Cross-Tabulation: Statistical Analysis

Transcription Global can provide tables, and customised cross-tab analysis, using data collected from quantitative surveys. We can also do this for structured data collected using qualitative surveys. We offer a fast and cost-effective service which is designed to meet our clients’ needs in a customised and accurate manner. We are responsive to how clients plan and intend to use the data, and are pro-active in suggesting useful cross-breaks within this process.

We work with SPSS to perform the cross-analysis, and report in Excel. Excel has become the standard platform for managing market research data, which can then readily be imported into PowerPoint, dashboards and other business analysis programs. We can analyse both simple and complex levels of quantitative survey data, with outputs designed to meet the analysis needs of our clients. In addition, we can also provide a range of multivariate analyses. These include mapping techniques, factor analysis, cluster analysis and other types of analyses offered by SPSS. Please contact us to discuss your needs and requirements further.

We have significant experience in analysing survey data from studies conducted for clients working in both the consumer and pharmaceutical sectors.

The following sub-categories describe the types of ‘tables and analytics’ services we can offer:

Experimental Design

Experimental design generally involves applying a Latin square plan which is tailored to the number of concepts to be tested, used particularly in the product testing phase. In pharmaceutical market research, samples sizes are often fairly small in comparison to other consumer research. If your sample size is reasonably similar, and has similar tastes, they may well answer in a pretty similar way and you will get statistically significant results. If their answers are very wide ranging, you will need a bigger sample to determine the statistically significant pattern.

Data Preparation

It’s recommended that prior to new data analysis, its quality must be tested and verified. Are there any routings which have been missed? Have any respondents not completed the interview or shown signs of respondent fatigue? How consistently have respondents answered questions?

Patterns in the data need to be assessed to understanding how different types of respondent answered the questionnaire and the distributions of responses, in order to shape subsequent analysis.

New variables may need creating in order to define meaningful respondent segmentations. For branding and advertising projects, it is important to categorise people according to their brand usage. If the study has included a cluster analysis, a new variable should be created, which defines respondents’ cluster membership. These segmentations could then be added to banners when generating the tables.

Coding work may also be required. For example the coding of open-ended responses, where verbatim comments are exported into Excel, then coded. These codes are then re-introduced into the SPSS data-set. Alternatively, it could be to do with coding responses against a secondary data-base, such as with geographical information, prescribing information or vehicle database. Weighting the data against reliable universe statistics might also be required.

All this work may be done in an accessible and transparent fashion. Key data can be exported from SPSS into Excel to be available for all members of the project team to consider.

Presentation Friendly Tables

The tables produced for some international quantitative projects can sometimes result in masses of numbers and data sets, of which only a very small proportion will actually be required for the end presentation. This usually results in research executives being tasked with producing a condensed and more relevant presentation after having sifted through this enormous data set.

The tables themselves are not the required end product which is why we typically work to produce the minimum required number of tables which can then easily be incorporated into a PowerPoint presentation or other reporting device, of course if the full data set is required, we can oblige. We do this by providing the tables within Excel and separating raw counts from percentages. We also work closely with the analyst in the exact definition of the tables, describing what the data looks like and producing a first dry run of the tables for comments and feedback, remembering that this is your presentation.

Strategically, using tables within Excel produces a standard platform for popular business visualisation tools and analytic approaches. With PowerPoint, the data entry sheet uses an Excel format, allowing the importation of data directly from Excel to PowerPoint. The SAP Crystal Dashboard Design software relies on Excel to store and work with the underlying data. There are also numerous other analytic approaches common to market research, such as modelling, forecasting, and conjoint, which are also based on Excel.

Significance Testing

Transcription Global can help advise on the appropriate significance testing for a data set. More often than not it might be to test for significant differences between percentages, or means scores, for two independent samples using a t-test. Relevant to the small universe sizes that we often work within pharmaceutical research, we can advise on whether the finite population correction is likely to make a noticeable impact.

If other tests are more appropriate to the data, then we can advise on this, such as with the use of chi-square, paired-samples t-tests, one-way ANOVA and when to use nonparametric tests.

Overall, we recognise that significance testing as a technical concept not always clearly understood, which is why we would prefer to explain the relevance of significance testing in simple language.

Multivariate Statistics

We can help with a wide range of multivariate analysis which can be useful to market researchers. Specifically, following types of analysis can be run:

• Brand mapping with Correspondence Analysis
• CHAID
• Cluster Analysis
• Factor Analysis

We also work with a wide range of modelling approaches such as AMOS (Structural Equation Modelling), regression and mixed models approaches. We also work with the new ‘Generalized linear mixed models available within SPSS 20. These can manage a nested structure of ordinal target data, which is common feature of healthcare treatment (hospital, doctor and patient levels). Indeed, it was this new GLMM offering which encouraged us to purchase the Advanced Statistics module, where they also specifically mention pharmaceuticals and market research in the product description. An attractive and potentially useful geographical map feature has also been introduced with SPSS 20.

We can advise on where and why to use these techniques and to generate results relevant for business decision-making.

Dashboards

We are also offer dashboards, principally using the Excel-driven SAP Crystal Dashboard Design platform.

Relevant to research industry, dashboards can help business executives to easily visualise market research information. Dashboards can then be used as an alternative to situations where Excel alone is designed to show interactive information, such as with tracking market dynamics, forecasting and conjoint-type data.

The dashboard software we use creates ‘.swf’ files which can be viewed within standard business software platforms, typically in PowerPoint or in a PDF or in a browser.

A key element of our design philosophy is to develop dashboards for our clients which are ‘low-maintenance’. What we want to avoid is creating something which is time-consuming and expensive to up-date.