RESEARCH DISCIPLINE
Marketing Research Project Statistics
From Questionnaire Items to Scale Scores
Consumer questionnaires analyzed from the first item check to the final model, with conjoint analysis in marketing research for choice data already collected: respondents screened, reverse-worded items scored correctly, brand scales reported with their reliability, and groups compared with estimates and intervals.
No Commitment Required
Strictly Confidential
In-House PhD-Level Statisticians
Fixed Project Fee
Same-Day Response
How Is a Consumer Questionnaire Analyzed?
Marketing research project statistics begins with the questionnaire as a measuring instrument: each respondent answers batteries of rating items meant to capture attitudes such as brand image, perceived quality or purchase intention. No group is compared and no model of how the scales relate is fitted until the responses have been screened, each scale scored and its reliability estimated. A choice-based conjoint task is a questionnaire of another kind, in which respondents choose among product profiles, so choice data already collected are screened and modeled with the same care. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, relies on in-house PhD-level statisticians for this work and documents every screening and scoring rule it applies. The questionnaire’s design and the brand conclusions drawn from it are decisions the researcher keeps.
WHEN TO ENGAGE
Stages of a Consumer Questionnaire Project
What a questionnaire study needs from statistics changes as it moves from draft items to returned responses, more than it changes by topic.
Draft Questionnaire The items and scales are drafted and not yet fielded, so scale lengths, the pretest plan and the target number of respondents can still be changed.
Responses Just Collected Data have come back from a panel or a customer list, and the responses need screening, scoring and a first look at each scale before anything is compared.
Brand Scales to Compare Brand perception was measured in several groups or for several brands, and the comparison needs scores that mean the same thing across them, reported with intervals.
Model Questioned in Review An advisor has asked whether the scales are reliable enough or the sample large enough for the planned model, and the answer has to come from this dataset.
MEASUREMENT AND MODELS
From Raw Responses to Brand Scale Scores
The work follows the order in which questionnaire data are normally trusted: the responses first, the scales second and the relationships among them last.
Response Screening
Straightlining, speeding, duplicate entries and impossible answers are flagged, and the number of respondents caught by each rule is reported beside the results.
Missing and Reversed Items
Skipped items and reverse-worded items are handled by a stated scoring rule, so every scale score can be recomputed exactly from the raw file.
Scale Reliability
Internal consistency is estimated with an interval through Reliability Analysis (Cronbach's Alpha), and its dependence on the number of items is stated.
Conjoint Choice Data
In choice-based conjoint analysis, part-worth utilities by attribute level come from a choice model that treats each respondent's repeated choices as clustered, as in Marketing research.
Relationships Among Scales
When the question links the scales, a model is fitted with Structural Equation Modeling (SEM) only if the sample and the scale quality can support its parameters.
Program and Journal Conventions
Tables and wording follow the conventions of your program or journal within Business, Management & Economics, with sample sizes shown for every estimate.
COMMON ISSUES
Questionnaire Problems That Weaken a Study
A questionnaire study usually weakens at the item or at the respondent, which is easy to miss once the answers have been summed into a score.
Straightlining Left In Respondents who give one answer to a whole battery inflate agreement and the correlations among scales, so they are identified and results are shown with and without them.
Reverse Items Scored Wrong A reverse-worded item left unrecoded lowers a scale's reliability or reverses its meaning, so the scoring is checked against the item wording and not only the output.
Importance Tied to Level Range In conjoint results, an attribute's relative importance depends on the range of levels offered, so a wide price range inflates price, and importance is reported with that range stated.
Alpha Quoted Without Context A reliability coefficient above a conventional cutoff is quoted as proof of quality, though it rises with the number of items and says little about what the items measure.
Agreement Bias Unchecked Respondents who tend to agree with any statement lift scores on positively worded scales, so the balance of item wording and the pattern of agreement across items worded in opposite directions are examined.
Familiar Brands Rated Higher Respondents tend to rate brands they know more favorably, so a comparison of brands is shown alongside familiarity or restricted to respondents who know each brand.
DELIVERABLES
Scored Scales and Tables for Your Report
Every scale score and comparison in the report can be recomputed from the raw responses with these documents, including which respondents were set aside.
Screening Log The screening rules, the number of respondents flagged by each, and the results with and without them, so the effect of every rule is visible.
Scoring Key For every scale, which items are summed, which are reversed before scoring, how skipped items are treated and the range the final score can take.
Reliability Table A reliability estimate with its confidence interval for each scale in this sample, with the number of items and respondents behind it.
Group Comparison Tables Differences in brand scale scores between groups or brands, or part-worth utilities from a conjoint study, with 95% intervals and the respondents behind each.
Structural Model Summary Where the question calls for it, the fitted model with its estimates, fit information and the reasoning on sample size behind it.
Questionnaire Analysis Code The screening, scoring and table steps in one runnable file, applied to the exported responses, from the first screening rule to the last table.
ILLUSTRATIVE EXAMPLE
Comparing Brand Perception Across Two Shopper Groups
An illustration of how a questionnaire dataset is screened and scored before two groups are compared.
ILLUSTRATIVE EXAMPLE · SPECIALTY COFFEE SHOPPER ATTITUDE RESEARCH · UNITED STATES
A master's researcher in marketing surveyed customers of a specialty coffee chain with a questionnaire on brand image and perceived quality, and wanted to compare customers of its downtown and suburban stores.
The scoring rules were recorded first, including which items were reverse-worded. The complication came at the reliability stage: one scale looked unreliable until a reverse-worded item was found to have been entered without recoding, and a group of respondents had given the same answer to an entire battery. The item was recoded, the straightlining respondents were flagged under a rule fixed before looking at group differences, and the comparison was reported with and without them. Downtown respondents were younger than suburban ones, so the difference was also shown adjusted for age, and the results were described as differences between store locations, not effects of location.
Outcome: The researcher received the screening log, the scoring key, reliability estimates with confidence intervals, the group differences with their 95% confidence intervals and the code that reproduces them.
This example does not describe a real client or project. We never share details of actual client work, in line with the confidentiality we promise every researcher. This illustrative example is included to show the type, scope, and standard of work involved in this kind of analysis.
SOFTWARE
Choosing Between SPSS, R and Latent Variable Tools
The choice depends on what the questionnaire analysis ends with, a table of group differences or a latent variable model, and on what your program expects.
HOW WE ASSESS YOUR PROJECT
We Review Your Actual Research
Your assessment is based on a review of your actual research materials and analytical requirements, not a generic project category.
You Submit Your Project Details
Share whatever is available at the current stage, such as the research aim, protocol or proposal, dataset, variable list, manuscript or statistical requirements.
We Review Your Research and Define the Scope
Your statistically trained Research Coordinator reviews the materials to understand the study design, research questions, data structure, analytical requirements and expected deliverables. If clarification is needed, additional documents or information can be requested before the assessment is finalized.
Your Analysis Is Completed and Delivered
Once the scope and fixed project fee are agreed, the in-house statistician completes the work and the Research Coordinator remains the communication point through delivery, with the analysis, deliverables and timeline as agreed.
RESEARCH RESPONSIBILITY
Your Questionnaire Study, Your Conclusions
Brand meaning is not a statistical output: the screening, scoring and models are delivered as agreed, and the questionnaire's content and brand conclusions belong to you.
Agreed Statistical Support OnlineSPSS provides the agreed methodological, analytical and statistical reporting support, defined in a scope that is confirmed before any work begins.
No Intellectual Contribution OnlineSPSS does not claim any intellectual contribution to your questionnaire study. Authorship of the thesis, project report or article remains yours.
Your Research Decisions The research question, the content of the questionnaire, the interpretation of brand findings and the conclusions remain yours, with overall responsibility.
Confidential Handling Research data and project details are handled confidentially, project information is not published, and an NDA is available where required.
FREQUENTLY ASKED QUESTIONS
Respondents, Ratings and Brand Comparisons
How many people to ask, whom to set aside and how to treat one person's ratings of several brands all have statistical consequences for every comparison.
What does statistical support for a marketing research project cover?
The analysis of consumer questionnaire data covers respondent screening, scale scoring, reliability estimation, group comparisons, conjoint analysis in marketing research and, where the question requires it, a model that links the scales. Each rule applied to the data is documented, so that every step can be reproduced and defended.
How many respondents does a brand perception questionnaire need?
There is no single number. It depends on how precisely the comparison must be estimated, which means how wide an interval the study can accept for a difference between groups, on the number of groups and, if a model is fitted, on the number of parameters. Rules of thumb tied to the number of items are rough guides, and the precision of the key estimates is planned instead.
Should respondents who rush through the questionnaire be removed?
Rushing is better flagged than deleted. Screening rules based on completion time, straightlining or failed attention checks are fixed before outcomes are examined, and the main results are shown with and without the flagged respondents. Removing people because their answers look inconvenient can bias the findings, so the rule has to be independent of the result.
Can ratings of several brands from the same respondents be compared directly?
Yes, but not as if different people had given them. When each respondent rates several brands, the ratings are correlated, so differences are estimated within respondents with a model that accounts for each person rating more than one brand, and a respondent’s general tendency to rate generously is separated from views about any one brand. Treating the ratings as independent groups misstates the uncertainty of every comparison.
How do you protect my research data and confidentiality?
Your data and project details are handled confidentially, and project information is not published. Researcher names and institutions are not used publicly without permission. An NDA is available where required and can be arranged before sensitive materials are shared. Read more about confidentiality and data security.
What happens when I request an assessment, and how long does it take?
You share your research question, design and relevant materials, such as the questionnaire, the response file and the sources of the scales, and a statistically trained Research Coordinator reviews the request and responds the same day. The analysis, deliverables, timeline and one fixed project fee are then confirmed before any work begins, so the timeline follows from the agreed scope.
Still have a question? Find more answers in our full FAQ.
NEXT STEP
Brand Scores Start
at the Item Level
A brand score inherits every reversed item and careless respondent summed into it, which is why the item level is examined before any test is chosen.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014