population.

- Mobility- Illegal immigrants- Budget constraints- Incomplete responses or nonresponses

Why cant the United States Census survey every

person in the population?


30/82

Sampling Concepts

Situations Where A Sample May Be Preferred:

Infinite PopulationNo census is possible if the population is infinite or of indefinite size(an assembly line can keep producing bolts, a doctor can keep

seeing more patients).

Destructive TestingThe act of sampling may destroy or devalue the item (measuringbattery life, testing auto crashworthiness, or testing aircraft turbofan

engine life).Timely Results

Sampling may yield more timely results than a census (checkingwheat samples for moisture and protein content, checking peanutbutter for aflatoxin contamination).


31/82

Sampling Concepts

Situations Where A Sample May Be Preferred:

AccuracySample estimates can be more accurate than a census. Instead ofspreading limited resources thinly to attempt a census, our budget

of time and money might be better spent to hire experienced staff,improve training of field interviewers, and improve data safeguards.

CostEven if it is feasible to take a census, the cost, either in time ormoney, may exceed our budget.

Sensitive InformationSome kinds of information are better captured by a well-designedsample, rather than attempting a census. Confidentiality may alsobe improved in a carefully-done sample.


32/82

Sampling Concepts

Situations Where A Census May Be Preferred

Small PopulationIf the population is small, there is little reason to sample, for the effort ofdata collection may be only a small part of the total cost.

Large Sample SizeIf the required sample size approaches the population size, we might aswell go ahead and take a census.

Legal RequirementsBanks must count allthe cash in bank teller drawers at the end of eachbusiness day. The U.S. Congress forbade sampling in the 2000 decennial

l ti

Database Exists

If the data are on disk we can examine 100% of the cases. But auditing orvalidating data against physical records may raise the cost.


33/82

Sampling Concepts

Parameters and Stat ist ics Statistics are computed from a sample ofn items,

chosen from a population ofNitems.

Statistics can be used as estimates ofparametersfound in the population.

Symbols are used to represent populationparameters and sample statistics.


34/82

Sampling Concepts

Parameters and Stat ist ics

Statistic Any measurement computed from a sample. Usually,the statistic is regarded as an estimate of a populationparameter. Sample statistics are often (but notalways) represented by Roman letters.

Parameter or Statistic?

Parameter Any measurement that describes an entirepopulation.

Usually, the parameter value is unknown since werarely can observe the entire population. Parametersare often (but not always) represented by Greekletters.


35/82

Sampling Concepts

Parameters and Stat ist ics The population must be carefully specified and the

sample must be drawn scientifically so that thesample is representative.

The target population is the population we areinterested in (e.g., U.S. gasoline prices).

Target Populat ion

The sampling frame is the group from which wetake the sample (e.g., 115,000 stations). The frame should not differ from the target

population.


36/82

N n

Finite or In f in i te? A population is finite if it has a definite size, even if

its size is unknown.

A population is infinite if it is of arbitrarily large

size. Rule of Thumb: A population may be treated as

infinite when Nis at least 20 times n (i.e., whenN/n > 20)

Sampling Concepts

Here,N/n > 20


37/82

Sampling Methods

Probability Samples

Simple RandomSample

Use random numbers to select itemsfrom a list (e.g., VISA cardholders).

Systematic Sample Select every kth item from a list orsequence (e.g., restaurant customers).

Stratified Sample Select randomly within defined strata(e.g., by age, occupation, gender).

Cluster Sample Like stratified sampling except strataare geographical areas (e.g., zipcodes).

l h d


38/82

Sampling Methods

Nonprobability Samples

JudgmentSample

Use expert knowledge to choosetypical items (e.g., which employees

to interview).

ConvenienceSample

Use a sample that happens to beavailable (e.g., ask co-worker opinionsat lunch).

S l h d


39/82

Sampling Methods

Simple Random Sample Every item in the population ofNitems has the

same chance of being chosen in the sample ofnitems.

We rely on randomnumbers to select aname.

=RANDBETWEEN(1,48)

S li M h d


40/82

Sampling Methods

Random Number Tables A table of random digits used to select random

numbers between 1 and N. Each digit 0 through 9 is equally likely to be

chosen.

Sett ing Up a Rule

For example, NilCo wants to award cash prizes to

10 of its 875 loyal customers. To get 10 three-digit numbers between 001 and875, we define any consistent rule for movingthrough the random number table.

S li M h d


41/82

Sampling Methods

Sett ing Up a Rule Randomly point at the table to choose a starting

point.

Choose the first three digits of the selected five-digit block, move to the right one column, downone row, and repeat.

When we reach the end of a line, wrap around to

the other side of the table and continue. Discard any number greater than 875 and any

duplicates.

82134 14458 66716 54269 31928 46241 03052 00260 32367 25783

Table of 1,000 Random DigitsStart Here


42/82

82134 14458 66716 54269 31928 46241 03052 00260 32367 25783

07139 16829 76768 11913 42434 91961 92934 18229 15595 02566

45056 43939 31188 43272 11332 99494 19348 97076 95605 28010

10244 19093 51678 63463 85568 70034 82811 23261 48794 63984

12940 84434 50087 20189 58009 66972 05764 10421 36875 64964

84438 45828 40353 28925 11911 53502 24640 96880 93166 68409

98681 67871 71735 64113 90139 33466 65312 90655 75444 30845

43290 96753 18799 49713 39227 15955 46167 63853 03633 19990

96893 85410 88233 22094 30605 79024 01791 38839 85531 94576

75403 41227 00192 16814 47054 16814 81349 92264 01028 29071

78064 92111 51541 76563 69027 67718 06499 71938 17354 12680

26246 71746 94019 93165 96713 03316 75912 86209 12081 57817

98766 67312 96358 21351 86448 31828 86113 78868 67243 06763

37895 51055 11929 44443 15995 72935 99631 18190 85877 31309

27988 81163 52212 25102 61798 28670 01358 60354 74015 18556

19216 53008 44498 19262 12196 93947 90162 76337 12646 26838

28078 86729 69438 24235 35208 48957 53529 76297 41741 54735

34455 61363 93711 68038 75960 16327 95716 66964 28634 65015

53510 90412 70438 45932 57815 75144 52472 61817 41562 42084

30658 18894 88208 97867 30737 94985 18235 02178 39728 66398

S li M th d


43/82

Sampling Methods

With or Without Replacement If we allow duplicates when sampling, then we are

sampling with replacement.

Duplicates are unlikely when n is much smallerthan N.

If we do not allow duplicates when sampling, thenwe are sampling without replacement.

S li M th d


44/82

Sampling Methods

Computer Methods

These arepseudo-random generators because even the bestalgorithms eventually repeat themselves.

Excel - Option A Enter the Excel function =RANDBETWEEN(1,875)into 10 spread-sheet cells. Press F9 to get a newsample.

Excel - Option B Enter the function =INT(1+875*RAND()) into 10spreadsheet cells. Press F9 to get a new sample.

Internet The web site www.random.org will give you manykinds of excellent random numbers (integers,decimals, etc).

Minitab Use Minitabs Random Data menu with the Integeroption.
http://www.random.org/http://www.random.org/


45/82

Using MINITAB to generate random numbers.

S li M th d


46/82

Sampling Methods

Row

Column Data A rrays When the data are arranged in a rectangular array,

an item can be chosen at random by selecting arow and column.

For example, in the 4 x 3 array, select a randomcolumn between 1 and 3 and a random rowbetween 1 and 4.

This way, each item has an equal chance of beingselected.

Sampling Methods


47/82

Dillard's K-Mart Saks

Dollar General Kohl's Sears Roebuck

Federated DeptStores

May Dept Stores Target

J. C Penney Nordstrom Wal-Mart Stores

Sampling Methods

Row

Column Data A rrays Use =RANDBETWEEN function to choose row 3

and column 3 (Target).

Sampling Methods


48/82

Sampling Methods

Random izing a List In Excel, use function =RAND() beside each row

to create a column of random numbers between0 and 1.

Copy and paste these numbers into the samecolumn using Paste Special | Values (to paste

only the values and not the formulas).

Sort the spreadsheet on the random numbercolumn.

Sampling Methods


49/82

Sampling Methods

The first n itemsare a random

sample of theentire list (theyare as likely asany others).

Random izing a List

Sampling Methods


50/82

Sampling Methods

Systemat ic Sampl ing

For example, starting at item 2, we sample everyk= 4 items to obtain a sample ofn = 20 items froma list ofN= 78 items.

Note that N/n = 78/20 4.

Sample by choosing every kth item from a list,starting from a randomly chosen entry on the list.

Sampling Methods


51/82

Sampling Methods

Systemat ic Sampl ing A systematic sample ofn items from a population

ofNitems requires that periodicity kbeapproximately N/n.

Systematic sampling should yield acceptableresults unless patterns in the population happen torecur at periodicity k.

Can be used with unlistable or infinite populations. Systematic samples are well-suited to linearly

organized physical populations.

Sampling Methods


52/82

Sampling Methods

Systemat ic Sampl ing For example, out of 501 companies, we want to

obtain a sample of 25. What should the periodicitykbe?

k = N/n = 501/25 20.

So, we should choose every 20th company from arandom starting point.

Sampling Methods


53/82

Sampling Methods

Strat i f ied Sampl ing Utilizes prior information about the population.

Applicable when the population can be divided

into relatively homogeneous subgroups of knownsize (strata).

A simple random sample of the desired size istaken within each stratum.

For example, from a population containing 55%males and 45% females, randomly sample 120males and 80 females (n = 200).

Sampling Methods


54/82

Sampling Methods

Strat i f ied Sampl ing Or, take a random sample of the entire population

and then combine individual strata estimates usingappropriate weights.

For a population with L strata, the population sizeNis the sum of the stratum sizes:

N= N1 + N2+ ... + NL

The weight assigned to stratumjis

wj= Nj / n For example, take a random sample ofn = 200

and then weight the responses for males bywM= .55 and for females by wF= .45.

Sampling Methods


55/82

Sampling Methods

Cluster Sample Strata consist of geographical regions.

One-stage cluster sampling sample consists of

all elements in each ofkrandomly chosensubregions (clusters).

Two-stage cluster sampling, first choose ksubregions (clusters), then choose a random

sample of elements within each cluster.

Sampling Methods


56/82

Here is anexample of 4

elements sampledfrom each of 3randomly chosenclusters (two-stage

cluster sampling).

Sampling Methods

Cluster Sample

Sampling Methods


57/82

Sampling Methods

Cluster Sample Cluster sampling is useful when

- Population frame and stratum characteristics arenot readily available

- It is too expensive to obtain a simple or stratifiedsample

- The cost of obtaining data increases sharply withdistance

- Some loss of reliability is acceptable

Sampling Methods


58/82

Sampling Methods

Judgment Sample A nonprobability sampling method that relies on

the expertise of the sampler to choose items thatare representative of the population.

Can be affected by subconscious bias (i.e.,nonrandomness in the choice).

Quota samplingis a special kind of judgment

sampling, in which the interviewer chooses acertain number of people in each category.

Sampling Methods


59/82

Sampling Methods

Convenience Sample Take advantage of whatever sample is available at

that moment. A quick way to sample.

Sample size depends on the inherent variability ofthe quantity being measured and on the desiredprecision of the estimate.

Sample Size

Data Sources


60/82

Type of Data Examples

U.S. general data Statistical Abstract of the U.S.

U.S. economic data Economic Report of the PresidentAlmanacs World Almanac, Time Almanac

Periodicals Economist, Business Week, Fortune

Indexes New York Times, Wall Street Journal

Databases CompuStat, Citibase, U.S. CensusWorld data CIA World Factbook

Web Google, Yahoo, msn

Data Sources

Useful Data Sou rces

Survey Research


61/82

Step 1: State the goals of the research

Step 2: Develop the budget (time, money,

staff)

Step 3: Create a research design (targetpopulation,

frame, sample size)

Step 4: Choose a survey type and method ofadministration

Survey Research

Basic Steps o f Survey Research

Survey Research


62/82

Step 5: Design a data collection instrument(questionnaire)

Step 6: Pretest the survey instrument andrevise as

needed

Step 7: Administer the survey (follow up ifneeded)

Step 8: Code the data and analyze it

y

Basic Steps o f Survey Research

Survey Research


63/82

y

Survey TypesType of

SurveyCharacteristics

Mail You need a well-targeted and current mailing list

(people move a lot). Low response rates are typicaland nonresponse bias is expected (nonrespondentsdiffer from those who respond). Zip code lists (oftencostly) are an attractive option to define strata ofsimilar income, education, and attitudes. Toencourage participation, a cover letter should clearlyexplain the uses to which the data will be put. Planfor follow-up mailings.

Survey Research


64/82

y

Survey TypesType ofSurvey

Characteristics

Telephone Random dialing yields very low response and is

poorly targeted. Purchased phone lists help reachthe target population, though a low response ratestill is typical (disconnected phones, callerscreening, answering machines, work hours, no-

call lists). Other sources of nonresponse biasinclude the growing number of non-Englishspeakers and distrust caused by scams andspams.

Survey Research


65/82

y


Characteristics

Interviews Interviewing is expensive and time-consuming, yeta trade-off between sample size for high-qualityresults may still be worth it. Interviews must becarefully handled so interviewers must be well-trained an added cost. But you can obtain

information on complex or sensitive topics (e.g.,gender discrimination in companies, birth controlpractices, diet and exercise habits).

Survey Research


66/82

y


Characteristics

Web Web surveys are growing in popularity, but are

subject to nonresponse bias because those whoparticipate may differ from those who feel too busy,dont own computers or distrust your motives

(scams and spam are again to blame). This type ofsurvey works best when targeted to a well-definedinterest group on a question of self-interest (e.g.,views of CPAs on new proposed accounting rules,frequent flyer views on airline security).

Survey Research


67/82

y


Characteristics

Direct

Observation

This can be done in a controlled setting (e.g.,

psychology lab) but requires informed consent,which can change behavior. Unobtrusiveobservation is possible in some nonlab settings(e.g., what percentage of airline passengers carryon more than two bags, what percentage of SUVs

carry no passengers, what percentage of driverswear seat belts).

Survey Research


68/82

Plan What is the purpose of the survey?Consider staff expertise, needed skills,degree of precision, budget.

Design Invest time and money in designing thesurvey. Use books and references to

avoid unnecessary errors.Quality Take care in preparing a quality survey

so that people will take you seriously.

y

Survey Guidel ines

Survey Research


69/82

Pilot Test Pretest on friends or co-workers to makesure the survey is clear.

Buy-in Improve response rates by stating thepurpose of the survey, offering a token ofappreciation or paving the way withendorsements.

Expertise Work with a consultant early on.

y

Survey Guidel ines

Survey Research


70/82

y

Gett ing Advice Consider hiring a consultant in the early stages.

Many resources are available to help- The American Statistical Association

- The Research Industry Coalition

- The Council of American Survey Research Organizations

Survey Research
http://www.amstat.org/http://www.researchindustry.org/http://www.casro.org/http://www.casro.org/http://www.researchindustry.org/http://www.amstat.org/


71/82

Use a lot of white space in layout. Questionnaire Design

Begin with short, clear instructions.

State the survey purpose.

Assure anonymity.

Instruct on how to submit the completed survey.

Survey Research


72/82

Questionnaire Design Break survey into naturally occurring sections.

Let respondents bypass sections that are notapplicable (e.g., if you answered no to question 7,

skip directly to Question 15).

Pretest and revise as needed.

Keep as short as possible.

Survey Research


73/82

Questionnaire DesignType of Question Example

Open-ended question Briefly describe your job goals.

Fill-in-the-blank How many times did you attend formalreligious services during the last year?________ times

Check boxes Which of these statistics packageshave you ever used? SAS Visual Statistics SPSS MegaStat Systat Minitab

Survey Research


74/82

Type of Question Example

Questionnaire Design

Ranked choices Please evaluate your dining experience

Excellent Good Fair Poor

Food

Service

Ambiance Cleanliness

Overall

Survey Research


75/82

Type of Question Example

Pictograms What do you think of the Presidentseconomic policies? (circle one)

Questionnaire Design

Likert scale Statistics is a difficult subject.Neither

Strongly Slightly Agree Nor Slightly StronglyAgree Agree Disagree Disagree Disagree

Survey Research


76/82

The way a question is asked has a profoundinfluence on the response. For example,

1. Shall state taxes be cut?

2. Shall state taxes be cut, if it meansreducing highway maintenance?

3. Shall state taxes be cut, it is means firingteachers and police?

Quest ion Wording

Survey Research


77/82

Make sure you have covered all the possibilities.For example,

Are you married? Yes No

Overlapping classes orunclear categories are aproblem. For example,

How old is your father? 35 45 45 55

55 65 65 or older

Quest ion Wording

Survey Research


78/82

Responses are usually coded numerically(e.g., 1 = male 2 = female).

Missing values are typically denoted by special

characters (e.g., blank, . or *). Discard questionnaires that are flawed or missing

many responses.

Watch for multiple responses, outrageous orinconsistent replies or range answers.

Follow-up if necessary and always document yourdata-coding decisions.

Cod ing and Data Screening

Survey Research


79/82

Source of Error Characteristics

Nonresponse bias Respondents differ from nonrespondents

Selection bias Self-selected respondents are atypical

Response error Respondents give false information

Coverage error Incorrect specification of frame orpopulation

Interviewer error Responses influenced by interviewerMeasurement error Survey instrument wording is biased or

unclear

Sampling error Random and unavoidable

Sources of Error

Survey Research


80/82

Enter data into a spreadsheet or database as aflat file (n subjects x m variables matrix).

Data File Format

Survey Research


81/82

Using commas (,), dollar signs ($), or percents (%)as part of the values may result in your data beingtreated as text values.

A numerical variable may only contain the digits0-9, a decimal point, and a minus sign.

To avoid round-off errors, format the data column

as plain numbers with the desired number ofdecimal places before you copy the data to astatistical package.

Advice on Copy ing Data


82/82

Applied Statistics inBusiness and Economics

End of Chapter 2

doane chapter 02

Documents