MNM03702 exam revision questions and answers

The four basic levels of measurement are: nominal, ordinal, interval and ratio scales.

Comparative scaling techniques require the respondent to compare two or more objects. They are: Paired comparison scales Rank order scaling Constant sum scaling Q-sort scaling only one object or standard is evaluated in non-comparative scaling techniques. Non-comparative scaling techniques consist of graphic rating scales and itemised rating scales. In the case of the graphic rating scale, the respondent rates a specified item by means of a mark on a line that stretches from one extreme to the other. In the case of the itemised rating scale, the respondent is given a short description of different object categories from which to choose the description that best suits his/her rating of the object

Paired comparison scale - the respondents are given products in pairs and asked to indicate their preference in each pair. - e.g - Which brand of chocolate do you prefer? - cadbury's or nestle Rank order scaling - ranks various objects in order of preference according to the attitude being tested. - e.g - rank five chocolate bars according to taste. Constant sum scaling - the respondent must divide a given number of points among the objects under discussion. - e.g. - divide 100 points among the follwoing cellphone brand name to indicate the quality that you associate with each one: q-sort scaling - quick discrimination among a relatively large number of objects -e.g - a number of statements (between 60-140) about attitudes are written, each on a separate card - these cards are given to the respondent who must, fro example arrange them in 10 piles according to preference from 'strongly agree' to 'strongly disagree'

Step 1: Determine the tolerable error – This is the tolerable difference between the sample estimate and the population parameter that the researcher is willing to accept. In other words, it is the precision or accuracy of the estimate that the researcher will allow. • Step 2: Specify the confidence level – This is the degree (or percentage) of certainty that the researcher wants to attain in the estimate of the population parameter. If the researcher wants to be 100 per cent sure of the results, the whole population must be studied. Step 3: Determine the Z value associated with the desired confidence level: – As soon as the confidence level has been specified, this is done using the standard Z Table • Step 4: Estimate the standard deviation of the population: – This can be done in three ways: • Carry out a pilot‐survey that will give an indication of the standard deviation; • Use the standard deviation of similar studies that were undertaken previously as an indication for the proposed study; • Guess the standard deviation, using the ‘rule of thumb’ that says the standard deviation is one‐sixth of the range of the distribution of the sample average. Step 5: Use the appropriate statistical formula: – The formula for calculating the sample size is: • Step 6: Draw the desired sample.