Preferences Window

Use Preferences to change the app language and customize reports, adjust some common analysis preferences, and more. Please note that not all options are available on all devices.

General

Language

When you have more than one language installed, this option allows to choose the app language. By default the language will be the same as the operating system language on the device if the app is available in this language. If your preferred language isn't yet supported - use the "Send Feedback" button to let us know.

Output Options

These options allow you to customize report formatting such as font, size, and numeric format.

Default font, Size

Use these options to select the font and size you want to use for reports.

Numeric format / Display format

This group of options tells the program how to display numbers in reports.

    Select a cell with a number and the formula bar will display the actual number value with no formatting applied.

    Decimal places (Precision)

    Use this option to specify the number of digits after the decimal point. This option controls how numbers look in reports, it does not affect calculations. The default value is 5, the minimum value is 2 and the maximum value is 15.

    Hide trailing zeros

    Use this option to hide trailing zeros (sequence of 0) after a decimal point. When the option is selected, trailing zeros are removed from the resulting string after rounding to the desired precision (decimal places). Leave unchecked for fixed-point notation. The default value is "OFF" (option is not active).

    For example, a cell contains the number 100.500.
    The option is turned OFF (fixed-point notation): the number will be displayed as 100.5000 when the precision (number of decimal places) is set to 4, and as 100.50 when the precision is set to 2.
    The option is turned ON (hide trailing zeros): the number will be displayed as 100.5 when the precision is greater than or equal to 1.

    Use scientific notation (1.23E4)

    Use this option to display all numbers in E-notation (also referred to as scientific notation).
    Numbers in E-notation are written in the form mEn: m multiplied by 10 to the n-th power. m, the number preceding to 'E', is a real number between 1 and 10 (1 ≤ m < 10) and n, the exponent, is an integer number.
    Scientific notation is useful to display numbers that are too big or too small to be conveniently written in a decimal format. A number that is too large or too small to display in the cell will automatically be displayed in scientific notation. The default value is "OFF" (option is not active).

    For example, when the scientific notation option is active and the precision (decimal places) is set to 3, the number the 8675309 is displayed as 8.675E+06, which is 8.675 times 10 to the 6th power.

    Percent-style cells for probabilities

    Use this option to apply percentage formatting to numeric values that represent a probability (like p-values) or a percentage. The default value is "OFF" (option is not active).

    For example, a p-value of 0.0123 is displayed as 1.23%, when the option is active and 0.0123 otherwise.

Add a timestamp to reports

Use this option to include in a report the following information: the timestamp that indicates the date and time when a report was generated, the name of a data source (workbook) and the ranges (columns) used as input.

Statistics

Alpha (significance level)

This option defines the alpha level (significance level or the level of significance) for tests. An alpha level is the probability of incorrectly rejecting the true null hypothesis (type I error). The default value is 0.05 (5%).
For confidence intervals the alpha value is transformed into the confidence level as 100×(1 - alpha)%. For example, an alpha of 0.05 indicates a 95 percent confidence level. As alpha increases the confidence level will decrease.

Missing values removal

Use this option to force the casewise deletion of cases (rows) with missing values instead of the pairwise deletion. This approach is known as a complete case analysis.
By default (the option is set to Automatic), missing values are deleted with one of the following deletion techniques.

  • The analysis by analysis deletion is used by univariate analysis methods and tests for independent samples, i.e. by commands that process each variable individually – descriptive statistics without group variable, normality tests, independent-samples t-tests, one-way ANOVA for unstacked data, nonparametric tests for independent samples, basic time series and data processing commands (differencing, mean removal, standardization), univariate charts (stem-and-leaf, box-plot, Q-Q plots, histogram).
    In this case, a case (row) with a missing value is excluded only from the analysis of the variable containing the missing value – and only that variable.
  • The pairwise deletion is used for correlation and covariance analyses unless the Force casewise deletion option is active.
    This approach excludes a pair of observations when either of the values is missing. For example, if a case (row) has data on VAR1, VAR3, and VAR4, but not on VAR2, that case would be included in computing r(VAR1,VAR3) and r(VAR1,VAR4), but not in computing r(VAR1,VAR2). The pairwise deletion allows to use more information from the sample but the correlation matrix elements are based on different sample sizes. This can result in biased estimates and inconsistent correlation matrix – the matrix may not be a positive semidefinite and may have negative eigenvalues (Kim & Curry, 1977), and thus, would not be suitable for multivariate analysis.
  • The casewise deletion (listwise deletion) is used for multivariate analysis commands and commands that use a group variable (or a by variable).
    A case (row) is excluded from analysis when it contains a missing value for at least one of the variables selected. In other words, the analysis is only run on cases (rows) with a complete set of data. For example, with Force casewise deletion option selected, the correlation command will calculate all correlations by excluding cases (rows) that have missing values for any of the selected variables and the analysis will be based on the same set of data (and therefore the resulting correlation matrix will be a true correlation matrix).

Random Seed

Like most modern software, the app uses the industry-standard Mersenne Twister pseudorandom number generator (MT19937 implementation). The Mersenne Twister generator is the most widely used general-purpose pseudorandom number generator, developed in 1997 by Makoto Matsumoto and Takuji Nishimura. The random number generator is started by specifying the initial state, commonly called the "seed" (also referred to as the base for the random number generator). Each time a command runs the initial seed could be either a specific value or a random number (set the value of 0 for the Random Seed). Specifying the seed allows to reproduce a sequence of random numbers and replicate an analysis that uses random numbers.


References
  • Kim, J.O. and Curry, J. (1977) The Treatment of Missing Data in Multivariate Analysis. Sociological Methods Research, 6, 215-240.