Tabulation comprises sorting of the data into different categories and counting the number of cases that belong to each category. The simplest way to tabulate is to count the number of responses to one question. This is also called univariate tabulation. The analysis based on just one variable is obviously meager. Where two or more variables involved in tabulation, it is called brivariate or multivariate tabulation. In marketing research projects, generally both types of tabulation are used.
The tabulation may be done by hand or by machine or some part by hand and the other by machine. The number of tabulations will depend on the number of variables, while the number of responses to a question will depend on the sample size. If the number of variables in a survey is not large and the sample size is small, then tabulation by hand is preferable. On the other hand, when the number of variables involved is large as a result of which there may be several relationships and the sample size is large, it will be more appropriate to opt for machine tabulation.
Hand Tabulation: Normally, hand tabulation is carried out with the help of a tally sheet. Consider, for example, the following question: How many movies did you see last week? For z sample of say, 60 respondents, the tabulation may be as shown in Fig.12.1
| No. of Movies seen | Frequency |
| 15 | 9 |
| 13 | |
| 10 | |
| 8 | 5 |
| 5 or more | 60 |
Fig.12.1 Form of Hand Tabulation, with Tallies The hand tally can also be used for preparing cross-classification tables. Consider, for example, the relationship between the economic status and the number of moves seen.
It may be noted that the total column at the extreme right in Fig.12.2 contains the same figures as shown earlier in the case of straight tabulation (Fig.12.1).
| No. of movies Seen | Rich | Middle | Poor | Total Seen |
| 9 | ||||
| 15 | ||||
| 13 | ||||
| 10 | ||||
| 8 | 5 | |||
| 5 or more | 10 | 20 | 30 | 60 |
Machine Tabulation: In the case of large scale surveys, where a good deal of data from a large number of respondents has been collected, hand tabulation will be bothersome and take much longer to complete. In such a case, the method of machine tabulation is used. For this purpose, mechanical sorting and tabulating equipments available with such concerns as the Hindustan Computers Ltd and Remington Rand are used. These equipments include key punches, sorting machines and tabulating machines.
The main advantages of mechanical tabulation are:
- Extensive and large-scale survey can be handled conveniently.
- It takes much less time than hand tabulation.
- Greater accuracy in the sorting and counting is achieved with the help of machine tabulation.
- Monotony and drudgery of tabulation can be avoided by passing on this work to the machines.
- In the case of large-scale tabulation, mechanical tabulation eventually turns out to be cheaper than hand tabulation.
- Cross-tabulation, depicting the inter-relationship between two or more variables can be undertaken easily.
One-way Tabulation
A one –way table, from the first hand tally given earlier, is given as Table 12.1
Table 12.1 shows both the absolute frequencies and the percentage of respondents. Although it may not be necessary to give percentages, it is a good practice to include them, as they facilitate comparisons. The researcher should ensure that percentage figures add up to 100. He has also to decide up to what decimal place percentages should be given. Generally, figures may be given to one decimal place. Rarely, if ever, do they need to be given to two decimal places. While such figures would be more accurate, they would also be more confusing to the reader. The guiding principle in reporting percentages is that unless decimals serve a useful purpose, they should be avoided.
Sometimes only percentages are shown in the table and corresponding frequencies are omitted. In such cases, it is necessary to indicate the total number of cases on which the percentages are based.
Table 12.1 Movies seen by Sample Respondents.
| No.of movies seen per week | Number of respondents | Percentage of respondents |
| 0 | 9 | 15 |
| 1 | 15 | 25 |
| 2 | 13 | 22 |
| 3 | 10 | 17 |
| 4 | 8 | 13 |
| 5 | 5 | 8 |
| 100 |
Cross Tabulation: It may often be necessary to tabulate responses to two or more questions simultaneously. Such tabulations are known as bivariate or multivariate tabulations, depending on whether two or more than two variables are involved.
In constructing cross classification tables, one has to first determine which data should be given primary emphasis and which should be given secondary emphasis. Data with primary emphasis are normally given in columns while those with secondary emphasis are shown in rows. This order is repeated for higher order tables, i.e. those having three or more dimensions. This convention is almost invariably followed because it is easier to see data when figures follow one another in a column rather than in a row.
Table 12.2 is an example of cross-classification.
| Income Level | Shopping centre A B | Total no. of households |
| Up to Rs.1000 Rs. 1000+ Total | 100 200 | 300 |
| 120 80 | 200 | |
| 220 280 | 500 |
Table 12.2 gives the break-up of respondent households, both by income level and by their preference for shopping centers. A table of this type is also known as contingency table. This is the simplest contingency table with two rows and two columns. The data contained in Table 12.2 pose an important question- Does the preference for shopping centers depend on the income level of households? To answer this question, it is necessary to analyse the two variables simultaneously.
Data shown in Table 12.2 can be transformed into percentages and then these percentages alone can be shown or they can be shown side by side with the original data. The question in a two-way tabulation is which base should be used for 100 per cent, as the data may be percentaged in either dimension. Percentages should be based on totals of rows or columns, whichever is relatively more important. One simple rule in this regard is the cause and effect rule. This rule states that the percentages should be computed in the direction of the causal factor. Table 12.3 and 12.4 give these data.
| Table 12.3 | Preference for Shopping Centres by Income Levels of Households | |
| Income Level | Shopping Centres A B | Total |
| Up to Rs.1000 Rs.1000+ | 33 67 60 40 | 100 100 |
| Table 12.4 | Income Levels and Preference for Shopping Centres (Percentages) | |
| Income Level | Shopping Centres A B | |
| Up to Rs.1000 Rs.1000+ Total | 45 71 55 29 100 100 |
In our example, income level appears to be the causative factor, which should influence the preference for shopping centers and not the other way round. Thus, the percentages should be computed in the direction of income level or across shopping centers. Table 12.3 presents these percentages and suggests that the choice of shopping centers is affected by the income level of households. This table indicates that 67 per cent of households in the lower income level prefer shopping centre B as against only 40 per cent of households in the higher income level. Further, 33 per cent of households in the lower income level prefer shopping centre A as compared to 60 per cent of households in the higher income level. It is apparent from Table 12.3 that there are marked differences in the choice of shopping centers on account of differences in household incomes.
Sometimes, the cause and effect rule leads to the conclusion that percentages might be computed in either direction. In such cases, the researcher has to use his direction. It may also be noted that the cause and effect rule is not always applicable. It may be advisable in a particular problem to compute percentages in a certain direction, but the data might not permit the researcher to do this.
