what is the once occasion where you must pay attention to sample size

the sample is small

what is the percent requirement a sample size must be of a population

must be 5%

are sample size and sample representativness related?

no

what determines REPRESENTATIVNESS?

the selection method

what determines ACCURACY?

sample size

accuracy of a sample

a measure of how closely it reports the true values of the population it represents

sample error

amount of error that exists within your sample

what does a large sample ensure?

that it has good accuracy; low sample error

what is the only perfectly accurate sample

a census

the higher the accuracy is the lower the ___ ____ is

sample error

can a random sample size be a very tiny part of the population but still be accurate?

yes

what does the size of a random sample depend on?

1. clients desired accuracy (acceptable sample error) 2. cost of data collection for that sample size

confidence interval approach

the most correct method by which to determine sample size

higher accuracy = lower___

lower error

higher error = lower____

lower accuracy

what does sample error depend on

the size of the population

what happens when a sample size reaches 1000?

very little gain in accuracy occurs as it gets even bigger... even if you increase it to 2000 or 3000.... law of diminishing returns

sample error formula

1.96 x (sq. rt( (p*q)/n)))

variability

amount of similarity or dissimilarity in respondents answers to a particular question

what does p mean

the percent of people saying "yes"

what does q mean

100-p

what is the maximum variability in a population

50/50

what is the minimum variability in a population

90/10

what does increased variability mean for sample size

the more variability in a population, the larger will be the required sample size

what does n mean

the sample size

confidence interval

defines endpoints based on knowledge of the area under a bell-shaped curve

95% confidence interval=

1.96

99% confidence interval=

2.58

confidence interval verbalized

I am 95% sure that 88% of the students enjoyed spring break +/- 20%

are sample size and size of the population related

usually no

3 items needed to compute sample size

1. variability 2. acceptable sample error 3. confidence level

what is e

acceptable sample error/ expressed as a percent

what is z

the confidence interval

acceptable sample error

the amount of sample the researcher will permit to be associated with the survey

how do you estimate variability?

use prior experience, research, and intuition

arbitrary approach to sample size

determines sample size by relying on erroneous rule of thumb

good/bad of arbitrary sample sizes

good- simple/easy to apply

bad- not efficient/not economical

bad- not efficient/not economical

conventional approach to sample size

looks at previous/similar sample sizes from studies and uses those sizes

good/bad of conventional sample size

bad- ignores special circumstances/may perpetuate other surveys mistakes/may be too small or too large-- doesn't seem to be anything good

statistical analysis approach to sample size

use particular analysis- such as subgroups- in order to get a good ample size

if you calculate n=(z^2(p*q))/e^2) and n=5% what do you need to do

run the answer through a new equation n(sq.root N-n/n-1))

if you're n=5% and you run it through the new N equation, what will likely happen to your new n

the n will likely decrease

data entry

the creation of a computer file that holds the raw data taken from all of the questionnaires deemed suitable for analysis

data coding

identification of codes that pertain to the possible responses for each question in the questionnaire

summarization def + example

use of measures and statistical values to describe the data matrix i.e. average, mean, median, mode

ex: the average respondent's age is 44

ex: the average respondent's age is 44

conceptualization def + example

use of words or graphics that managers can relate to

ex: the pie graph shows that few respondents are younger than 30 years of age

ex: the pie graph shows that few respondents are younger than 30 years of age

communication def + example

describes the underlying patterns or relationships.. basically describe patterns that you see

ex: More women are in college than men

ex: More women are in college than men

generalization def + example

indicates how sample findings relate to the population

ex: this means that from 32% to 40% of the target market purchases our brand on a regular basis

ex: this means that from 32% to 40% of the target market purchases our brand on a regular basis

data matrix

coded, raw data from the survey

5 types of statistical analysis to reduce the data matrix

1. descriptive analysis 2. inferential analysis 3. differences analysis 4. associative analysis 5. predictive analysis

4 benefits of data summarization

1. It condenses the data 2. It applies understandable conceptualizations 3. it communicates underlying patterns 4. it generalizes sample findings to the population

descriptive analysis def + example

summarizes the basic findings of a sample (mean median mode frequency etc)

ex: describe the typical respondent/ describe how similar respondents are to the typical respondent

ex: describe the typical respondent/ describe how similar respondents are to the typical respondent

inferential analysis def + example

generalize results from target variable and sample population and paste them over the entire population (confidence interval)

ex: estimate the population values

ex: estimate the population values

differences analysis def + example

determine if differences exist; evaluate statistical significance of difference in the means of two or more groups in a sample (t test of differences analysis of variance)

ex: evaluate the difference in means between two groups in the sample

ex: evaluate the difference in means between two groups in the sample

associative analysis def + example

determine how two variables are related (correlation/cross-tabs)

ex: determine if two variables are related in a systematic way

ex: determine if two variables are related in a systematic way

predictive analysis def + example

find complex relationships within the variables in the data set to Forecast the future (multiple regression)

ex: determine the dispositions of several variables' influences on a key variable

ex: determine the dispositions of several variables' influences on a key variable

3 measures of central tendency

mean, median, mode

measures of variability

reveal the typical difference between the values in a set of values

3 examples of measures of variability

1. frequency distribution 2. range 3. standard deviation

what equals the standard deviation squared

variance

(what scale¢ral tendency used with this ques)

what is your gender?

what is your gender?

measurement level: nominal scale

central tendency: mode

central tendency: mode

(what scale¢ral tendency used with this ques)

rank these 5 brands 1-5

rank these 5 brands 1-5

measurement level: ordinal scale

central tendency: median

central tendency: median

(what scale¢ral tendency used with this ques)

on a scale of 1-5 how does Starbucks rate on customer service

on a scale of 1-5 how does Starbucks rate on customer service

measurement level: interval scale

central tendency: mean

central tendency: mean

(what scale¢ral tendency used with this ques)

about how many times did you go shopping last week

about how many times did you go shopping last week

measurement level: ratio scale

central tendency: mean

central tendency: mean

statistics vs. parameters

statistics are sample values vs. parameters are corresponding population values

parameters

values that are computed from a complete census and are precise and valid measures of the population

statistics

values that are computed from a complete census

inference

a form of logic where you make a generalization about an entire class based on what you have observed from a small set of members from that class

2 things an inference is based on

1. sample size

2. variability

2. variability

what format are population parameters in?

greek letters

what format are sample statistics in?

roman letters

what does statistical inference take into account?

that large, random samples are more accurate than small ones

as "n" increases, what happens to error?

error decreases

as variability increase what happens to error?

error increases

2 types of statistical inference

1. parameter estimation

2. hypothesis testing

2. hypothesis testing

standard error is a measure of what

variability

what do you never solve for with SPSS

PERCENTAGE

s_xbar

standard error of the mean

s_p

standard error of the percentage

what does the range of your estimate of the population depend on? (2 things)

sample size

variability

variability

5 steps to computing a confidence interval

1. determine the sample statistic 2. identify the sample size 3. determine the variability 4. decide on the level of confidence 5. perform computations to determine the upper and lower boundaries of the confidence interval range

5 steps to intuitive hypothesis testing

1. state what you believe (90% of students hate marketing research)

2. draw a random sample and determine the sample statistic (from a class of 100, 90 hate marketing research)

3. compare the statistic to the parameter (90 is the same as 90!)

4. decide whether or not the sample supports the hypothesis (everyone hates marketing research as much as I thought!)

5. If the sample doesn't match the hypothesis, revise it to match (nope, its clear that nearly everyone hate marketing research)

2. draw a random sample and determine the sample statistic (from a class of 100, 90 hate marketing research)

3. compare the statistic to the parameter (90 is the same as 90!)

4. decide whether or not the sample supports the hypothesis (everyone hates marketing research as much as I thought!)

5. If the sample doesn't match the hypothesis, revise it to match (nope, its clear that nearly everyone hate marketing research)

sampling distribution concept

our sample is one of many theoretical samples that comprise a bell-shaped curve with the hypothesized value as the mean

directional hypothesis

one where you say the hypothesized mean is greater than or less than some amount (i.e. not equal)

market segmentation is based on

differences between groups and consumers

statistical signfigance of differences

the differences found in the sample can be generalized to the population

4 requirements of differences between groups in order for them to be useful

must be:

1. statistically significant

2. meaningful

3. stable

4. actionable

1. statistically significant

2. meaningful

3. stable

4. actionable

when should a t-test be used

when the sample size is 30 or less

what test do you use if the sample size is greater than 30?

z tests

what is x

the mean