2 Survey QuestionsThere are three basic types of questions used within an EBI survey:CategoricalScaledText
3 Categorical Question Allows the data to be divided into groups Example: gender or raceOften referred to as a “demographic question”A categorical question allows respondents to place themselves into exactly one categoryA demographic categorical question characterizes the respondent’s personal attributesA demographic question is a type of categorical question
4 Scaled QuestionsAllow respondents to indicate how strongly they feel about a question or statement using a rating scaleGood for gauging the respondent’s attitudeExample:
5 Text Question Is considered “open-ended” Allows respondents to write an answer in their own wordsExample:
6 Basic Statistical Terms Mean responseStandard deviationFrequency distributionFactor means
7 Mean Response Is an average Is only used on scaled questions Is computed by taking the sum of all the numerical responses to a question and dividing that sum by the number of people who answered the questionN/A (not applicable) responses are omitted from the mean
8 Standard DeviationTells how tightly the responses are clustered around the mean; a measure of dispersal, or variation, in the responsesResponses are clustered near the meanResponses are dispersed from the mean
9 Standard DeviationIf the responses are close to the mean, we expect to see a low standard deviation.This shows the respondents answered the question in a similar manner.In contrast, if the responses are spread across a greater range, the standard deviation will be higher.
10 Standard DeviationHigher standard deviation is often interpreted as higher volatility, because more respondents answered one extreme or the other.In comparison, lower standard deviation would likely be an indicator of stability.The most consistent responses are those with the lowest standard deviation.
11 Frequency Distribution Shows the number or percent of individuals that selected each of the responses for the corresponding questionN showsthe number ofpeople whorespondedKey Text showsthe answersto the question
12 Frequency Distribution Frequency distributions can also be formed by grouping responses together.
13 Frequency Distribution If each category represents an equal portion of the participants, it indicates there is little agreement among the participants.If one category represents a large portion of the participants, then there is a large degree of agreement among the participants. This reflects a common level of satisfaction.
14 Factors A group of questions statistically tied together A factor describes a broad concept more accurately than a single question doesDining Services FactorHow satisfied are you with the following aspects of the dining services?Food qualityCleanlinessAtmosphereHours
15 Dependent and Independent Factors Overall program effectiveness or overall satisfaction is typically the dependent factorThe remaining factors are typically the independent factors
16 RegressionA statistical method used to predict the value of the dependent variable (typically the “Overall Program Effectiveness” factor) by studying its relationship with the independent variables (typically all other factors)Tells us the level of impact the independent factors have on the overall factor
17 Statistical TestingSuppose you examined the mean response to a question asked of male and female participants.In all likelihood, the mean for males will be different than the mean for females.Statistical testing reveals whether the difference occurred as the result of random chance, or if there is a real difference between the way men and women perceive the question.
18 Statistical TestingCategorical – tests for a real difference in the mean of two categoriesExample: men vs. womenLongitudinal – tests for a real difference in the mean from one year to the nextExample: 2009 vs. 2010External – tests for a real difference in the means of your institution and comparison institutionsExample: your institution vs. select 6
19 BenchmarkingAn improvement process in which an institution compares its performance against other institutions and uses the information to improve its own performanceTwo types of benchmarking-internalexternal
20 Internal Benchmarking Longitudinal – shows where you have made improvements and where improvements could be madeUnit/Category – shows which units or categories performed at a higher level
21 External Benchmarking Select 6 – compares your institution to six participating peer institutions of your choiceCarnegie Class – compares your institution to a group of similar institutionsAll Institutions – compares your institution to all other participating institutions combined
22 Contact Us If you have questions or need more information, please contact:EBI Help Desk