-generalizability to a larger population 

-the determanation of statistical, quantifiable differences between groups is important

-statistical analysis of data is required 

• Provides best means for understanding populations of individuals and for describing this understanding in terms of numerical descriptors 
• proper use of sampling: findings observed can confidently be generalized to the population from which the sample was drawn 
• proper sampling plus large sample size provides the basis for numerical and statistical descriptions 
• data analyzed through descriptive and inferential statistical procedures, providing insight into pop attitudes, perceptions and behaviors 

Research is the systematic collection of info from respondents in order to better understand and/or predict some aspect of their attitudes or behaviors 

• Analytical 
• descriptive 
• Both are important to ad researchers in both the private and academic sectors 

Methods for collecting survey info:

• Personal interviews
• Telephone interview 
• mail survey 
• online survey 

• occur when interviewer admins a survey to a respondent in a face-to-face setting 
• intercept/prerecruited 
• when questionairre is very long/complex 
• when use of visual stimuli 
• high cost and time consuming 
• difficult to obtain random sampling 
• (CLOSED ENDED 20 MIN IS TOO LONG) 

• entail the admin of a survey questionaire via phone 
• Assisted by CATI( computer assisted Telephone Interviewing) 
• Lower marginal cost ( not neccissarily true) 
• true random sampling 
• questionaire must be relativly short and simple 
• difficult to collect sensitive data, like income 

• entails mailing each potential respondent a package containing a cover letter, the survey, questionaire, instructins for completion and return, a stamped envelope adressed to the research co. 
• cost efficenecy 
• respondent convenience 
• low response rate (20-25% is good) 
• limited questionaire length and complexity 
• extended timing 

• Using the internet 
  • Survey monkey 
  • social networking 
  • research co. in second life 
• cost and time efficient
  • panel survey w/ incentive 
• no true random sampling 
• need technical support 

Criteria for selection data collection method: 

• Cost?- online and mail 
• timing requirements?-online and telephone 
• sample, interview, admin control?-personal interview 
• sample characteristics? 
  • small homogeneous- personal I and mail 
  • large heterogenous-oline and telephone 
• Accuracy/Complexity of topic & Q? online, mail phone
• interview length? 
  • long-mail and online 
  • short-online and telepone 
• response rate-personal>interview>mail 

• what the researcher is interested in explaining 
• The DV in an experiment is the criterion or measure used to evaluate the influence of the IV 

used to explain changees in the dependent variable.

• manipulated in some systemic way 

• experiments go beyond the what/why of problem in order to arrive at the "why"

• Correlation-variable 'b' has to change in value when variable 'a' changes 
• sequence- vb must appear after va has appeared, or must change after va has changed 
• absence of other causal factors 

Refers to the extent that one can eliminate alternative explanations for the observed experimental results 

-natural 

-survey 

outside:better 

-experiment research as is manipulated 

-inside:better 

• refers to the steps involved in conducting an experiment. an experiment design does not have to be a complicated series of statments, diagrams and figures; it may be as simple as" 

Pretest ---->Experimental treatment----->Postest 

• 2 designs discussed here are often reffered to as "experimental" design 

1. Control group 
2. random assignment 

• Control groups are important becuase almost nothing starts at '0'

AKA one group post-test only 

AKA, one group pre-test post-test 

• Bad design, still cant know for sure wether the experimental treatment was responsible for the change in the DV or if change was a result of extraneous factors 

EXPERIMENTAL  0₁ X 0₂ 

CONTROL         N  N  N 

AKA pretest only with control 

• a good design, both post measures are take at the same time, controlling for history. maturation, instrumentation and testing effects 

BEST EXPERIMENTAL DESIGN 

• AKA pretest-postest with control 
• best, establishes all three criteria for causality, correlation, sequence and control 

• the way that abstract, unobservable concepts are linked to observable events 
• the way researchers use an observable events to draw inferences about concepts that may not be directily observable ie attitude toward a brand 

Step 1: ID and define concept of interest 

Step 2: specify an observable event 

Step 3: evaluate and revise the observable event 

• ID concept of interest 
  • concept-an invented name for property of object, person, state, event 
• Develop conceptual definition 
  • conceptual def-the central/ core idea of concept
• Develop operational defination 
  • operational D: translation of the concepy into one or more observable events by explicity describing the most important observable, defining characteristics of the concept 

• ID the appropriate level of measurement 
  • consists of "rules for assigning #'s to objects in such a way as to represent quantities of attributes 
  • we measure the attributes/characteristics, of objects and not the objects themselves 
  • the definition is broad and does not specifiy exactly how the numbers are to be assigned 
  • numbers must be assigned so they properly reflect the properties of the attributes 

• Nominal 
• ordinal 
• interval 
• ratio 

LOWEST->FREQUENCY->RAW #'S->PIE CHARTS/ BAR GRAPHS 

• #'s merely used to ID attributes, also implies equivalence 
• classification of the measured characteristic or attribute..
• gender, age, income, etc 
• dichotomous, multiple choice, checklist, etc 

% and or bargraph 

• posses ID, equivalence and meaningful order 
• mathmatical meaning: order 
• rank, etc 
• ordinal measures can't provide any insights into the relative distance between ranked objects 

MEAND AND STANDARD DEIVATION 

• all the properties of ordinal scale and intervals btwn adjacent 
• mathmatical meaning with arithmetic operations such as addition/subtraction 
• themometer, time 
• ranking, likert, semantic differential 

HIGHEST 

• scale posses a natural or absolute zero 
• age
• constant sum 

• develop the mesurement instrument 
• evaluate and revise instruments 
  • pilot study should be conducted to asses the reliabilty and validity of any untested measures 

• Nominal
  • dichotomous questions 
  • multiple-choice questions 
• Ordinal 
• interval 
• ratio 

• in its most common form it asks a respondent to rate an object on a number of itemized, seven-point scales bounded on each end by one of two bipolar adjetives :

The Celebrity was..

-Attractive  1 2 3 4 5 6 7 Unattractive 

-Not Classy 1 2 3 4 5 6 7 Classy  

• Asks a respondent to indicate his degree of agreement or disagreement with a series of specific statements: 

The Celebrity  St. Agree  agree  neutral .....

Trustworthy: 1       2          3     .....

Attractive   :               1       2          3 

• Question must be truly open-ended 
  • how satisfied are you w/ product____ 
  • what are your thoughts towards____
• Must explicity probe responses: 
  • is there anything else? 
  • does anything else come to mind? 
• should be single-minded:
  • do any strengths of sauve come to mind? 

• Explicitly state respondents task in simple language 
• use simple active sentences and common language 
• avoid bias 
• avoid multiple informational requests in a single question 
• avoid assumptions 
• justify requests for personal info 

• An explicit/ implicit reference to importance 
• general info on rationale and goals of research 
• an explicit request for participation 
• reassurence that he task of participating is not too burdensome or time consuming 
• need for truthful answers 
• promise of confidentiality 
• reassurence that this is real, legitimate research 

• Begin with simple, non threatening, interesting, easy to answer questions 
• group questions on the smae topic together; complete one topic before moving to another 
• within a topic move from the most general to specific 
• place difficult/ sensitive questions at end 
• avoid biasing q's appearing later in the questionaire with q's asked earlier in the questionaire 
• address the most important topics first 

• Paper stock 
• quality of paper stock makes important impression 
• white always acceptable 
• paper should be 8.5x11" 
• standard number 10 business envelope best for most surveys. 

• Formatting affects the general appearence of the instrument and the likelihood that respondents will complete and return 
• layout 
• break up into sections 

• Dont send a fat envelope 
• think as to how enclousures are inserted 

A portion, piece, or segment regarded as representation of the whole; a small part of anything designed to show the style, quality, and nature of whole 

• The population of interest is small and identifiable 
• sampling migh eliminate important cases from the study 
• credibility requires the consideration of all memebers of the target population

• Can't survey the whole population or universe 
• characteristics of samples can be generalized to populations 

• a group of or class of subjects, variables, concepts or phenomena 
• parameter: a characteristic of population 

• A subset of the population that is representative of the entire population
• static: characteristic of a sample 

1. sample can consist of animated or unanimated objects 
2. can be evaluated in terms of efficiency
3. can be evaluated in terms of generalizations (most important in sampling) that it provides 

• The demographic of the target pop definition specifies the relevant age, gender, income or other related characteristics of the pop of interest 

• specifies the geographic areas in which the target population resides 

• specifies relevant category-or product-related behavior 

are quick and inexspensive to obtain. Research conducted amoung nonprobability samples is easy to design and carry out. ( cannot calculate sampling error and limited confidence in generalizing the findings) 

• Simple random 
• systematic random 
• stratified random 

• frequently used in ad research 
• each member of the pop has a known equal chance for inclusion in the research 
• a simple random sample can provide comparable estimates w/o having to survey entire population

• typically provide data identical to simple random samples woth added advantage of simplicity typically provide data identical to simple random samples woth added advantage of simplicity 
• step: -count the number of elements on lst  -determine desired sample size                   -compute a skip interval                             -select a random place on list                     -select and interview each element at            appropriate skip interval

• It must be used when you suspect that large variation in the variable being studied and that this variation is due to or correlates with observable characteristics in the universe being sampled 
• three step process 
  1. the classification criteria that defines the strata are ID 
  2. each element in the sample frame is assigned to one and only one stratum 
  3. independent random samples are selected from each stratum 

• an estimate of how statics may be expected to deviate from parameters when sampling randomly from a given population. 
• it occurs when measurement taken from a sample do not correspond to what exists in the population.
• an error free sample is highly unlikely 
• process of determining with a certain amount of confidence the difference between a sample and the target population 

an estimate, plus or minus, of the value of the population estimate; it states the range in which we beliebethe true population estimate lies. 

• the mathematical expression of our confidence that the population lies within the confidence interval 
• a confidence level of 95% means that there is a 95% probability that the op estimate from research lies within the defined confidence interval 

the frequency with which observations in a sample are assigned to each category or point on a measurement scale 

Basic Concepts for Sampling Error: Population Distribution 

The frequency with which all units or observations n a population would be assigned ot expected in each category or point on a measurement scale 

• The frequencies with which a particular value of a static would be expected when sampling randomly from a given pop
• distrubutions of sample means (frequency) 
• standard ettor is the Standard deviation of the sampling distribution 

• A definition of a particular functional relation between 1) deviations about the mean of a distribution and 2) the probability of these different deviations occuring 
• theoretical distribution

SE=√pq/n

p=sample %

Q=100-p

• Data Analysis plays an important role in the turning quantity of data into defensible, actionable sets of conclusions and reports. 
• Obtain info and insights from data 

• An unorderd array of responses 

• A typical frequency distribution table consists of two columns- the possible values of the variable in the form of categories and the number of occurences, or frequency, of each value 

mean 

median 

mode 

• Mean is perferred in general 
• Median
  • when thare are one or more extreme value in a data set 

• Measures describe the way in which the scores are spread out about a central port. 
• The range is the difference between the highest and lowest scores in a distribution of scores. 
• variance provides a mathematical index of the degree to which scores deviate from, or are at variance with, the mean. 

s²=∑(x-X)²

    N-1

Problem: the variance is expressed in terms of squared deviations from the mean rather than in terms of the original measurements 

• the square root of the variance 

√S²=√ ∑(x-X)²

     N-1

• SD overcomes the problems of variance not being calibrated inthe same unit as the original data by taking the sq root of the variance 
• SD has the values that are smaller and easier to work with than variance 
• SD provides a basis for estimating the probability of how frequently certain scores will occur based on sampling. 

• The main purpose for summary and descriptive methods of quantitative analysis are to reduce data and the enable us to see patterns and trends.
• What can be done with the data depends on which level of measurement they are.