English      Chinese      French      German      Indonesian      Spanish
Diagnostic test evaluation and comparison

Analyse test repeatability

Paste or enter data for critical limits (if required) below*.

Download example data

Paste column(s) of data to be summarised in the space below.
Include a header row of series IDs, a first column of run IDs and a column of data for each sample or series to be analysed.

Download example data

*One column is required for each data series. Include a header row and a column for row-names as shown in the example spreadsheet. Enter one of the following, depending on the option selected above:

  • If you chose to calculate limits from the data: leave blank;
  • If you chose to calculate limits from specified mean and sd: enter 2 rows of data, with first row the specified mean and second row the specified sd; and
  • If you chose to specify expected values and limits: enter 5 rows of data, in the following order, upper critical limit, upper warning level, expected value, lower warning limit and lower critical level.

This utility evaluates the repeatability of a test, based on the results of repeated testing on a series of 1 or more samples for a test producing a continuous outcome.

Data required is a series of test results for a panel of one or more samples. The first row of the data is expected to contain column headers identifying each of the sample series. The first column is assumed to contain identifiers for each test occasion (for example plate numbers, dates or just a sequential number). Testing data is assumed to be temporally ordered and matched across samples. Additional columns (after the first) are assumed to contain test results for the samples to be evaluated.

Outputs include (for each sample):

  • Numerical and graphical summaries of testing results;
  • Calculation of warning and critical limits and of numbers and percentages of occasions when these limits are exceeded;
  • Time series analysis, including plots of raw data, moving average and lowess trend line;
  • Summary of regression analysis for univariate linear trend; and
  • Autocorrelation plot.

The width of the window for calculating moving averages can be set by the user. Warning and critical limits can be either:

  • Calculated from the data (mean +/- 2 and 3 standard deviations);
  • Calculated from mean and standard deviation values specified by the user (mean +/- 2 and 3 standard deviations); or
  • Specified explicitly by the user.