If distributions are not normal; i.e., non-parametric, such as those that are flat, peaked, or strongly skewed, non-parametric statistics are recommended. These statistics are particularly relevant in the IT realm where data frequently does not fit into a normal distribution.

T-tests are typically used for determining whether or not one group significantly differs from another on some type of metric. For instance, we may discover that females, on average, spend significantly more hours than males at health-related Internet sites. When conducting numerous t-tests the probability of reporting that a result is significant when it actually is not, dramatically increases. In such cases, one should use Anova to help control for chance findings.

One of the most useful z-test applications in market research is determining whether or not one proportion significantly differs from another. For example, we may discover that the proportion of Internet users in one geographic region exceeds that of another.

Correlation measures the degree of relationship between one variable and another. There may be, for instance, a high correlation between those who have two or more phone lines in their household and time spent on the Internet. One must note, however, that correlation is a measure of linear (i.e., straight line) relationships. If the two variables of interest have a non-linear relationship such as an inverted U, the correlation coefficient (r) will fail to detect a relationship when one is actually present.

This procedure is useful for detecting mean differences among three or more groups. ANOVA is a viable alternative to conducting numerous t-tests because the analysis controls for chance findings (Type I error). To assess differences in the average number of hours spent on the Internet among PC owners in four countries, ANOVA would be an appropriate tool. Similar to other statistical techniques, ANOVA is not immune from Type I error when used repeatedly with a data set. To address this problem, MANOVA (explained below) should be employed.

This analysis can detect mean differences among a number of different groups on several different measures while protecting for chance findings. The method is an efficient and powerful analysis for large research studies in which there are a variety of segments being assessed on a number of different measures.

This procedure is useful for detecting mean differences among three or more groups while holding one variable constant. To assess differences in the average number of hours spent on the Internet among PC owners in four countries, while controlling for access speed, ANCOVA would be an appropriate analytic tool. ANCOVA is particularly useful in research situations where a variable, such as income, gender, education, or age, can potentially obscure or bias the results.

This analysis can detect mean differences among a number of different groups on several different measures, while holding one or more variables constant. The method is useful for research studies where there are a variety of segments being assessed on a number of different measures, where one or more variables needs to be controlled for that may potentially bias the results.

In ANOVA, ANCOVA, MANOVA, MANCOVA, the respondent is assessed once for each measure. In repeated measures ANOVA-based designs, the respondent is measured several times. For instance, data collected through measuring the number of online purchase transactions made by different buying segments per quarter would be appropriate for this analysis. Accuracy is increased when measuring a respondent on several occasions as opposed to one, thus making a repeated measures approach one of the more powerful analytic techniques.

These techniques identify buyer preferences for product features, the most desired set of features for a product, and what tradeoffs buyers are willing to make for their desired product. The techniques are thus effective tools for developing a successful product design and bundling of product or service offerings.

These are all tree-based tools that segment groups of respondents that share similar characteristics. CHAID (Chi-squared Automatic Interaction Detector) and Exhaustive CHAID are ideal for visualizing large data sets for consumer profiles and segments. C&RT (Classification and Regression Tree) and QUEST (Quick Unbiased Efficient Statistical Tree) provide similar results but, unlike CHAID techniques, produce trees with binary splits which are more appropriate for some types of research. All four techniques are effective variable reduction tools and precursors to other types of analyses, such as regression and higher-order predictive models.

Discriminant analysis is useful for finding a group of variables (i.e., a discriminant function) that distinguishes one group from another. Although it works well for group membership situations, it is not as robust to statistical violations as, for example, logistic regression that will provide similar information.

In the realm of market research, these variable reduction schemes identify underlying dimensions of what respondents may be thinking when, for example, evaluating a product or service. Please note that these analyses do not test whether the dimensions that surface relate to a specified outcome (e.g., an online purchase). Regression or higher-order predictive models, such as RPM, are required to assess whether the dimensions have any predictive value.

This technique is particularly effective for exploring branding issues. Several brands can be compared and contrasted, on a number of different attributes, in one comprehensive picture. A perceptual map may indicate that several brands of laptops are perceived similarly in terms of price, performance, and wireless capabilities, but not in terms of reliability and warranty coverage. Another advantage of perceptual maps is that the data required to construct them is straightforward and typically not difficult to collect - consumers usually rate the product/service attributes on simple Likert-type scales (e.g., ranging from strongly agree to strongly disagree).

Through Cluster analysis, you receive exact information about your target audience. Their objective is to identify transparent representation of the information contained within the data. From this, it is possible to develop target group typologies. The process is based on similarity categories of the survey elements within homogenous groups. The objective of cluster analysis is to identify groups that resemble each other. Generally, several variables that appear to be similar are selected to ascertain "Similarity" and establish the measure of similarity of the respective characteristics. As there is not only one process of cluster analysis, a process for the individual survey objective is constructed.