Data preparation Chapter 5 109 n n n n n Check missing values. Some missing values may be coded as "0". It's important that all missing values be represented as blank or null values. Otherwise, you will be factoring missing data values in the analysis. Take a look at the task times. Make sure the times are expressed in seconds and look reasonable. If there are some task times with 0 seconds, you should be suspicious, particularly if the task was performed successfully and verbatim comments were entered. Check clickstream data, particularly with respect to page views. Make sure there is a reasonable range of page views, particularly in the case where participants were successful in a task but had far fewer page views than the minimum necessary to complete the task. Also, actual pages visited should make sense from a task perspective. If there are a large percentage of par- ticipants who seem to have visited a Web page that does not make sense, there could be an issue with how that page was captured. Check the variables that are based on conditional logic. For example, some questions may only be given if the participants indicate they had difficulty on a task. If you see verbatim comments entered for tasks that were highly rated, there may be a problem. As a final check, we recommend you take a glance at automatically generated