All too often, research projects flop because the data creation steps fail to end in a useful dataset. Even if you are a non-statistician commissioning or reading statistics done by others, you should interrogate the integrity of the data if you can.

Basic statistics, such as the descriptive measures described in Chapter 7, give you various relevant things that help with a pre-assessment of each variable for data checking purposes. For each variable, a descriptive analysis can give you information such as the number of observations with data (N), amount of missing data, mean (average), median, standard deviation, interquartile range (data points at 25th and 75th percentiles of answers), and also minimum and maximum.

These descriptive statistics can help with data checking as discussed in the next sections. In addition, basic association statistics such as correlations also have a role to play in checking and finalizing many datasets.