PaceXL Version 2: Summary of options and features
- Use the jumps/links below to go to broad topics.
- Use "Find" to search for a specific concept,
for example, "histogram".
- Use "Back" to return.
- (The PaceXL Help system includes more details on each
option).
Key features of PaceXL
Charts and Graphs
Ungrouped Data
Tabulations and Histograms
Grouped Data
Probability Distributions
Intervals and Tests
Analysis of Variance
Regression and Correlation
Time Series Analysis
Index Numbers
Quality Control Charts
KEY FEATURES OF PACEXL
PaceXL
is a statistics and charts add-in for Microsoft Excel, designed for
introductory and intermediate courses.
PaceXL as an Add-in
- PaceXL operates inside Excel using its own added-in
menu and toolbars.
- The full capabilities of Excel are retained.
- Excel is used for the input of data, and for the output
for calculations, graphs and reports.
- PaceXL. works 'inside' Excel to streamline statistical
analysis.
- It provides an extended range of statistical and
graphical options.
- Several Excel workbooks using PaceXL can be open at
once.
- Several distinct data sets can be stored on the one
worksheet.
- In theory, over 200 variables and over 15,000
observations are permitted in a data set.
Menu, toolbar and dialogs
- PaceXL has its own menu item in the main Excel menu.
- Special PaceXL toolbars are used for general options
and for charts.
- PaceXL options are activated by selecting from the
PaceXL menu or toolbars.
- Each selection opens its own dialog window, from which
calculations and charts can be generated.
Data Area and data types
- PaceXL works by selecting a data matrix (called a Data
Area) before calling up any option.
- Column headers/labels in the Data Area are used as
variable names for easy selection.
- The user can switch between different options using the
same variables and data set.
- Variables in the Data Area itself can be numerical
(age, height, weight, etc) or categorical (gender, country of residence,
etc).
- PaceXL identifies each of the two types.
- PaceXL will attempt to allow for missing values
automatically: a blank cell is taken to be a missing value.
Analysing subsets and transformations
- As columns are identified by name, variables are easily
selected / deselected.
- The 'Unstack' option breaks one variable into groups
according to another index, or grouping, variable. (For example, weights
of adults can be grouped into 'weights for males' and the 'weights for
females'.)
- The 'Select' option enables certain rows from the Data
Area to be used in an analysis or for certain rows to be excluded. (For
example, if we wished to perform a regression analysis on just 'males'; or
to exclude an outlier observation set.)
- Common mathematical transformations (square, square
root, logarithms, reciprocal, etc) are available.
Statistics options and results
- PaceXL has specific statistics routines (see Contents).
- All calculations from these routines are written to a
single worksheet called the Results Sheet.
- New calculations are placed under each other in the
Results Sheet.
- The Results Sheet can be renamed/saved. A new one is
automatically created.
Charts and graphs
- Hundreds of different types of charts and graphs can be
generated using PaceXL.
- The Charts and Graphs dialog is a separate dialog which
generates a wide range of charts not available directly in Excel (box
plots, standard histogram, etc) and combination charts such as a box plot
on top of a histogram, or a normal distribution over a dot plot.
- Each of the statistics routines also generate charts
and graphs specific to those routines, for example, residuals for
regression.
- Charts can be displayed in "large format"
(suitable for lecture demonstrations).
- The user still has access to all the Excel chart/graph
options, and the Chart Wizard.
- Many standard Excel charts (column, bar, etc) can also
be generated through PaceXL.
Saving and opening files
- PaceXL workbooks are normal Excel workbooks, and saved
as disk files, and reopened, in the normal manner.
- A Data Area still remains set, and the workbook can be
used immediately with PaceXL.
- The Results Sheet and associated results are also
retained.
Printing and print preview
- PaceXL output is all captured in standard Excel
worksheets.
- Use normal Excel functions for printing and print
preview.
Help system
- PaceXL has an extensive help system, including
step-through instructions.
- Help is context sensitive, meaning that if F1 is
pressed while in a particular routine, Help in that routine is displayed.
- See the Home page and Download page for information on
using the Help system in recent versions of Excel or Windows.
Sample data sets
- Several sample data sets are available with PaceXL.
- Some are used for demonstrating PaceXL concepts via the
tours-tutorials.
- Others may be useful to instructors for assignment and
case study work.
PaceXL - CHART AND GRAPH – TYPES
The
Charts and Graphs dialog provides the options detailed below, grouped by page
tab from the dialog window.
One Variable
Notes:
- The charts from this tab use just one variable (column)
from the Data Area
- Charts are based around frequency counts of occurrences
of individual values (or categories)
For
numerical variables:
- Column chart
- Bar chart
- Pie chart
- Dot plot
- Polygon
- Frequency curve
- Frequency column
- Ogive
- Box plot
- Pareto chart
- Normal probability plot
- Quick histogram
For
categorical variables:
- Column chart
- Bar chart
- Pie chart
- Pareto chart
One or More 'Y'
Notes:
- These charts require at least two variables (columns)
from the Data Area
- Only one variable may be drawn on the horizontal (X)
axis
- The variable on the X axis may be either numerical or
categorical
- More than one may be drawn on the vertical (Y) axis
- The variable(s) on the Y axis must be numerical
- Many of the charts on this tab are matched directly by
charts available from Excel's Chart Wizard
Cross-tab
(contingency table) or grouped (frequency distribution) input format
recommended:
- Column chart
- Bar chart
- Pie chart
- Line chart
- Area chart
Numerical
variables (raw / ungrouped) required or recommended:
- Scatter
- Multiple box plot
- Confidence intervals
Two
or More 'X'
Notes:
One
numerical selected for the Y axis, one numerical variable selected for the X
axis, one categorical / coded variable used as a grouping variable. In effect,
a different scatter plot is drawn for each category / value in the grouping
variable.
Notes:
Multiple
pairs of Y and X variables are selected. Only numerical variables permitted.
Scatter 'Matrix'
Notes:
Able
to draw a scatter plot of each pair of variables included in the selection
list.
'Line'
of best fit options:
- Linear
- Parabolic / quadratic
- Exponential
- Cubic
- Power
- No fit
PaceXL - UNGROUPED DATA
Ungrouped Data - Broad functions
The
Ungrouped Data routine is used for analysing 'raw' data, in particular, for
calculating summary measures such as the mean and standard deviation, but also
for ranking, checking for outliers, etc. for numerical variables, and for
frequency counts for categorical variables.
If
you wish to group your raw data into frequency distributions or cross-tabs, use
Tabulations and Histograms. If your data are already grouped, use Grouped Data.
The
Ungrouped Data routine provides the options detailed below, grouped by page tab
from the dialog window.
Numerical Data
- Sample data - Summary measures (mean, standard
deviation, etc.)
- Population data - Summary measures (mean, standard
deviation, etc.)
- Percentile - for the percentage entered
- Trimmed Mean - for the percentage entered
Choose Measures
- Add to or reduce the range of summary measures
calculated
Rank/Sort/Count
- Rank and Percent Rank
- Sort, Rank and Percent Rank
- Frequency Count
- Check Outliers
- Normal Distribution Ranks
- Test for Normal (goodness of fit)
Categorical Data
Plots
For
numerical variables:
- Column chart
- Bar chart
- Pie chart
- Dot plot
- Polygon
- Frequency curve
- Frequency column
- Ogive
- Box plot
- Pareto chart
- Normal probability plot
- Quick histogram
For
categorical variables:
- Column chart
- Bar chart
- Pie chart
- Pareto chart
PaceXL - TABULATIONS AND HISTOGRAMS
Tabulations and Histograms - Broad functions
The
Tabulations and Histogram routine proceeds from the Ungrouped Data routine. It
is used to group raw data into frequency distributions or two-way cross-tabs
(or cross-classification tables or contingency tables). It also plots
histograms and column graphs of those tables. If the data set is already in
grouped form, use the Grouped Data routine.
The
Tabulations and Histograms routine provides the options detailed below, grouped
by principal options from the dialog window.
Frequency Interval option
- Frequency distribution table for equal or unequal
classes
- Histogram for absolute, relative and cumulative
frequencies
This
is the default tab of the dialog. It operates on just one variable at a time.
Users can specify intervals for grouping values by entering the lower bounds of
each class. (Only the first two are required if Equal Classes are selected).
The histogram plot includes a wide range of combination plots (such as
histogram-dot plot).
Frequency Count option
- Frequency count table of individual values (no grouping
by intervals)
- Frequency count plots (includes column, dot, box plot,
quick histogram)
Operates
on one variable at a time. Counts how often each value occurs. This option is
suitable for discrete numerical data or categorical data and is equivalent to
the Frequency Count and Categorical options in 'Ungrouped Data'.
Multiple Plots option
- Frequency curve, polygon or histogram for two or more
numerical variables grouped into classes
Useful
for comparing two or more distributions. Because histograms plot on top of each
other, obscuring underlying plots, the default is the frequency curve. Classes
are set in a similar way to Frequency Interval.
Two Variable Cross-Tab option
- Cross-tabs for total frequencies, percent of total,
percent of rows and percent of columns
- Column graphs of the above tables
Operates
on two variables at a time and in interval form. Numerical or categorical
variables permitted. Either discrete or continuous numerical data can be used.
For example, we may want a cross-tabulation of 'Gender' by 'Hours Worked' in
10-hour groups. The lower bounds of each interval for each variable are entered
as for Frequency Interval. A 'Chi-Square Test of Independence' is given as an
option: note that it uses an alpha level of 5%. The p-value from the test is
also shown, which can be used to compare to other alpha levels if required.
Cross-tabulations can be graphed. For example, Percent of Row and Percent of
Column tables can be plotted as column graphs, thus providing an alternative to
the tables for identifying whether relationships exist between two variables.
PaceXL - GROUPED DATA
Grouped Data - Broad functions
The
Grouped Data routine is used when data are already in grouped form, either as a
frequency distribution or as a cross-tab. (A cross-tab is also known as either
a two-way cross-tabulation or a contingency table).
With
Frequency Distributions, the Data Area may include more than one frequency
column (for example, frequencies for males and for females and for the two
combined).
(Note
that the Tabulations and Histograms dialog is used for grouping ungrouped/raw
data into frequency distributions and cross-tabs. Once the data are grouped,
the output from 'Tabulations and Histograms' is similar to that obtained from
'Grouped Data'.)
The
Grouped Data routine provides the options detailed below, grouped by page tab
from the dialog window.
Frequency Distribution
With
frequency distributions, the Data Area must include two data columns, including
one column for the midpoints of the classes for each interval. (The midpoints
should not be included as the Row Headers.) Output available includes:
- Sample data - Summary measures (mean, standard
deviation, etc.)
- Population data - Summary measures (mean, standard
deviation, etc.)
- Percentile - For the percentage entered
- Table/Plot: Single column - this generates a frequency
distribution or histogram for the selected column of frequencies
- Goodness of fit: Normal
- Multiple Plots - Frequency curve, polygon or histogram
for two or more frequencies plotted on the one graph
Choose Measures
- Add to or reduce the range of summary measures
calculated
Two Variable Cross-Tab
- Cross-tabs for total frequencies, percent of total,
percent of rows and percent of columns
- Column graphs of the above tables
- Test of independence
PaceXL - PROBABILITY DISTRIBUTIONS
Probability Distributions - Broad functions
The
two broad functions of the Probability Distributions routine are to:
- calculate probabilities for a given probability
distribution
- calculate values of the 'variable' corresponding to
given probabilities for a particular distribution
A
Data Area is not required for any calculations.
Probability distributions available
The
eight distributions available are:
- Uniform distribution
- Binomial distribution
- Poisson distribution
- Normal distribution
- Standard normal distribution
- t distribution
- Chi-Square distribution
- F distribution
Uses of this routine
The
probability distribution routines can be used in a number of ways but in
particular:
- to replace or extend probability tables of the type
given in the back of textbooks
- to help solve standard textbook-type problems involving
probability distributions
Performing calculations
In
each case you need to:
- enter the parameters for the distribution, and
- either a probability value or the values of the
'variable'
Calculations
are displayed in the dialog itself. Updating is immediate, once a value of a
parameter or input value is updated.
Use
the Update button to update values once an input has been altered.
You
can repeat calculations for the same distribution or change distributions.
However, after changing distributions, defaults are reset.
Saving results
- Current calculations will be written to the Results
Sheet when the Save Results button is clicked.
- Plots of distributions can be saved using the normal
Save Plot options
Plots of distributions
Plots
are available for each distribution as calculations are performed. Plots can be
in:
- normal format on the Results Sheet, or
- large format as a Chart Sheet.
Plots
are updated immediately with any change in input or with the scroll bars. That
is, the plots are linked to the dialog window.
The
Dialog box can be reduced in size using the 'Less Dialog' button in order to
view more of a particular plot.
Scroll-bars
The
scroll-bars can be used to examine incremental values in key parameters and
input values. For example, the standard deviation for the normal distribution
can be altered gradually by clicking the scroll-bar. If the plot is shown, the
height/width of the distribution will gradually alter.
Note
that the scroll bars may jump to a default value before returning to your
required value. This can apply to calculations and to plots.
Two X values
Two
values for X may be entered for the normal, uniform, binomial and Poisson
distributions (and two z values for the standard normal distribution). The
probability between these two values is calculated and can be shown on a plot.
What-if possibilities
Using
the scroll-bars, and other display options, 'what-if' situations can be
investigated. For example, gradually changing a parameter and viewing the
effects on a result or plot. Or plotting a normal distribution over a t
distribution and altering the degrees of freedom, etc.
PaceXL - INTERVALS AND TESTS
Intervals and Tests - Broad functions
The
Intervals and Tests routine is used for inferential statistics for one or two
sample situations. It is one of the most powerful of the PaceXL statistics
options. Input may come from a Data Area or from Summary Measures. Numerical
variables and categorical variables are permitted.
The
Intervals and Tests option provides the options detailed below. (Functions are
grouped by operation, type of calculation, etc.)
Variable type
Calculations
may be performed on:
- Numerical variables
- Categorical variables
One or two samples
Calculations
can involve:
- 'One Sample' data: just one variable or
category/attribute involved
- 'Two Sample' data: two variables or
categories/attributes involved
Calculation types
There
are three main types of calculations:
- Confidence Interval
- Hypothesis Test
- Sample Size
Numerical variables
In
this tab, calculations are performed for numerical variables with the following
options:
- One Sample, Mean: using the t distribution
- One Sample, Mean: using the z distribution
- One Sample, Variance/SD: using the Chi-square
distribution
- One Sample, Median: Sign test
- Two Sample, difference between Means, Equal Variances:
using the t distribution
- Two Sample, difference between Means, Unequal
Variances: using the t distribution
- Two Sample, difference between Means, Dependent Samples:
using the t distribution
- Two Sample, difference between Variances/SDs: using the
F distribution
- Two Sample, difference between Medians: Wilcoxon Rank
Sum, using the z distribution
Note
that PaceXL uses the t distribution as the default distribution, even for large
sample sizes (30 or more).
'Categorical' variables
In
this tab, calculations are performed for categorical variables (including
numerical coded categorical variables) with the following options:
- One Sample, Proportion: using the z distribution
- Two Sample, difference between Proportions: using the z
distribution
With
categorical variables, an attribute value may need to be selected, eg 'Male' or
'Female', or '0' or '1'.
Parametric v nonparametric methods
Statistical
inference techniques are sometimes categorized as:
- parametric methods
- nonparametric methods
The
Intervals and Tests routines includes both categories of techniques.
Finite population correction
If
a population is not large:
- the Finite Population Size correction factor can be
employed for one sample tests
Data source
The
input source for the sample data can be:
- via the Data Area (the variables in the Data Area are
listed for selection and PaceXL automatically calculates sample summary
measures for each variable selected)
- via Summary Measures (if the sample summary measures
are already known the user enters the values directly into the dialog)
Extra input required
Extra
input may be required in the form of selecting or entering the:
- Confidence coefficient
- Level of significance (alpha)
- Value of the population parameter for the Null
Hypothesis
- Required Margin of Error
- Estimates of population parameters
- Sample size, mean, standard deviation, etc.
Results
Calculations
are written to the Results Sheet and include:
- Full details of the type of calculation performed
- Details of the data input
- A summary of the intermediate calculations and key
results
- A final result or suggested conclusion
Unstack/Select
Note
that Unstacking and Selecting from a Data Area are options for this routine.
These options provide a powerful means of analysing subsets of your data, for
example, testing for any difference between Male and Female respondents
Calculating the sample size
Two
choices for calculating the required sample size, n are:
- numerical variable and estimating the population mean
- categorical variable and estimating the population
proportion
Plots
The
following plots are available:
- hypothesis tests (for selected numerical variable tests
or categorical variable tests)
- confidence intervals for one or more numerical
variables
- box plots for one or more numerical variables
- single variable plots for numerical variables (box
plots, dot plots, quick histogram) and categorical variables
PaceXL - ANALYSIS OF VARIANCE
Analysis of Variance - Broad functions
The
Analysis of Variance routine is used for one factor and two factor analysis of
variance.
The
options available with the Analysis of Variance routine are given below.
One Factor
- Standard one factor ANOVA output
- Summary measures for input data
Options
include:
- Confidence interval calculations
- Multiple comparisons
- Repeated measures
Two Factor
Data
are assumed to be in Stacked format and an Index variable must be included in
the Data Area.
- Standard two factor ANOVA output with interaction
- Summary measures for input data
Plots
- Box plots for all samples, including values
- Confidence intervals using pooled standard deviation
- Confidence intervals using individual sample standard
deviations
PaceXL - REGRESSION AND CORRELATION
Regression and Correlation - Broad functions
The
Regression and Correlations routine is used for exploring for relationships
between variables using regression modelling, correlation and scatter plots.
Models
are generated by simply selecting and deselecting variables as required from
the dialog window.
Examples
of the types of output include:
- regression model with standard output
- optional full table of residuals
- correlation matrices
- 'scatter matrix' plot option
- prediction intervals
- residual plots, including histogram, normal curve and
box plot options
- automated regression options including stepwise and
best subsets
The
Regression and Correlation routine provides the options detailed below, grouped
by page tab from the dialog window.
Correlation
Produces
the following matrices for the pairs of variables chosen:
- pairwise linear correlation coefficients
- t statistics for zero correlation
- probability values (p-values) for zero correlation
- r-squared (optional)
Matrices
can be printed as three (or four) separate tables, or printed as a table with
three (four) rows per combination.
Scatter
- Draws scatter plots for every pair of variables chosen
in the list
- Fit options include linear, quadratic, exponential,
cubic and no fit
Regression
This
is the default, and principal, tab of the Regression and Correlation dialog. It
is used to generate all regression models, including automated methods. Once a
mode has been generated, predictions and plots can be made for that model.
Input
includes:
- Y (dependent variable)
- X (independent or explanatory) variable(s)
- confidence % level required
- which residuals to be analysed for normality
Standard
Output includes the following:
- regression equation and model
- regression coefficients, including p-values and
confidence limits
- variance inflation factor (VIF) for each explanatory
variable
- summary measures, such as R-squared, adjusted
R-squared, standard error of estimate, Durbin-Watson
- ANOVA table
Optional
output including:
- fitted results
- residuals, including standardized residuals,
studentized (TResiduals), leverage, Cook's distance
- flagging of observations for influence analysis
Optional
table of:
- intermediate calculations, including sums of
cross-products
Optional
calculation:
- with intercept
- no intercept
Optional
input:
confidence
interval %
Predictions
Provides
intervals for:
- mean of Y (confidence interval)
- individual Y (prediction interval)
Predictions
relate to the latest regression model generated.
Regression Plots
Produces
the following plots:
- Fitted
- Residuals
- Adjusted residuals
- standardized residuals
- Residuals v Normal
- Prediction Bands (for simple regression - one
explanatory variable)
Automated Methods
This
tab is used to select, and set up, 'automated' regression methods. The default
is multiple regression. Automated options are:
- stepwise regression
- forward selection
- backward elimination
- best subsets (with Cp statistics)
The
output level can be set to:
The
variables to be included in the model are selected from the Regression tab.
Select and New Variable
- Select can be used to choose / exclude rows from the
Data Area
- Unstack cannot be used
- Temporary variables can be created (squared, logs, etc)
for inclusion in regression models
PaceXL - TIME SERIES ANALYSIS
Time Series - Broad functions
The
Time Series Analysis routine is used for analysing time series data. A range of
calculation and plotting methods are available for exploring past behaviour,
and for forecasting. Annual and sub-annual (monthly, quarterly, etc) data can
be analysed.
Data
period:
- PaceXL attempts to keep track of time periods and
sub-periods
- Except for 'Seasonal Analysis' the data can be annual
or sub-annual, or some other period (such as, five-yearly)
Number
of variables:
- Except for Plot Series, only one variable is analysed
at a time
Accuracy
of fit measures given:
- R-squared
- MSE (Mean Square Error)
- MAD (Mean Absolute Deviation)
- RMSE (Root Mean Square Error)
- MAPE (Mean Absolute Percentage Error)
Forecasts
available for:
- trend fitting
- exponential Smoothing
- autoregression
Plots
available for all tabs:
- including fitted results and residuals
- semi-log options
- trend fits
The
Time Series Analysis routine provides the options detailed below, grouped by
page tab on the dialog window.
Set Period
Data
period choices:
- year
- half year, quarter, month, week or day
- other (for example, hourly data for a day, or daily
data for a five-day week, or five-yearly)
Data
period input:
- year/period of first observation
- number of first year or sub-period
Plot Series
Options
for plotting past data for one or more variables:
- arithmetic scale
- semi-log scale
- percentage change
- index form
- trend
- trend: semi-log
Differences
Calculation
and/or plot options:
- first differences
- second differences
- percentage differences
- index form
Moving Average
Calculation
and/or plot options:
- moving average
- percent of moving average
- absolute and percent residuals
- moving average differences
- trend fits
Moving
averages are centered for an even number of periods per cycle.
Seasonal Analysis
Calculation
and/or plot options:
- seasonal indexes
- seasonally adjusted (deseasonalized) results
- absolute and percent residuals
- trend fits
Conditions
and notes:
- with Seasonal Analysis, annual data are not permitted
- seasonal indexes are calculated using the modified mean
method
- there must be 2 or more sub-periods per main period
(year/cycle)
- there must be at least four complete years/cycles of
data because
- cycles may be lost due to use of a centered moving
average
- two cycles are lost due to the modified mean method
- thus if the data being analysed are quarterly data,
there must be 16 or more data points
- note that 'Actual Y as a Percentage of Moving Average'
calculations, which are used to calculate seasonal indexes, can be
obtained from the Moving Average option
Trend Fitting
Trend
types:
- linear
- exponential
- quadratic
Calculation
and/or plot options:
- trend equation only
- full calculations
- seasonal adjustment option for sub-annual data
- forecasts only
- all trends summary (for the trend types given above)
- absolute and percent residuals
- seasonalized trends (sub-annual data)
Exponential Smoothing
Calculation
and/or plot options:
- single exponential smoothing
- double exponential smoothing
- triple exponential smoothing models
- seasonal adjustment option for sub-annual data
- different alphas summary (0.1, 0.2 to 0.9)
- absolute and percent residuals
- seasonalized forecasts ('sub-annual' data)
Conditions
and notes:
- one constant (alpha) for single and triple smoothing
- two constants (alpha and gamma) for double smoothing
- the triple exponential model is not a specific model,
but is included for demonstration/example purposes
Autoregression
Calculation
and/or plot options:
- lags of 1 to 12 permitted
- one predictor lag included for lag value
- all predictor lags included
- autocorrelation coefficients
- fitted and residuals plots
PaceXL - INDEX NUMBERS
Index Numbers - Broad functions
The
Index Numbers routine has two broad functions.
The
first is the calculation of index numbers. This option includes calculating
unweighted and weighted index numbers (price or quantity) based on inputs of
prices and quantities.
The
second is the use of price index numbers. This option allows for deflating a
value series using a price index series, calculating percentage changes in an
index series (for example, calculating the percentage rate of change in the
Consumer Price Index to provide an estimate of the period inflation rate), and
deriving other index series.
The
Index Numbers routine provides the options detailed below, grouped by page tab
from the dialog window.
Calculating Indexes
Index
types:
- price indexes
- quantity indexes
Weighting:
- unweighted
- weighted (based on selection of a quantities column)
Output
options:
- indexes only
- full calculations
Notes:
To
calculate price (quantity) indexes, select the column for base period prices
and the given (latest) period prices, and the quantities period.
Unweighted
indexes
- unweighted aggregate of prices index
- unweighted average of price relatives index
Weighted
indexes
- weighted aggregate price index
- weighted average of price relatives index
PaceXL
cannot detect the period to which the weights belong. However, by selecting the
appropriate weights in the Data Area, a Laspeyres Index (base year weights) or
a Paasche Index (latest year weights) can be calculated.
Using indexes
One
value series and one price series:
- deflated series using a value and price series
- percentage changes in the value and price series
- various index forms of both series
- plots of these calculations
Two
or more value series and one price series:
- deflated results
- percentage changes
- index forms of results
- plots of these calculations
PaceXL - QUALITY CONTROL CHARTS
Quality Control Charts - Broad functions
The
Quality Control Charts routine is mainly a plotting routine, but calculations
underlying these plots can also be generated. Control charts can be drawn for
numerical variables and categorical variables (or attributes).
The
Quality Control Charts routine provides the options detailed below, grouped by
page tab from the dialog window.
Numerical Variables
Chart
types and calculations:
- X-Bar chart
- s chart
- R chart
Options
with one/some of the above:
- Pooled SD (standard deviation)
- SD-Bar
- R-Bar
- Population SD (standard deviation)
- Population mean
- Tests for non-conformity (X-Bar chart only)
- Exclusions permitted
Test
available for X-Bar chart for non-conformance:
- Test 1: One or more means more than 3 sigmas from CL
(center line)
- Test 2: Nine means in a row on same side of CL
- Test 3: Six means in a row, all increasing or all
decreasing
- Test 4: Fourteen means in a row, alternating up and
down
- Test 5: Two out of three means more than 2 sigmas from
CL on same side
- Test 6: Four out of five means more than 1 sigma from
CL on same side
- Test 7: Fifteen means in a row within 1 sigma either
side of CL
- Test 8: Eight means in a row more than 1 sigma either
side of CL
Attributes (categorical variables)
Chart
types and calculations:
- p chart
- np chart
- C chart
- U chart
Options
with one/some of the above:
- Population proportion, p
- Population mean
- Tests for non-conformity (p chart only)
- Exclusions permitted
Test
available for p chart for non-conformance:
- Test 1: One or more proportions more than 3 sigmas from
CL
- Test 2: Nine proportions in a row on same side of CL
- Test 3: Six proportions in a row, all increasing or all
decreasing
- Test 4: Fourteen proportions in a row, alternating up
and down
Factors
- Prints table of factors for control charts
Exclusions
- Samples can be excluded from the analysis from the
Numerical Variables or Attributes tab