Advanced Crosstab
Contents
1. eee 25 26 Special TUNCUIONS 29 ZEE 29 T gs E E E E E E 32 aieo jor cece cet et Gece eeepc 33 Silo liL Re RE a 34 Rank Format Separate Rank Columns eaaeee eee ea aaaaaaaa aaa aaa aaaaaaaaaaa aaa ennnnnnn nnn nnns nnnna irn n snis nsns aai nnns risa 35 New Functions have been Provided ccccssseccccceeeeceeeeeeeeeeeceseeeeeeeeeeeeeeesseeeceessaeeeeeeeseaueeeesseaeeeessaaeeeeeesaeeeeessaaees 36 SAVING ONN COUE Se sds OTC Eau i O de Ue ee Llosa 87 gol Nolo a A E 38 Code Product Usage or Purchase cccccccccecessseeeeeeeeeeaeeeeeeeeeeeeeeeeeeeeeeeeeseeeeeeeeeeeeeeeesseeeeeseseaeaeeeeeeeeseaeaeeeeeeeeseaas 38 go NO COG E 40 zuo 40 i A A 41 Page 3 Categories The codebook has been created in categories by default all categories are displayed2. 19 cala 20 MNE MONIC 9 20 i es I E 21 lej 816 rz Ulla AR Aie oem 21 geli nO mmm PEE 21 Household Decision Maker 21 FIOUSENOIG Purchaser Weight EE uU 21 OIG 22 Week PN 22 ee ER eee eee ee 22 Average Quarter u 010 CHCE 23 TELMAR SPO Page 2 Total Listenership T u I 24 Average Quarter Hour Listenership eueseee aaa aaaae aaa aa aaaaa aaa aaaaaa nennen nennen nnn nnn ann risa a annes nna 24 gt 6 S 25 Msn Ez rH 25 MEDAN TN 25 VB 25 xl scores 25 pp 25 Bel
3. Intent to Purchese a New vehicie Cade 146191 Semple 1505 Population 45255 98 Explaned Variance O 677720204511348 Cramers v 0 1207242516071 62 Bitmap The process of creating a CHAID analysis 1 Select a study 2 Table Base your target market 3 Columns items you want to review in your CHAID analysis Note It s good if you re going to a conference to have a set of saved own codes for 100 200 demographics or other product items for quick access during your convention The speed is god if you keep your column limit to around 250 4 rows 5 When completed click on Tools CHAID The analysis will run and display a tree style CHAID analysis by default TELMAR ASP Page 34 Rank Format Separate Rank Columns n lt General J Surveys e Coding 2 Printing A Exports th M ultiE asing Email Restore System Defaults Use these settings to configure result options General Stability Style TABBED RESULTS RANKED RESULTS V Show column totals Show row totals I Fixed row totals Separate columns v ward wrapping Query highlight colour uj Ignore zero weighted respondents Size all columns at the same time Use survey codes Show trend colours Cancel An analysis can also be ranked where each column is displayed in a separate window This
4. FF File Edit Codes Options Prototypes Tables Help DE Surveys Coding Tab Rank Prnt When tabbed the description it will include Default Data Items Stability Decimal places Filte the boundaries for the N Tile In our Total example the last column title would appear Audience as Cold Drinks P7D 80 100 6 99 This esos means that for 80 100 the values found Cold Drinks P7D audience per respondent ranged from 6 to 99 The N quantity 0 20 Mesos 1 2 Tile can be used as any other function to be Cold Drinks P7D Audience AND ed and OR ed with In the example tity 20 40 R E sd a Row and Table coding expressions may ISIBEETEUEENEERES EHE also exist to be crosstabbed quantity 40 60 3 4 Cold Drinks P7D Audience quantity 60 50 Resps 4 6 aca Cold Drinks P7D Audience quantity 8090 100 esos 6 99 AMPS 200240 Total Radia TELMAR SPO Page 31 QCodes QCodes are special way of looking at attitudinal questions relating to media and product usage side by side They may only be used as rows and they may not be mixed with other coding statements and if they are used all rows must be defined as a Qcode Additionally there must be at least one column defined The coding for the Qcode statement appears as follows Beeld Read Right Through Read qcode BEE BEE 6071 6072 6073 6074 6 Most Read Some Read Selec
5. This will work out a frequency distribution against all respondents in the survey i e including those that did not answer the particular question you are analysing NtilesZ This will work out a frequency distribution against those respondents that answered the particular question only It will ignore all zero respondents The coding expression to perform the N Tile coding NTiles on is the uppermost text in the window The window begins witha e esmmm breakdown of 5 NTiles You may edit this field and then click on the Set NTiles Tles s Set NTiles button This will reset the Start and End percentages 20 40 40 60 60 GU 80 100 zi Clear Calculate w OK 3 Cancel TELMAR amp se 30 The percentages represent the tile you wish to extract from the coding expression For example 0 to 20 will extract the first 2096 of Cold Drink users You may also manually edit the percentages within the grid However your Start percentage may not be less that 0 and your End percentage may not be less than the Start percentage or higher than 100926 Cold Drinks PFD Quantity CDRINK Click OK to then place your NTiles into your selected grid The NTile function parameters are as follows Start 96 End 96 and expression to NTileZ This can be done E m DL manually as opposed to using the wizard Untitled Table
6. Winning Media Solutions NEW INT ADVANCED CROSSTABBING USER MANUAL 2008 Page 1 Table of Contents 672 216 CZE T mm 4 eR 4 E E EE E milieu 6 Other Advanced Features 7 lis c M 7 Colour He eera E eld Geese E E A E E 8 6721 610 6 9 NN T E E N A EE ERE 9 Calculation FUNCIO SERERE Um T 10 v YAN SO mr 11 20M Advanced Soning ERR 11 Oe 6 0 Ct 12 Arithmetic OCT ANON S icm Ment ERR Hexen dca muni i diae EE CEC 13 Ie m NM m 13 OP OE NR o m 14 10e mutply OST ERR Emm 14 ANSTAMIE OPEN Kc 14 Relational NRI ER 15 Coding Logical Operators PE CEA 16 WNEFANO SS NOR secie m 16 TE ACE 5 cM E 17 ol en ee eons eer rere eee eee re 18 OPETIOS soe EET Emme 18 aei 18 NV OUTING 1S
7. 34 o e and the brackets are used in this coding expression The order of priority is firstly BRACKETS In the absence of BRACKETS the order of priority is firstly the AND then the OR Brackets should always be present when more than one operator is being used TELMAR SPC Page 18 Volume Coding Volume coding is different to the normal card column punch coding where the answer is either positive or negative With the volume question there is a range of possible numeric answers To cope with this range of possible answers TELMAR uses a special type of coding namely the use of N meaning any number EXAMPLE How many TV sets are there in your household This particular question is a volume question where the number of TV sets is the volume There is a range of possible answers for that particular question The coding for this question would be MO7N DS where MO7N would give the total NUMBER of TV sets in all households How does it calculate total number of TV sets Respondent 1 had 2 sets Respondent 2 had 1 set Respondent 3 had 4 4 sets Total number of sets ANSWER EQUALS THE NUMBER OF SETS The coding for people who have 2 TV Sets MO7N 2 1 3 TV Sets MO7N gt 1 MO7N lt 3 ANSWER EQUALS THE NUMBER OF PEOPLE WITH SETS Number of TV sets can be coded using mnemonics as follows TVSETS total number of TV sets TVSETS 2 no of people with 2 TV sets TVSETS gt 1 TV
8. 5 16 30 17 30 Mon Fri The code 13067 0 13167 0 13267 0 13367 0 13467 0 DIV 5 A breakdown of the code The bold numbers above indicate the different days of the week Mon Fri DIV 5 the division of the number of days What is actually happening The code is measuring everyone that listened each day Regardless of how long a person listened or how many times they switched on the set during that day they would be counted as one person It then adds all the days together and divides that figure by the number of days e g 5 to get an average per day TELMAR SPC Page 22 Average Quarter Hours Average quarter hour figures are obtained by adding together all the quarter hour segments in a time channel and dividing by the total number of quarter hours The result is a number but when used with a demographic will give you the number of people who listened in the average quarter hour EXAMPLE 1 Consider the following Radio South Africa average quarter hour on Monday from 06 00 07 00 The code is 11031 4 DIV 4 A breakdown of the code 11 The code for Radio South Africa Card1 1 03 Monday 0600 0900 Column 1 06 00 06 15 Punch The addition of the other quarter hours within the 06 00 07 00 hour 4 Represents the last quarter hour in the 06 45 07 00 i e 11034 DIV4 In the hour 06 00 07 00 there are four quarter hours therefore to get the average quarter hour one must divide by 4 EXAMPLE 2 Consider the fo
9. Coverage Rank Index Rank Scal Rank Plow Rank Table Table Rank Projected Proj Rank Resps Rank Sigma Tolerence 2 Sigma Tal Rank 2 Sigma 2 Sigma 95 Rank Cost 1 Unit Cast Cost 1 Unit Cost Rank Cost 1 CPT Rank Cost 1 Val Imps Cost 1 Val Imps Rank Audience Rank 5Caomp Rank Page 28 Special functions Ntiles Definition In an R amp F analysis quintiles are the division of all people reached into 5 equal parts that is each part represent 2096 of all people reached or all people shown in the frequency distribution The data arranged in an increasing order of magnitude Schematic representation of quartiles of a data set Relative frequency r Qu lt 2 Lower quartile Medjan UPPST quartile Schematic illustration of quartiles of a distribution In this instance the X horizontal axis is showing the quartiles with quartile 1 being the lowest reached group On the Y vertical axis you see the amount of frequency Therefore for each of the quartiles if there were 5 on the Y axis you d know the amount of frequency that was obtained for each of the quartiles If you re doing a print R amp F in MediaPlanner you can print a quintile analysis showing the amount of impressions GRPs and other info for each quintile Since quintile 1 is the low exposure low frequency group they get the lowest amount of impressions GRPS etc for their group For quintile 5 t
10. The OR button will add the word or to a line of coding as you enter individual codes e g Male OR Female And The AND button will add the word AND to a line of coding as you add individual codes e g Female AND 12 17 Not The NOT button will add the word NOT to a line of coding as you add individual codes e g Drink Beer AND NOT Drink wine SUM The SUM button will SUM add all items selected to make them usable with comparison or numeric signs The Double Bracket button will bracket full line of coding or part of the line if is selected 8 The Open and Close Bracket buttons allow you to bracket sections of coding as you build you codes Double Arrow The double arrow button will replace any selected items in the editor grid with any selected items in the codebook as long as the number of items selected is the same TELMAR xSP 26 Assigns values 1 9 for punches 1 9 in a column BIN To see the number of items purchased rather than the number of people purchasing only available for some surveys CLUS Similar to Num statement but extracts columns differently MEAN Mean of any numeric vector can be filtered by a prefix within brackets e g MAL Mean Age MEDIAN Mid value in a numeric vector May also be prefixed as Mean MEDIANZ Mid value in a numeric vector excluding zero values May also be prefixed as Mean or Median SOLUS Exclusive SUM OF Same as Sum VAL Assigns given values to different punches in a co
11. but can be deselected using the categories button There are two ways of accessing the Category Selection window 1 Click on CATEGORIES at the bottom of the Selection Tree 2 Right mouse click in the Selection Tree window and select Categories The items included in each category are differentiated by the colour of the text E g Items included in the General category are coloured yellow Demographics By Group Category selection Bl Available Categories Prototypes Demographics No Group Own Codes Avera ge Data Own Publications _ Demographics Home Information General er Products Activities Branded Praducts Travel Media _ _ _ Publicatians Financial Services Yehicles fTransport Modes of Transport shopping Appliances Products Personal Usage Products HH Purchased Brands Personal Usage Brands HH Purchase TY Information Radio Information Cinema show Filtered m Cancel Outdoor Advertising Print General Print Readership By Group Print Readership No Group Show all When a specific category is selected in the Category Selection window all other categories disappear from the Selection Tree To bring all categories back into the Selection Tree go back into Categories Click on Show All and then OK TELMAR KSP 4 Search Clicking on Options below the search function will show the detailed search options The user is able to save a selected set of options as defaul
12. code of the audience structure Rank of the audience vs all the audiences in the Colum Average Issue Readership Rank of the AIR vs all the AIR in the column Rank of the Comp vs all the Comp in the column Reach in 96 Rank of the 96 Cov vs all the 96 Cov in the column Rank of the Index vs all the Indexes in the column Rank of the 96 Col vs all the 96 Col in the column Rank of the 96 Row vs all the Row in the column Audience vs the total audience of the Table Ranks of the 96 Table vs all the Table in the column Audience weighted Ranks of the Projected audience vs all the Projected audience in the column Rank of the respondents vs all the respondents results in the column 2 sigma or two standard errors this means the chances are 95 out of 100 that the reported audience e g the rating would not differ from the findings of a complete census of the population by more than the number reported For example if a 10 rating has a 2 sigma of 2 and you conducted the same audience survey among 100 different samples of people you would find a rating of between 8 and 12 10 or 2 in at least 95 of them In the remaining 5 samples you might find a rating that varies from 10 by more than 2 Therefore you can be confident of the range 95 percent of the time a 95 percent confidence level Data ltem Selection Selected data iterns Available data items 000 Resps w C gl camp
13. format is activated via the Edit Preferences menu switching Separate Rank Columns on In this format each window is sorted File Edit Codes Options MultiEasina Tables Tools separately Right mouse clicking on the title Table All Women will sort each window into the most appropriate order Right mouse clicking again will reverse the order In the example below the data has been sorted by Audience Help Surveys Coding CrossTab Rank Switch Default Data Items Stability Decimal places Filter Highlight Ce Women Audience 000 Resps 96Col 96ROW 15 608 12 407 100 0 50 2 1 16 24 4 024 2 845 25 8 49 0 2 25 34 3 574 2 698 22 9 47 8 3 4 201 3 360 26 9 50 4 4 3 808 3 504 24 4 Table All Men Default Data Items Stability Decimal places Men Audience 000 Resps 15 501 12 405 16 24 4 190 3 808 25 34 3 897 2 550 35 49 4 131 2 849 50 3 283 3 198 Filter Highlight Ce Col Row 100 0 49 8 27 0 51 0 BEL B2 26 7 49 6 21 2 46 3 TNT Untitled Print olour Tiles Index 100 0 37 6 95 4 100 5 107 0 olour Tiles Index 100 0 102 4 104 7 99 5 92 9 Winning Media E Page 35 New Functions have been provided SOLUS results in a count of one and only one of a number of items For example if you want people who read only one of two publications but not both you can code SOLUS PUB1 PUB2 VAUD works like VIND but operates on Audience inste
14. you the mid point in a range of numbers For Example MEDIAN AGE middle age in a range NUM NUM is used in volumetric coding to show a number For Example NUM M 07 1 will give you the number of television sets in South Africa RAN RAN is used to combine a range of coding For Example AGE RAN 20 30 will give you people between the ages of 20 and 30 VDIV VDIV is used to INDEX one product code against another For Example BEER H VDIV WHISKY E will show where beer has a stronger market than whisky DEC n DEC n You can add the coding DEC 2 at the end of a coding statement so you can see 2 decimal places TELM R k SPC Page 25 Coding Summary Numeric multiplication Numeric division Numeric addition Numeric subtraction GT gt Greater than GE gt Greater or equal to LT Less than LE lt Less or equal to EQ Equal to NE Not Equal VDiv Compare audiences e g if 15 44 vdiv age45 if you have 1000 15 44 year olds and 2000 45 year olds using vdiv the answer will be 0 5 divide one by another MEANZ Same as mean excluding respondents with zero values NUM Extracts Numeric vector from CCP files DNUM Extracts floating point vector from CCP files Number To integrate a number RAN Extracts respondents that fall within a range Any The Button will take any selected codes and create a bracketed OR statement All The All button will take any selected codes and create a bracketed AND statement Or
15. 0 Search Desktop start Inbox Micro i91 NEW TNT 20 al W untitled Paint Ye Tranmene TNT2001 Ge Telmarupdater 2 PQ 12 59 PM TELMAR KSP 8 Calculations This item is valid only if there is more than one survey tabbed A calculation can be performed on one survey against the other TNT Default Switch Filter Rank CrassTab Data Items Stability Coding Surveys Highlight Colour Calculations urvey order EEE Decimal places Calculation C Difference Percentage Difference Average C sum Perform calculation against all Survey Display wf Calculate C Calculate all surveys against a base f Calculate all surveys against a previous More Options t cancel Winning Media Solutions C Page 9 Calculation Function Perform Calculation against all o If ticked on this will calculate all surveys in two ways 1 Against a base The user can select which survey is to serve as the base against all the other surveys This is done by selecting the relevant base survey from the Primary Survey list 2 Against a previous If this is selected each survey will be trended against the previous survey in the Primary Survey list e g In the example below 200746 will automatically be trended against 2006A6 o If ticked off this will calculate the second selected survey against the first This optio
16. DIDAS d Page 16 The Operator The operator tells the software that the result is POSITIVE if either or both of the recorded responses is POSITIVE eg People who read the Femina or the Cosmopolitan Femina OR Cosmopolitan o FEM COSM The or the OR is used to join the mnemonics The computer would then count all people who said they read the Femina or the Cosmopolitan or both the Cosmopolitan and the Femina says they may be two things either one or the other getting bigger With OR the computer takes both universes with the result and is never smaller that the largest of the two Group 2 CLEAVAGE mu SALAS A DANII HOLIDAY 54 A VRLA SETY ADIDAS TELMAR e 17 The Not Operator Unlike the AND and operators the NOT operator does not tell the software how to combine the responses from two answers to get a single result The NOT operator tells the software to exchange POSITIVE for NEGATIVE and vice versa The result is that the software counts the NEGATIVE responses eg People who do not have a television set in their households ALLeNOT TV ANY The eNOT is used to join the mnemonics Combining Operators The software has a strict order or priority when using more that one operator i e using both the AND and the OR operators eg Females between 16 and 34 e Female AND Age 16 24 OR Age 25
17. OW 100 0 index 100 0 TELMAR SPC Page 7 Colour Tiles x Tiling Colours Colour Tiles will colour each block of data according to aM Ed ET mm E size So a high index would be Red a low index would Lowest Medium High Highest be Blue Any data item can be used for the tiling M tles per eslumn If Exclude the Total Row Data Item to colour 000 Resps Scal SR ON TNT Untitled Table Western Cape File Edit Codes Options MultiBasing Tables Tools Help TNT Surveys Coding CrossTab Rank Switch Print BX Excel fr Default Data Items Stability Decimal places Filter Highlight Colour Tiles Calculations Survey order Total Men Women otal Audience 000 3 074 1 505 1 569 Resps 3 340 1 666 1 674 Co 100 00 100 00 100 00 R0W 100 0 48 9 51 1 Index 100 0 100 0 100 0 16 24 Audience 000 623 313 310 Resps 805 461 344 96 20 27 20 81 19 76 Clear Row 100 0 50 2 49 8 Index 100 0 102 7 97 5 25 34 Audience 000 700 353 348 Resps 561 268 293 Col 22 78 23 44 22 15 wf Apply to all tables wf Apply 3 Cancel Row 100 0 50 4 49 6 Index 100 0 102 9 97 2 35 49 Audience 000 965 478 487 Resps 886 407 479 31 40 31 79 31 02 Row 100 0 49 6 50 4 Index 100 0 101 2 98 8 0 Audience 000 785 361 425 Resps 1 088 530 558 25 55 23 96 27 07 Row 100 0 45 9 54 1 Index 100 0 93 8 106 0 xl Survey AMPS 2006 2007 Main Branded Weight Population Report Units 100
18. SETS lt 3 no of people with 1 3 TV sets Volume coding can also be used when wanting to look at exact age groups For example if one wanted to look at those in the age group 20 30 one could code this in two ways TELM R k SPC Page 19 Card Column Punch If a person replied that they were 21 years old there would be 2 columns punched i e a mark at 2 in the first column and at 1 in the second column AGE OF RESPONDENT COLUMN AGE OF RESPONDENT COLUMN The first column is in tens and the second column is in units i e 21 equals 2 X 10 1 Therefore the age group 21 30 would be coded as follows S15N 10 S16N gt 21 S15N 10 S16N lt 30 OR IT COULD BE CODED AS FOLLOWS NUM S 15 2 gt 21 NUM S 15 2 lt 30 MNEMONICS AGE 21 all people who are 21 years old AGE gt 21 all people who are over the age of 21 AGE gt 21 all people 21 years or older AGE lt 30 all people under the age of 30 NOTE in AMPS only 16 29 would be included AGE lt 30 all people who are 30 years old or less AGE gt 21 AGE lt 30 all people who are 21 30 years old Page 20 Weightings The way that questions are asked will affect the weights used In general there are four weights Population Household Household Decision Maker Household Purchaser Population Weight This weight is used for questions that are directed at the individual in the household For example How old
19. ad of Index In the example below there are 390000 Cape Argus and 270000 Cape Times readers with a duplication of 143000 The SOLUS function will result in readers who read one and only one of these two publications i e 374000 readers The VAUD function will result in a ratio between the number of Cape Argus and Cape Times readers i e 390000 270000 144 or there are 4496 more readers of the Cape Argus than the Cape Times NI Untitled Table 2 PY File Edit Codes Options Tables Help TNT Surveys Coding CrossTab Rank Switch Default Data Items Stability Decimal places Filter Highl Total 16 24 Audience Resps ROW index Cape Argus Audience 390 68 Resps 553 129 ibis 0 8 9 6Row 100 0 17 5 Index 100 0 64 3 Cape Times Audience 270 32 Resps 376 70 0 9 0 4 Row 100 0 11 9 Index 100 0 43 7 Cape Argus or Cape Audience E 84 imes Resps 736 166 Co hak 1 0 R0W 100 0 16 3 Index 100 0 60 1 Cape Argus and Cape Audience 143 16 imes Resps 193 33 0 5 0 2 100 0 11 0 Index 100 0 40 5 SOLUS Cape Argus Audience 374 69 Cape Times Resps 543 133 9 eCol alos 0 8 9 e6Row 100 0 18 4 index 100 0 67 6 Cape Argus IND Cape Value 100 147 imes Resps 376 376 Co AROW index Cape Argus YAUD Cape Value 144 212 imes Resps 376 70 Co R0W Index TELMAR A SP 36 Saving own codes Saving Own Codes now has the option of saving only the highlighted entries in the coding gr
20. alysis by Household Decision Maker Weight is meaningful Household Purchaser Weight Any respondent of either sex who claims to be solely or partly responsible for the household purchases These respondents weighted to households form the basis of the values relating to household FMCG categories TELMAR k SPC Page 21 inning Media Solutions Radio Coding Radio Coding is split into quarter hour segments in the survey Information is for an average day Monday Sunday The examples below pertain to the Radio Diaries In AMPS the cards are all alpha letters digits e g Card B Because of the size of the Diaries there are more than 26 cards therefore the cards are all numeric numbers digits e g Card 2 Week Cumes A week cume calculated the maximum number of people that can be reached in one week for a particular station during a particular time period the net listenership Consider the following Radio 5 16 30 17 30 Mon Fri The code is 13067 0 13167 0 13267 13367 0 13467 A breakdown of the code The bold numbers above indicate the different days of the week Mon Fri What is actually happening The code is measuring everyone that listened at least once during the week Regardless of how long or how often they listened they would be counted as one person Day Cumes A day cume gives the average number of people per day who listen to a particular time period Consider the following Radio
21. are you What car do you drive Have you read Bona in the past 6 months you personally tried to lose weight by using slimming preparations in the past 12 months Household Weight One person or a group of persons usually but not always members of one family who live together and whose expenditure on food and other household items is jointly managed Boarders or lodgers may be included as members of a household provided that they have at least one main meal a day communally Resident domestic servants are excluded and regarded as forming a household of one or more persons in their own right The weight is used for questions that refer to the household and is used with regard to certain products such as having a stove in the home Household weights can only be used on the following demographics Geographical Location A household due to it being all members living in that home cannot have an age sex or read any particular magazine or watch a particular TV station Household Decision Maker Weight If the analysis is to be done to establish media usage or other personal demographics for household questions then the Household Decision Maker Weight should be use after filtering on the HH Decision Maker Male Head of Household or Female Housewife Example If one needs to find out People with a Household Income who lives in Gauteng The wording of the question and the purpose of the analysis determines whether an
22. ces of various publications or various demographics To launch the Prototypes screen click on Edit Prototypes within the Prototypes menu item Prototypes Survey for Results gt Reload Pro file Title Coding Mnemonic AIR RespgjAlR Proj The Prototype screen appears and 2 hosts BOB and DOF PROTOtmp 402 B1 automatically loads the prototype file of 1 host BOB temp the selected survey The survey combo box shows all loaded surveys in your current run Prototypes may at any time be reloaded for the selected survey by clicking on the Reload Pro File button Test Protototypes Cancel Save You may add delete or edit the prototypes All mnemonics given to prototypes must be prefixed by PROTO Existing prototype files will however still work The prototypes in the grid may be tested by clicking on the Test Prototypes button at the bottom of the screen All errors with prototypes will be reported and be highlighted in red If there are no errors the AIR respondents and projected values can be seen Remember that only mnemonics provided by Telmar for prototyping may be used here Hence ALL and other such mnemonics are not supported in prototypes Once completed click on SAVE to save the prototypes Prototypes will only save if they are all valid To use prototypes simply type the mnemonic of the prototype in your coding To use Prototypes via the selection tree they appear at t
23. es DF v Save cancel Clear e shart tie Add to Tables Columns Rows Go Go Options Test All Big Grid Selection Tree Selection Decision 5 2003BP Main TEL ARX SP 37 How dol Code Product Usage or Purchase 1 Amount Used 2 Average Usage Take Average over Take Quantity over into coding grid into coding grid Click on AND Take Light Medium Heavy Over Take Light Medium Heavy Over Only use Audience Only use Audience Figures Figures Page 38 3 Ave number used If you want to find out how many bottles a heavy consumer consume TELMAR k SPC Page 39 How do I code Exact Age Page 40 Glossary A Adults A person over the age of 16 years as defined by AMPS AMPS The All Media amp Product Survey examines the readership listenership television and cinema viewing levels as well as product amp service consumption usage and purchase habits F Frequency The number of times on average that a person within the target market is supposed to have been exposed to the advertiser s message Freq Cumulative total of ratings Total Unduplicated Audience Reach H Household One person or a group of persons usually but not always members of one family who live together and whose expenditure on food and other household items is jointly managed Boarder
24. h the second operand The result being a number EXAMPLE WHI TVSETS When WHI TVSETS is used WHI converts to 1 as itis POSITIVE The result is a number The divide operator The DIVIDE operator divides the first operand with the second operand The result being a number This is used primarily for radio coding EXAMPLE Total Community Radio Average Quarter Hour from 05 00 06 00 The code is Y71Y 1 4 4 In the hour 05 00 06 00 there are four 1 4 hours therefore to get the average quarter hour on must divide by 4 Works the same way as In this example it is looking at all the punch positions from Y to 1 can use because this is an arithmetic code Relational Operators The result of a relational operator is logical Each operand must be numeric as you are comparing numbers If either operand is not numeric it is converted to numeric on the basis that POSITIVE would be 1 and NEGATIVE 0 The relational operators are 1 gt Greater than The result is positive if the first operand is greater than the second operand 2 gt Greater than or The result is positive if the first operand is greater than Equal to or equal to the second operand 3 Lessthan The result is positive if the first operand is less than the second operand 4 lt Lessthan The result is positive if the first operand is less than Equal to or equal to the second operand 5 Equalto The result is p
25. has to click on Data Items and select Cost 1 CPT The user either double clicks on the Cost 1 CPT option or highlights it and use the arrow key Data Item Selection Cf x Default Data Items Stability Decimal places Filter Highlight Colour Tiles Calculations Survey ort Total 16 24 and 25 34 and 35 49 and 50 and Women Women Women Audience 000 31 109 4 024 3 575 Available data items Selected data items Comp Rank Audience 000 Resps 21 068 2 478 2 195 Coverage 100 0 100 0 100 0 100 0 100 0 Cov Rank Row 100 0 12 9 11 5 13 5 12 2 Index 100 0 100 0 100 0 100 0 100 0 Index Rank Cost 1 CPT 0 0 0 0 0 8 O m Cosmopolitan Audience 000 Table Resps 1 057 261 153 Table Rank Col 3 0 5 7 4 3 i Rank Index 100 0 186 8 143 1 pan qu Cost 1 53 220 323 2 Sigma Tolerence Load Bona Audience 000 2 214 379 341 2 Sigma Tol Rank Resps 1 110 216 174 185 71 2 Sigma 6Col wozi 9 4 95 8 8 4 3 Aero dig 5 100 0 1324 1341 123 6 59 8 i naex EE on E Cost 1 CPT 14 82 91 84 192 Cost 1 ul Rank Elle Audience 000 389 107 87 79 33 Cost 1 Val Imps Resps 407 114 65 87 52 a al Imps Rank Col 1 3 2 7 24 1 9 0 9 Cost 1 CPT gt ROW 100 0 27 4 22 2 20 3 8 5 Index 100 0 212 1 193 6 150 2 69 2 PM p Cost 1 CPT 97 352 435 477 1 142 pply to all tables a Rows C Columns Essentials Audience 000 220 17 28 73 48 Re
26. he first node Page 33 Chaid Analysis CHAID chi square Automatic Interaction Detector is used to simplify the process of recognizing significant natural divisions within the market This will quickly identify niche markets This distinguishing method facilitates immediate and accurate analysis of an extensive set of demographics categories against a desired target CHAID e Identifies those characteristics that are key to explaining the variation in responses to a question e Produces a tree diagram showing the order and magnitude of effect e Could be used for example to identify which characteristics are key to determining whether physicians are high or low prescribers of a new product e gt visual impact Results can be exported to a Bitmap file for easy insertion into presentations Simultaneously examines all of the demographics selected their interrelationships and their predictive power of the target This approach goes beyond a simple cross tabulation sorted by index Telmar s CHAID analysis provides a user friendly graphical representation of the statistical data 2 IDEV_INIPlus01 Melat rame Presentation Server Client TMT Chaid Analysis ExpancAll Contract ll s a Ch sis T Views Ong Char Protie C Hora nai A fh Very Likely Colour Tree Level mcm fe Colour Tes Fiom mes tm Wei Use Index E saiuadajd Selected Irem Details
27. he high exposure group you see more GRPs and impressions It sort of shows what you already know people who read a lot of magazines are seeing more impressions opportunities to see your ad than those people who only read one or 2 magazines Same with TV the people in the high exposure quintile 5 are going to get more of the GRPs and impressions in a TV R amp F quintile analysis the people who watch the most TV of all the people you ve reached are the people who have more opportunities to see your ad impressions Page 29 1 In TNT if you ask for quintiles you re asking to take all of the volume of a product purchased and divide the users into equal sized groups most clients use quintiles for 5 groups or terciles for 3 heavy medium light users The common usage is to make these Ntiles of product usage in columns in a crosstab then media or demographic info as rows You could find then for example that heavy peanut butter purchasers also tend to be most like to have 2 children OR that the heavy high quintile group users of healthy cereal also tend to be most likely to enjoy camping or hiking To use the N Tiles wizard you firstly code the expression on which you wish to base the N Tile This can be done on the Table Column or Row grid Right click on the code in the row you wish to perform the N Tiles on and a popup menu will appear You can either click on the N Tiles or NTilesZ menu item Ntiles
28. ids In order to save own codes e he userclicks on Codes and then Save own codes A dialog box appears which the user can either specify to only save the codes that are highlighted or to save all the codes in either Tables or Columns or Rows o After selecting the above options the user needs to give a description for the saved codes TNT Untitled gt ARA 8 x File Edit Codes Options Tables Help TNT Surveys Coding CrossTab Rank Switch Print Excel Email Coding Survey AMPS 2003BP Main AO3BP 5 gt gt lt lt lt gt VDIV MeanZ Num DNum Number RAN More 0 Tables 4 Columns Selection decision Selection tree Own Codes Demographics By Group Summary i 16 24 Demographics No Group 25 34 Average Data 35 49 a Home Information 50 Activities B Daily Travel Beeld Financial Services z Die Burger OR Vehicles Transport Business Day Shopping Cape Argus AND Appliances Cape Times NOT Products Personal Usage The Citizen Products Personal Purchased ja Daily Dispatch SUM Products HH Purchased Daily News TY Information Daily Sun Radio Information Cinema Outdoor Advertising Print General Print Readership By Group amp Newspapers Daily Daily Supplements c Bi Weekly c Weekly Weekly Supplements Monthly Newspaper amp Magazines amp Newspaper Regular Reader gt Reset Categori
29. llowing Radio South Africa Tuesday 08 00 10 00 The code is 11139 Y 11141 4 DIV 8 A breakdown of the code 11 the code for Radio South Africa Card 11 13 Tuesday 06 00 09 00 Column 9 E 08 00 08 15 Punch Note If the column changes e g from 3 to 4 then use one and start the code again DIV8 In the hour 08 00 10 00 there are eight quarter hours therefore to get the average quarter hour one must divide by 8 TELM R k SPC Page 23 Total Listenership Total listenership measures the gross listenership of a station It simply adds up all the times the station was listened to without taking duplication into consideration Consider the following Radio 5 12 00 14 00 Mon Fri The code is 13051 8 13151 8 13251 8 13351 8 13451 8 Average Quarter Hour Listenership Average listenership measures the average listenership of a station It adds up all the times the station was listened to without taking duplication into consideration and divides it by the total number of quarter hours of the time channel Consider the following Radio 5 12 00 14 00 Mon Fri The code is 13051 8 13151 8 13251 8 13351 8 13451 8 DIV 40 Statistical Coding MEAN amp MEANZ MEAN amp MEANZ both give you the average of a range of numbers MEAN calculates the average based on the total sample MEANZ calculates the average for the people who actually answered the question For Example MEANZ AGE Average age MEDIAN MEDIAN gives
30. lumn Vind Returns the index of the 2 weighted vectors taking their universe into account e g age15 44 vind age45 you will get a comparison index showing how they compare e g if 15 44 year olds index at 150 and 45 at 75 the answer will be 200 Only available in monobase VMul Returns the product of 2 tabbed weighted vectors multiplies audiences to compare them only use in some specific surveys Only available in monobase VAUD Compare audiences as for Vdiv but gives results as a e g if 15 44 vaud age45 if you have 1000 15 44 year olds and 2000 45 year olds using the answer will be 50 Only available in monobase Repeat Only use for some specific UK surveys Page 27 Data Items Audience 000 Resps Col Row Index Survey Comp Audience Rank A I R Rank Comp Rank Coverage Cov Rank index Rank Col Rank Row Rank Table Table Rank Projected Proj Rank Resps Rank 2 Sigma Tolerance 2 Sigma Tol Rank 2 Sigma 2 Sigma Rank TELMARKsPO This is the grossed up value based on the total Respondents to represent the total population Respondents The actual amount of people who answered the questionnaire and answered positively to the question The percentage of the Column entry based on the row entry The percentage of the Row entry based on the Column entry A percent in relation to a norm of 100 Any value above 100 are more likely and anything below 100 is less likely Survey s
31. ly sections with the words Print Cinema or Outdoor in them Selection Tree Demographics By Group Demographics No Group Average Data Home Information Activities Travel Financial Services Vehicles Transport shopping Appliances Products Usage Products Purchased TV Information Radio Information Favourite Radio Station Cinema Outdoor Advertising m m m Cinema Outdoor Advertising Print General Print Readership By Group Print Readership Ho Group sf Reset Categories Use short Reset Categories Use short Print Cinema Outdoor Go f Selection Tree Selection Decision Gu fe Selection Tree Selection Decision TEL ARX SP 6 Other Advanced Features Untitled Table FF File Edit Codes Options fulifesing Tables Tools Help Surveys Coding CrassTab Rank Switch Taf Excel E PDF Default Data Iterns Stability Decimal places Filte Highlight Colour Tiles Calculations gt irvey order Sort Audience 000 Mesos eo Seo Index Blacks Audience 000 23 446 6 766 5 893 6 097 4 689 Resps 10 363 3 520 2 720 2 517 1 606 75 4 a 76 9 foe 66 1 2550 100 0 28 9 25 1 26 0 20 0 Index 100 0 109 3 104 2 97 1 GZ Highlight Data IET Data Item AMPS 2007 Main Bra In order to quickly find data that meet certain criteria am al SR Highlight Col
32. n will only calculate two surveys at a time To trend more than 2 surveys click on More Options and select the surveys you want to trend For example to trend A01BP on AO1AP click on A01BP in Primary Survey click on in Survey to Subtract and click on AO1AP in Dividend TAT Untitled Table Western Cape Pf File Edit Codes Options MuliEasing Tables Tools Help TAT Surveys Coding CrossTab Rank Switch Print BX Excel TE POF Em Default Data Items Stability Decimal places Filter Highlight Colour Tiles Calculations Survey order AMPS AMPS AMPS AMPS AMPS 2006 2007 2005 2006 2006 7007 2005 2006 2006 7007 2005 2006 Main Branded Main Main Branded Main Main Branded Main Total Non Branded Men Non Branded Women Non Branded Women Audience 000 Mesos Index Audience 000 Mesos eco ADOM Index Audience O00 Ress ROW Index Audience 000 esos GC z GROW Audience O00 TELMAR KSP 10 Survey 4MPS 2005 2006 Main Non Branded Display Order 4 This item is valid only if there is more than one survey tabbed Then the Delete survey displayed and the order of display can be selected within this dialog box Calculations created under Calculations can be removed from wf Ok cancel here Sort Advanced sorting iol xi Sorting options Show 10 Rows Show Bottom 10 Rows Item Sor
33. ng Research Foundation Conducts advertising research on behalf of the advertising industry SABRE South African Business Research Evaluation A media research survey specifically aimed at the business market Sample In statistics a group of individuals regarded as representative of a whole population that are selected for study questioning Sample Area In research the geographical area within which a respondent sample is obtained Sample Error In statistics the deviation between the observed characteristics of a sample and the characteristics of the population from which it is drawn inherent in all samples and inversely related to sample size Solus A guaranteed position as the only ad on a page or in a commercial break W Weighting In general the application of ration estimation to adjust in tab sample data from samples so that the weighted sample is in balance with the universe for various household persons characteristics The way that questions are asked will affect the weights used In general there are four weights 1 Population 2 HouseholdHousehold Decision Maker 4 Household Purchaser Winning Media wer 3 Page 42
34. ositive if the first operand is equal to the second operand TELM R k SPC Page 15 Coding Logical Operators Up until now the sample coding statements have consisted of a single code This will often be sufficient but the software becomes most useful when joining codes to establish characteristics of target markets For example you might want to know not only how many people drink tea but how many of them are men how many of them are women how many of the men are within a certain age group The operators used to join codes or mnemonics are OR NOT The And Operator The AND operator tells the software that both answers must have POSITIVE responses recorded eg Females who read the Cosmopolitan Female AND Cosmopolitan o FEMeCOSM The e or the AND is used to join the mnemonics The software would then only count females who said they read the Cosmopolitan The operator can be used to join a number of mnemonics eg Females 16 24 who read the Cosmopolitan and live in Johannesburg Female AND Age 16 24 AND Cosmopolitan AND Johannesburg In the definition AND they must be both one thing and the other i e getting smaller In the AND one universe falls within another i e beauty conscious within married females The result can never be larger than the smallest of the two Group 1 Group 2 IPI n E WIN HOLIDAY 4 PRIVATE VILLA SERY A
35. our Highlight will allow the user to specify a single or a indes B series of criteria e g index 100 and all the rows that meet that criteria will be highlighted The highlight colour can be changed by clicking on Highlight Colour gt lt Columns and selecting a preferred colour Tatal Blacks Coloureds Indians Whites White Coloured Indi Default Data Items Stability Decimal places Filter Highlight Colour Tiles Calculations Survey or LSM 1 AND statements HIM qM Total 16 24 25 34 35 49 50 2 otal Audience 000 31 109 8 214 7 471 8 332 7 091 OR statements Resps 21 068 5 685 4 380 5 352 5 651 Co 100 0 100 0 100 0 100 0 100 0 Row 100 0 264 24 0 26 8 22 8 Index 100 0 100 0 100 0 100 0 100 0 AMPS 2007 Main Branded AB Feb 2007 Nov 2007 000 16 24 Col 20 OR AMPS 2007 Main Branded Feb KPE 2007 Mav 2007 25 34 Col gt 203 OR AMPS 2007 Main 6Col Branded AB Feb 2007 Hov 2007 35 49 94 20 OR Row w Index Coloureds Audience 000 2 734 643 655 849 588 Clear wf Apply to all tables Apply 9 Cancel Resps 3 116 838 632 866 Co 8 8 7 8 8 8 10 2 Row 100 0 23 5 23 9 31 1 Index 100 0 89 0 99 7 115 9 Indians Audience 000 826 172 190 233 232 Resps 1 405 Co 100 0 Index 100 0 Whites Audience 000 4 102 Resps 6 184 96Col alpes 9 6RoWw 100 0 Index 100 0 White Coloured Audience 000 7 663 Indians Resps 10 705 Co 24 6 AR
36. reative execution in ant given medium Page 41 Index A percent in relation to a norm of 100 For instance a 123 Index means that a number is 23 higher than the norm An 83 index means that a number is 1796 below the norm LSM Living Standard Measurement is one type of segmentation tool based on wealth access and geographic indicators Wealth is used and dependable differentiator and the AMPS variables are particularly suited to this P Population Also Universe Group of persons being investigated form which some are selected in a systematic fashion to form a sample Population weight This weight is used for questions that are directed at the individual in the questionnaire The question normally includes the word PERSONALLY or YOU For example How old are you Psychographics 1 Theclassification of people in the audience of media on the basis of their personality characteristics and attitudes which impact on their life styles and product purchasing behaviour see Life Style and demographics 2 Amethod of population segmentation based on attitudes values beliefs and lifestyles R Reach The number of persons within your target market who are exposed to the advertiser s message at least once This is usually reflected as a percentage and is often referred to as coverage or cover Reach No of Readers Viewers Listeners x100 Number of targeted consumers S SAARF South Afican Advertisi
37. s or lodgers may be included as members of a household provided that they have at least one main meal a day communally Resident domestic servants are excluded and regarded as forming a household of one or more persons in their own right Winning Media 3 Household decision maker weight If the analysis is to be done to establish media usage or other personal demographics for household questions then the Household Decision Maker Weight should be use filtering on the HH Decision Maker Male Head of Household or Female Housewife The wording of the question and the purpose of the analysis determines whether analysis by Household Decision Maker Weight is meaningful Household purchaser Any respondent of either sex who claims to be solely or partly responsible for the household purchases These respondents weighted to households form the basis of the values relating to household FMCG categories Household Weight The weight is used for questions that refer to the household and is used with regard to certain products such as having a stove in the home Household weights can only be used on the following demographics 1 Household Income 2 Geographical Location 3 Living Standards Measures Housewife A housewife is the female household purchaser only The term Housewife does not imply that she does not work l Impact The relative degree of assumed awareness or measured ad noting achieved by a particular c
38. sps 291 25 36 84 Co 0 7 04 0 8 1 7 1 3 ROw 100 0 77 12 9 33 2 21 8 Index 100 0 59 5 112 7 245 5 177 8 Cost 1 103 1 344 799 312 475 TELMAR SP 12 Arithmetic Operators When using arithmetic operators the result is a number Each operand is treated as a number It is converted to numeric on the basis that POSITIVE is 1 and NEGATIVE is O There are four arithmetic operators 1 PLUS 2 MINUS 3 MULTIPLY 4 DIVIDE DIV The PLUS operator The PLUS operator adds two operands together A respondent is counted each time he she qualifies The result being a number EXAMPLE Consider the following COSM STY COSM 1 STY 0 1 Read Cosmopolitan and not Style COSM 1 STY 1 2 Read Cosmopolitan and Style COSM 0 STY 1 1 Read Style and not Cosmopolitan COSM 0 STY 0 1 Read neither Cosmo nor Style 4 TELM R k SPC Page 13 The minus operator The MINUS operator subtracts the second operand from the first operand A respondent is counted each time he she qualifies The result being a number EXAMPLE TV COL TV BW If you subtract the number of black and white television sets from the number of colour television sets the result will give you the additional number of colour televisions to black and white televisions NOTE there must be a space on either side of the sign The multiply operator The MULTIPL Y operator multiplies the first operand wit
39. t Once the required string has been entered into the field clicking on GO and the programme will retrieve the matching data Search Options OF X Search text contains Only Search Selected Search Again Any keyword C All keywords C Exact phrase Starting with Tree options sese Description Clear choice tree C Code Search Selected Save settings as default w OK 3 close There are multiple search options Any keyword Enter one or more words and the search engine will find any descriptions by default or codes if selected that match the word s Equivalent to coding or All keywords Enter more than one word where all must be found in a description in order to meet the search criteria For example in this instance Beer Heavy will only find heavy beer users whereas in Any keyword it would find all heavy users of anything plus all beer drinkers Exact phrase To search for an exact match There is the option of searching for multiple exact phrases by entering a semi colon between the phrases Start With Enter one or more words and the search engine will find a description starting with what you ve entered in TELMAR SPC Page 5 Filtering The program can filter on multiple items by placing a semi colon between the required items For example Filtering by Print Cinema Outdoor in the Selection Tree will collapse the display from the full listing to on
40. t order Audience 000 C Sort Ascending Sort Descending Starting at row 1 Apply to all tables Sort options Set as Default Reload Defaults wf Done 36 Cancel TELMAR KSP 11 CPT Calculation In order for the user to do a CPT Calculation they need to following the following steps e Place the target market in columns e Select the required publications in rows e Crosstab the run e Take out data items not needed via the option Data Items e user then clicks on Edit cost THT Costs A blank screen appears The user Bleed Filename PENES Add Cost needs to click on Add Cost Now ZER Delete Cost 1 1 the user needs to click on the block Edit Costs FH TNT Costs with the 3 dots in to activate the latest rate cards The user needs to know where the saved print rates are saved on either the network or locally If the user want Filename Cost 1 4 Color Page Mo FATELMARAiPrintRates AIRY200812 AIR TRC Add Cost to use the latest rate card it will Delete Cost appear at the end of the list of rate Edit Costs cards o Note Ifthe user selected AIR Average Issue Reader titles they have pese Done to select AIR rate cards The rates are not different for Regular Readers vs Average Issue Readers just a different calculation e The TNT grid will not show the CPT calculation yet The user now
41. ted Page hraugh Just Glance Business Day Read Right Through Read qcade BDAT BDAT 6091 6092 6093 605 Most Read Some Read Selected Page hrough Just Glance The two publications in the above example are Beeld and Business Day Within the title you must code the different attitudes you wish to tab Firstly you define the title for this row followed by a double separator as Thereafter a definition for each of the attitude items to compare separated by a single Within the coding cell enter the Qcode function with parameters as follows The first parameter is the question asked the second is a filter code that is AND ed to the attitude statements The parameters thereafter correspond to the attitudinal titles given in the Title cell In our example Beeld Read Right Through corresponds to code 6071 A successful tab of QCodes will be a grid similar to the one below Default Data Items Stability Highlight Significant digits Male Filter Read Right Read Most Read Some ReadPage Through Just Glance Through Selected Beeld audience Resps ROW Business Day Audience 108 46 D TF B ot Ten Resps 117 54 26 22 9 3 3 Too 100 0 42 9 27 5 18 3 5 9 2 5 2 8 The only data items available are Audience Respondents and Row percentage TELMAR K SP 32 Prototypes The Prototype function allows the user to create a prototype vehicle publication from another publication or pie
Download Pdf Manuals
Related Search