AML Documentation

<g>

- Start a graphic presentation or define <g> attributes

<xt>

- Start a cross table or define <xt> attributes

<dig>

- <dig> Create a Dig analysis

a

Defines the main graph address to be analysed and presented

Example: The adress expression a='\4' will show a graphic presentation of question 4.

autocode

Ascending or Descending will make an address vector of all occurencies of a numeric data type (F) in the a attribute and the values will appear ascending or descending in the graph. The H address component may be present in the a attribute in order to limit the value vector.

Example: Question 2 is age. The graph <g a='\2' autocode='descending'/> will show the frequency of all age groups in the survey in a descending order. AUTOCODE IN INTERVALS? (Link: autocode.htm)

autocodeinterval

Specify the autocode clustering

Example: If autocodeinterval='2' is specified, the autocode feature will show categories for every 2nd value

b

Defines an address that adds new dimension(s) to the graph. This address is combined with the address of the 'a' attribute and the appearance of the graph is determined by the organizing rule found in the a attribute address.

Example: If question 5 is household income, question 1 is gender and 4 is age groups, the graph the following expression <g a='\5' b='\1 \4'/> will show the household income split down to gender and age. If you want to include also the total, you can specify ALL, e.g. b='ALL \1'

basedoc

Documents the size of the chart.

The base documentation will show together with the chart title.

catdoc

Documents the size of the categories in categorical questions.

The category documentation will show together with the categary labels.

chartonnewslide

Show chart as chart one on a new PowerPoint slide instead as a chart number two on the same slide together with previous chart.

This attribute will work together with the attribute: pppSlideLayout='ppLayoutTwoObjects'.

colorindex

Specify the colors in the chart.

The colors will override the existing template in e.g. a PowerPoint presentation.

deciavg

The number of decimals to be used for average and fractile values

decipct

The number of decimals to be used for percentage values

f

The logical expression that requests the filter population

The filter will be a subset of the 'universe' attribute in the <xtgroup> element.

filtertext

A template text expression to generate the text associated with the logical expression

If no filtertext is given there are two options: 1) When the filter (the 'f' attribute) is a BASIC logical expression (not combined from more questions) the standard template content will be that the generated text will be the text describing the right hand side of the logical expression (answer codes). 2) When the filter is a more COMPLEX logical expression, the logical expression will be shown as a filtertext. Known issue: The filtertext ONLY works on question level, NOT on a global level (xtgroup)

max

Maximum shown in the value axis

min

Minimum shown in the value axis

pptslidelayout

Specifies whether you want a standard PowerPoint layout (Title and content) or two charts on one slide (Two content)

qno

Should the question box number (name) appear as a suffix to the question text?

rolldir

The direction of the rolling to be performed, if any.

Example: Question 9.A is a brand awareness question, where the last code is 'Don't know'. The expression <g a='\9.A' b='ALL' rollexclude='1' rollend='5' rolldir='desc' /> will show the TOP (rolling='yes' rolldir='desc') 5 BRANDS (rollend='5') at the TOTAL LEVEL and the 'don't know any brand' (rollexclude='1') will also be shown at the bottom of the graph. Known issue: The filtertext ONLY works on question level, NOT on a global level (xtgroup)

rollexclude

Should a number bottom bars be excluded from the rolling prosess?

Example: Question 9.A is a brand awareness question, where the last code is 'Don't know'. The expression <g a='\9.A' b='ALL' rollexclude='1' rollend='5' rolldir='desc' /> will show the TOP (rolling='yes' rolldir='desc') 5 BRANDS (rollend='5') at the TOTAL LEVEL and the 'don't know any brand' (rollexclude='1') will also be shown at the bottom of the graph.

rollend

The top N bars of each graph block will be shown.

Example: Question 9.A is a brand awareness question, where the last code is 'Don't know'. The expression <g a='\9.A' b='ALL' rollexclude='1' rollend='5' rolldir='desc' /> will show the TOP (rolling='yes' rolldir='desc') 5 BRANDS (rollend='5') at the TOTAL LEVEL and the 'don't know any brand' (rollexclude='1') will also be shown at the bottom of the graph.

rolling

Should the values be sorted in the graph

sat

Are <sat> definitions to be applied for this graph?

seriesdoc

Documents the size of the series for categorical questions.

The series documentation will show together with the chart legend.

seriesin

Swap row/columns in the analysis: Specify which series (a='\xx' or b='\xx') that should be analyzed at the first level.

show

The sequence of show attributes in the bar blocks.

Note: Average will be shown for all quantitative values

stacked

Should stacking be made?

step

Specify the steps for the gridlines in the chart.

If you specify e.g. a step value of 20, and a max value of 100, there will be a gridline for every 20 in the chart (20,40,60 and 80)

title

Overrides the chart title.

Note: The title will also override any filters or basedocumentation that is in together with the original title

titlein

Will show the chart title and any documentation inside the chart itself. Else it will be shown as the PowerPoint SLIDE title

This may be useful if you want to use the Slide title to comment the chart.

type

Specifies the chart type to be produced

universe

The logical expression that defines the global filter for the graph

If combined, the filter will be a subset of the 'universe' attribute in the <xtgroup> element.

values

Should values be displayed within the graph?

wset

A weighting set id to be applied in this graph.

a

Defines the row margin of the cross table

Example: The adress expression a='\4[1,,4,-,+]' will show a table of question 4. The first four rows, no answer and sum of answers will ble shown. (Link: a.htm)

acc

Defines an address or address vector to be applied as a table accumulator

Example: If question 4 is age groups and question 2 is age (quantitative) the following expression <xt a='\4' acc='\2.1'/> will show the age average in the different categories. (Link: acc.htm>

autocode

Ascending or Descending will make an address vector of all occurencies of a numeric data type (F) in the a attribute and the values will appear ascending or descending in the table. The H address component may be present in the a attribute in order to limit the value vector.

Example: Question 2 is age. The table <xt a='\2' autocode='descending'/> will show the frequency of all age groups in the survey in a descending order. AUTOCODE IN INTERVALS? (Link: autocode.htm)

b

Defines the column margin of the cross table (break).

Example: If question 5 is household income, question 1 is gender and 4 is age groups, the table the following expression <xt a='\5' b='ALL \1 \4'/> will show the household income in the rows, and TOTAL; gender and age in table columns. (Link: b.htm)

bar

Specify yes or a list of column numbers if bars should be drawn in each table cell

Note: Not implemented in print, hence it works only in html. Furthermore, only one value is implemented. Example: If question 4 is age groups and question 5 is household income, the table the following expression <xt a='\5' b='ALL \4' bar='3'/> will show the results of column 3 also in a graphic representation. (Link: bar.htm).

basec

Default base for categorical values/computations. If given in the address this will take precedence.

baseq

Default base for quantitative values/computations. If given in the address this will take precedence.

baserows

When bases that effect the show attributes change within the table there is an option to determine whether these bases should be printed or not.

Note: The value none is exclusive to the others, all other combinations are possible.

cdtext

User defined text that should appear in the CD cell (Cell Documentation cell) for this table. If not present this cell will contain standard generated text.

chi2

Should chi2 test be performed? And at which level?

Can be combined with two other attributes: <wchi2> and <chi2calc>.

chi2calc

Should chi2 test be performed with normal (symmetric) or single class calculation?

Must be combined with the <chi2> attribute, and can be combined with the <wchi2> attribute.

deciacc

The number of decimals to be used for accumulated sum values

deciavg

The number of decimals to be used for average and fractile values

decichi2

The number of decimals to be used for Chi2 statistics

decidev

The number of decimals to be used for quantitative deviation measures

decipct

The number of decimals to be used for percentage values

decipop

The number of decimals to be used for population values

decimalsymbol

excludepair

Specify a series of two-letter combinations representing the pairs of columns that normally would be tested by the T test, but will now be excluded from the test.

f

The logical expression that requests the filter population

The filter will be a subset of the 'universe' attribute in the <xtgroup> element.

filterdoc

Should the filter box be printed?

filtershow

Which figures should be present within the filter box:

filtertext

A template text expression to generate the text associated with the logical expression

If no filtertext is given there are two options: 1) When the filter (the 'f' attribute) is a BASIC logical expression (not combined from more questions) the standard template content will be that the generated text will be the text describing the right hand side of the logical expression (answer codes). 2) When the filter is a more COMPLEX logical expression, the logical expression will be shown as a filtertext.

includepair

Specify a series of two-letter combinations representing the pairs of columns that NORMALLY would not be tested by the T test, but will now be included from the test.

introw

Should the number of cases row at table level be printed?

lowcol

The minimum number of cases that a column should have in order to show any values in that column.

pctsign

Which percentage figures should have the % suffix? Specify the figures you want to show with the % suffix:

Example: <xt a='\4' showc='pctvert int pop pcthori' pctsign='pctvert pcthori'/> will show both horizontal and vertical percentages with a % suffix

poprow

Should the table top population row be displayed?

qno

Should the question box number (name) appear as a suffix to the question text?

rolldir

The direction of the rolling to be performed

Example: Question 9.A is a brand awareness question, where the last code is 'Don't know'. The expression <xt a='\9.A' b='ALL \1.A \4.A' rollitem='1' rollvalue='pctvert' rolldir='desc' rollexclude='1' rollend='5' /> will show the TOP (rolling='yes' rolldir='desc' rollvalue='pctvert') 5 BRANDS (rollend='5') at the TOTAL LEVEL (rollitem='1') and the 'don't know any brand' (rollexclude='1') will also be shown at the bottom of the table.

rollexclude

Should a number bottom bars be excluded from the rolling prosess?

Example: Question 9.A is a brand awareness question, where the last code is 'Don't know'. The expression <xt a='\9.A' b='ALL \1.A \4.A' rollitem='1' rollvalue='pctvert' rolldir='desc' rollexclude='1' rollend='5' /> will show the TOP (rolling='yes' rolldir='desc' rollvalue='pctvert') 5 BRANDS (rollend='5') at the TOTAL LEVEL (rollitem='1') and the 'don't know any brand' (rollexclude='1') will also be shown at the bottom of the table.

rollend

The top N rows of each table block will be shown.

Example: Question 9.A is a brand awareness question, where the last code is 'Don't know'. The expression <xt a='\9.A' b='ALL \1.A \4.A' rollitem='1' rollvalue='pctvert' rolldir='desc' rollexclude='1' rollend='5' /> will show the TOP (rolling='yes' rolldir='desc' rollvalue='pctvert') 5 BRANDS (rollend='5') at the TOTAL LEVEL (rollitem='1') and the 'don't know any brand' (rollexclude='1') will also be shown at the bottom of the table.

rollitem

Which column or pair of columns determine the rolling prosess?

Example: Question 9.A is a brand awareness question, where the last code is 'Don't know'. The expression <xt a='\9.A' b='ALL \1.A \4.A' rollitem='1' rollvalue='pctvert' rolldir='desc' rollexclude='1' rollend='5' /> will show the TOP (rolling='yes' rolldir='desc' rollvalue='pctvert') 5 BRANDS (rollend='5') at the TOTAL LEVEL (rollitem='1') and the 'don't know any brand' (rollexclude='1') will also be shown at the bottom of the table.

rollvalue

Which cell show attribute should be used for the rolling?

Example: Question 9.A is a brand awareness question, where the last code is 'Don't know'. The expression <xt a='\9.A' b='ALL \1.A \4.A' rollitem='1' rollvalue='pctvert' rolldir='desc' rollexclude='1' rollend='5' /> will show the TOP (rolling='yes' rolldir='desc' rollvalue='pctvert') 5 BRANDS (rollend='5') at the TOTAL LEVEL (rollitem='1') and the 'don't know any brand' (rollexclude='1') will also be shown at the bottom of the table.

sat

Are <sat> definitions to be applied for this table?

showbreak

Should the cases and population table top row show horizontal percentages?

showc

The sequence of show attributes in categorical row blocks.

showcall

The sequence of show attributes in ALL columns of categorical row blocks.

showq

The sequence of show attributes in quantitative row blocks.

showqcall

The sequence of show attributes in ALL columns of quantitative row blocks.

summary

This option requests that only the calculated scaling rows are printed when the scaling feature is defined as part of the address.

Example: Question A1 is a grd question with a scale from 1 (disagree) to 7 (agree). The follwing table <xt a='\A1[1,,7<>]' summary='yes' showq='avg'/> will show ONLY the average of the scaling, where the codes has been given the code value

textcheck

When a scalar Q component of an address is composed of several question boxes or when a scalar S component through several question boxes are detected there is an option on checking whether the text elements should all be equal to allow the analysis to be done.

Example: Three questions \Q1 (First mentioned), \Q2 (Second mentioned) and \Q3 (Mentioned later) are brand awareness questions covering the same brands. To see total awareness, the table <xt a='\Q1:Q3' textcheck='yes'/> will produce the total awareness of the different brands, and at the same time check that all rows ar the same.

title

Using this attribute will have these affects: The text will appear at the top of the table heading box The text will appear as an index label to the Live Table application. If NEWTABLE=NO then this text will appear at row block text to this table section

ttest

Should ttest be performed? And at which level?

universe

The logical expression that defines the global filter for the cross table

If combined, the filter will be a subset of the 'universe' attribute in the <xtgroup> element.

wchi2

Should the weighted Chi square test be performed?

Must be combined with the <chi2> attribute, and can be combined with the <chi2calc> attribute.

wset

A weighting set id to be applied in this table.

zerocell

A character string to be shown in cells with no cases.

zerorow

Rows that have no cases may be left out

show

Which values are to be shown in the Dig analysis

Specify the values that will be shown in a space separated list. Default values are: 'desc addr vert nvert size chi2d'

a

Specify the question(s) you want to analyze. Set a='\' to get the complete questionnaire as-is, or use an expression to get parts of it.


You may also use the attribute <addr>. It does the same. Best practice is often to specify topbox or bottombox in grid questions, in order to make the groups more robust, e.g. a='\Q5[0:6,9:10]' will only show the topbox and bottombox of a NPS question from 0 to 10.

f

Filter. Specify the target group you want to analyze.

If you combine the filter definition with the use of the <Universe> attribute, the filter (or a part of it) must be a real subset of the universe definition

include

Should positive or negative significant findings be shown

Standard is showing both positive and negative findings

limitchi2

At what Chi2-significance level should the report stop showing results

A standard limit of 4 is implemented, i.e. all significant findings that are listed in the report are significant at a 95% level.

limitgamma

At what gamma-significance level should the report stop showing results

The standard report uses a limit of 0 as the limitation, showing only positive values.

limitindex

At what index level should the report stop showing results

The standard report uses a limit of 0 as the limitation, showing only values above 100.

limitnumber

How many findings should be presented in the Dig-report.

Default is to show all categories that satisfies the significance criterias

decimalsymbol

Any symbol/text to use as decimal symbol

decipct

The number of decimals to be used for percentage values

decichi2

The number of decimals to be used for Chi2 statistics

lang

In which language will the standards texts be shown?

Example: lang='no' will show the table in norwegian. Unless other is specified, default language will be English

filtertext

A template text expression to generate the text associated with the logical expression defined in the target group (<a>-attribute)

If no filtertext is given there are two options: 1) When the filter (the 'f' attribute) is a BASIC logical expression (not combined from more questions) the standard template content will be that the generated text will be the text describing the right hand side of the logical expression (answer codes). 2) When the filter is a more COMPLEX logical expression, the logical expression will be shown as a logical expression.

filterdoc

Should the filter box be printed?

filtershow

Which figures should be present within the filter box:

Specify using a space separated list. Default values are :'int pop pctfilt'

rolling

Which value should be used for the sorting?

The sorting value does not neccesarily need to be a part of the table. Default value is sorting by Chi2 value.

rollsign

Should the list be sorted by using absolute values or actual values.

Sorting by absolute values will show the positive and negative findings mixed in the sorting order

universe

The logical expression that defines the global filter for the dig analysis

If you combine the filter definition with the use of the <Universe> attribute, the filter (or a part of it) must be a real subset of the universe definition